4343 crossover modifications

[B&KMan]

---And the batteries are where?

---Maybe solder the connection carefully, or figure a way to seal the clip without contaminating it.

[Less than potato, even.... :p ]

Thanks for falcon eyes ;)

-----the dc cables and bat is not actually there... I just omit to buy a resistor 2. or 3.3 mOhms :p
Actually I just submit my layout to expert eyes for detect problem before final soldering... and maybe I not all soldering now because I try to test circuit before at low power and analyser....

it is a real hobby !!! It eat all my time... :D

--- for the clip : is appear in pict because this is step roug set-up...
it is not step when i solder...

see flower clip for time soldering
(sorry for pict)

[4313B]

Solen explain for fast caps, the Dc charge is affecte the response but in worse side... the dc charge on film and foil is downgrade in regards standart circuit... TITAN shock because Giskar is expose a point of view of G.Timber where is consider the best is DC charge in Film and foil....


:banghead: :biting: :banghead:

The best is build 2 versions but the price for extreme version is just over thousand...

so, I hope Giskar the great and Mr JBL ingeneer is wind this TITAN shock...I'm not quite sure how to take this so I'm not going to respond directly.

I skimmed over this thread again briefly and noticed reference to G.T. and the statement that biased film & foil would be best. Please forward me to the original statement so I can review it.

[B&KMan]

I skimmed over this thread again briefly and noticed reference to G.T. and the statement that biased film & foil would be best. Please forward me to the original statement so I can review it.

http://www.audioheritage.org/vbulletin/showpost.php?p=6210&postcount=13

I hope Is not again mis-interpretation of language fence...


Jean

:D

[4313B]

mis-interpretation of language fence

Ah! Nope, makes sense now. This comes back to the point I made about various capacitors with the same published dielectric being "different". Your Solen guy may very well be right that biasing his Solen poly and foil may be detrimental. Unfortunately it is up to you personally to make that call after listening tests. I've had several long conversations with guys who live and breath caps and they have mentioned that once you get to a certain level things become merely different as opposed to "better" or "worse", if that makes sense. It would also appear that there are fewer capacitor manufacturers than there are branded capacitors. Dayton is allegedly Bennic for instance.

I did bypass the Dayton MPP with Dayton PP&F and the result wasn't very nice. I replaced the Dayton PP&F with the AudioCap PP&F and the difference was astounding.

Same published dielectrics with different sonic signatures. One sounded great to me, the other sounded not so great to me. Someone else might think the Dayton PP&F sound perfectly fine. That's primarily why I've bowed out of recommending any specific combinations. I have no idea what capacitor combinations work best with which transducers and which power amplifiers. I am thoroughly confident no one else does either. :p The combinations are staggering to contemplate. Hopefully you will stumble across a combination that works great for you. I'm personally perfectly satisfied with

biased Solen MPP
or
straight SoniCaps
or
biased Dayton MPS bypassed with AudioCap PP&F

H/K back end. JBL direct radiator transducers. I have yet to try the RelCaps straight for JBL compression drivers and that project is months away.

[4313B]

I'm personally perfectly satisfied withFor now that is. ;)

:p

[B&KMan]

For now that is. ;)

:p

Hi, Giskard,

For a moment, I haved high level stress because I buy many bock of caps based on this remark and a couple other of great members...


You are lucky man... :applaud:

I hope to go in good way me too...

p.s. I found a 3.3 Mohms 1/4 and 1/2 watts resistor for circuit battery...

what is best ???

thanks again.


Jean.

[4313B]

1/4 is fine


You are lucky man... :applaud:Why? Did I narrowly miss another beating or something? :applaud:

[Ian Mackenzie]

Jean,

How is your crossover going.

I tested the original equivalent 3145 built auricap crossovers today and they perform to JBL attentuations specs. I was busy today but played them on a movie "Hunt for Red October" in DTS tonight. The Russian choir work comes across very well.

Ian

[B&KMan]

Jean, How is your crossover going.


Hey, I try to send info this week-end but this www is not aivailable :blink:

Well first is big jobs and I hope this build is fast but it too much time for quick build-up so I install and try to works as off time... but I have a familly with 2 young kids, and camping time and party kids, and, no time for follow project...

So I have one section (LF) completed (90%) and rest of one speakers is good implementation...

--------

Well many parameters is important as you know...

The network is put in 3 plates granit 12" X 12" so where you place this monster ???

what is a best build box for maximum inertia and less transfert vibration and try to better overall vibration speaker...


Well after many time reflexion I send pict of my box project...

This idea is simple but little difficult to realize...

1--- create mass inertia with less damping factor modal response.
2--- fixation network with improvement back speaker inertia.
3--- no adding box with adding resonnance speaker or other coloration....
4--- TRy not create a box radiator generator vibration....

so My idea is put back box with curve top panel... (see pict) This box contains network fixing epoxy elements...

1--- mass inertia...
2--- fix vibration caps and create fast leaking path...
3--- create mass inertia and fast path for speaker vibration.
4--- create diffusor back speaker for dissipation of back standing wave cerated on corner...
5--- protection box

What your feeling of this idea greats members ?

Jean...

[B&KMan]

Why? Did I narrowly miss another beating or something? :applaud:


hey !! the nirvana is futile ???

:D



I'm personally perfectly satisfied with ....

Jean...

[B&KMan]

I tested the original equivalent 3145 built auricap crossovers today and they perform to JBL attentuations specs. I was busy today but played them on a movie "Hunt for Red October" in DTS tonight. The Russian choir work comes across very well.

Ian

Do you have a pict ???

More taste report ???

mmmmmm I want more meat...

[Ian Mackenzie]

I do but would prefer to keep under wraps until its fully complete and then start a project thread.

Ian

[Ian Mackenzie]

More meat, a taste test..some vino . Yes.

Oky docy,

I play again tonight.

Lesson number 1. 4 way speaker easy to make phasing error.:confused:

Nothing more annoying easily fixed...but we have a surprise to eliminate such things in the project.:D.

Actually I can't tell you what it sounds like because I can't hear it......yes the power was plugged in...Muhhahahahaha....a joke.:banana:

No really incredible....as a well known friend says like peanut butter and chocolate.......

There was no inkling to tweak, its was so relaxing but at the same time engaging where you really start to listen to the music and forget about the speakers and everything else....Call it being there.

Sez below is the crossover board and my assistant Otis. The board is yet to be installed in the external enclosure. At the moment its just hooked up for some tests. For those interested I would describe this as a difficult thing to make and requires patience and a lot of TLC. Details to be discussed in a future thread.

The interesting thing and reason for posting the 3rd image is all the faders are set at zero (0) db. I can't recall one instance on the forums where anyone plays their 43XX monitors with the faders set flat, particularly 4 way systems.

I have measured the response and its absolutely flat and sounds perfectly balanced. I attribute this to the crossover parts quality, careful setup and calibration of the L pads and of course a brilliant amplifier.

This will be discussed in the project thread to be published soon.


Ian

ref system;

JBL 4345
BOSOZ sys X Preamp
Passlabs Aleph 60 Power Amp

[rek50]

Ian, Nice component mounting method. If Otis is anything like our three "Meeeezers" (Siamese), He has trained you plenty. LOL.

[Ian Mackenzie]

Well lets just say he's very smart and rungs rings around some people.

Ian

[B&KMan]

Well lets just say he's very smart and rungs rings around some people.

Ian


Hey hey Cool picts and info... and I feeling this is jimmy smith the cat album inspiration ????




of course it is better classical presentation board... look better...

I'm surprise you put foam computer part under caps ???

nice finish on mdf....

I hope you not evaluate sound with horibily cheap crocodile clip cable ??

muaaaaaa :D

next soon...

I plug the LF section on I try to perform a couple test mesure in next days...

[Ian Mackenzie]

Looks are not everthing visible.

The panel is aluminium. The pads are medium density cell rubber with the primary purpose of providing a moulding for the capacitor to seat on without bruising.

The crock clips are only temporary..I find them particulary useful for inital testing before final assembly

It was not an acid test and only as narrow insight for a mad scientist to get his rocks off.

We will disclose the wider perspective on upgrading a 4343 to a 4344 soon enough.:bouncy: .

Ian

[B&KMan]

Lesson number 1. 4 way speaker easy to make phasing error.:confused:

Nothing more annoying easily fixed...but we have a surprise to eliminate such things in the project.:D.



You talk really right !!! maestro

Hey I just finish a simple test work on LF section...

Test impulse reveal invert phase response in regard of the traditionnal 3141 Network !!!!!

with the 3141 network the cable black and green is for LF and black put to black and green (+) put to red
In this: the driver is go back if positif impulse s run....

Big surprise !!!
if I put same set-up with my new network the result is 180 invertion of phase.

after check multiple time the connections all is OK !!!

Anybody have a hypothesis of this phenomenon???


Thanks Jean.

[B&KMan]

Looks are not everthing visible.

The panel is aluminium.

Ian

AHHHHH... honor is save.... :D :D

well, the rubber is interesting approach in particulary if you put alu. plate...

double aspect: (1) absord and dampind vibration and (2) save deformation caps by tensile stress due to tarap... :applaud:



Ian, you expose your set-up audio but you not expose your instrumental set-up...

My curiousy is infame :D but I'm not resist of this question...

so please.... :applaud:

Jean


p.s. do you have not stressed to lead caps is internal short-circuit by unfortunatly contact with alu plate ???

[Ian Mackenzie]

Please post picture and diagram

[B&KMan]

Please post picture and diagram

This diagram according to Giskard works. modifying for cascade caps in Dc set-up

[Ian Mackenzie]

The outputs of the mid, hf and uhf filters should all have inverted phase 180 degree with respect to the woofer.

Here is my lab in theTardis!

Ian

[B&KMan]

The outputs of the mid, hf and uhf filters should all have inverted phase 180 degree with respect to the woofer.
Ian

Thanks for pict...

Well, black post on all driver is hot the black cable on 3141 is (-) is connect to black post driver (+) and green is red post driver (-)

the rest of driver of 3141 is (+) signal to (+) post driver (black in all case)

but with my new LF the resonse phase is inverted !!! the (-) signal is connect to (-) post for same motion of 3141 ???

Yiah the rest of new network is not connected... It is possible this affect phase invertion ???

SEE wiring pict

[Ian Mackenzie]

I am not quite sure what you are saying... its saturday morning..I need a coffee...black.

You should alway have load on each crossover band when testing or risk blowing up amplifier.

The 3143 wiring is confusing.

Just think of it as woofer in phase, mid, horn and slot all out of phase.

Ian

[B&KMan]

I am not quite sure what you are saying... its saturday morning..I need a coffee...black.

The 3143 wiring is confusing.

Just think of it as woofer in phase, mid, horn and slot all out of phase.

Ian

Yiah coffee needed and yes 3143 is confuse but one thread just for this expose clearly the 15" is out of phase ans the rest is in phase....

The black post driver of each is positive post. So normally the black (neg) cable is connected in positive post (black post) 15" driver ....

My feeling is the coil wiring is not same lengh and change the phase (not polarity)

normally the 15 with positif impulse the driver cone is go back... :blink:

[B&KMan]

Hi everybody not too tired by this long long thread... :applaud:

here my new back-end pannel with figure the network support...

Sorry for dust .... :D :D :D

[Earl K]

Hi Jean,

Since I don't have a 4-way of any sort , I've ignored a lot of this and tried to keep out of the way ( because, at the very least I hate phase issues ) but now ,,,


,,, Hey I just finish a simple test work on LF section...

- with the 3141 network the cable black and green is for LF and black put to black and green (+) put to red
In this: the driver is go back if positif impulse s run....

- That's just what I would have expected ( for the direction of cone-movement, when a positive impulse is applied ).

- When I look at the 3141 network schematic , the green (LF) wire is (+) as you state,
- FWIW; I would hook up the woofer ( when using the stock network ) to get a positive movement from that woofer diaphragm / with an applied positive impulse . Also meaning, I would hook up that green wire ( from the N3141 ) to the black speaker terminal and flip all the other 3 sections around to create the necessary opposite polarities.

- Why ? I don't like woofers that suck air :D ( rarefaction sounds quite unnatural to me in the lower frequencies) Anyways, this is somewhat beside the point .


Big surprise !!!
if I put same set-up with my new network the result is 180 invertion of phase.

after check multiple time the connections all is OK !!!

Test impulse reveal invert phase response in regard of the traditionnal 3141 Network !!!!!

- Please be more specific Jean , ie ;

- What does "same setup" mean ? ( your new network wires have different colors )
- Did you connect your new, single section lowpass filter directly to the woofer ( after first disconnecting the original N3141 network ) ?
- Did you connect the wire with the red dots ( on white ) to the woofers red terminal or black terminal ?
- Have you looked at the original N3145 schematic ( for the 4345 ) (http://www.jblproservice.com/pdf/Network%20Schematics/3145%20Network.pdf) to find out how JBL labels the LF leg of that network? ( It's the primary reference, forward-movement leg of the crossover )



Anybody have a hypothesis of this phenomenon???

- I've just suggested a couple of hypothesis ( but I am lacking some necessary detail about your test methods to go further ) .

- Since you've added a Zoebel to the original network , I would also consider the N3145 as a partial model for your phasing setup.

- For my own interest, ( since I don't want to reread all the posts in this long thread to find out when,where & why ),,, please tell me why the original 72uf cap in the N3141/3 ( woofer leg ) was increased to slightly more than 90uf, as in the N3145 ?

;)

[B&KMan]

Hi Jean,
- Why ? I don't like woofers that suck air :D ( rarefaction sounds quite unnatural to me in the lower frequencies) Anyways, this is somewhat beside the point .

Hi Earl and thanks for your reply...

Well I refer to
http://www.audioheritage.org/vbulletin/showthread.php?t=487&page=6&pp=15&highlight=4343+polarity

post 83 and 84.... If you look the result the LF is better respons if you put invertion polarity: better dynamic range ans correct phase response....

for my test I perform new test maybe next day and I post pict and result...

But I have a dual channel analyser and generetor impulse integrate... I put + signal into red post speaker and (-) signal into black post speaker

in this the (-) signal is appear into black cable (on network 3143) and (+) signal into green cable.

The (+) post driver on 15" is black and (-) is red (this is same for all driver of 4343....

Well in factory JBL for 4343 the (+) signal is connected in (-) post driver (red in this case) so if you put (+) impulse the driver cone is back ...

Well is I connect oput amp signal to new LF section network, ans respect the invertion signal polarity for driver : the drive cone is go foward !!!

so for the same electric path connection the phase is inverted of 180 degree...

But erratum humanum es so I reperform test tomorow...

:cheers:


Jean.

[B&KMan]

Here a little pict of test physical set-up box for this godzilla network :D

:cheers:

[Earl K]

Earl K asked : - For my own interest, ( since I don't want to reread all the posts in this long thread to find out when,where & why ),,, please tell me why the original 72uf cap in the N3141/3 ( woofer leg ) was increased to slightly more than 90uf, as in the N3145 ?

- I think I've answered my own question . Jean, according to the above posted Charge-Coupled™ schematic , your network is a take-off on Giskards' evolved N3145 network .
- Is that correct ?

- Anyways, in looking at the above ( referenced ) network created by Jean, there seems to be a typo ( at least I hope it is )

- Right now , that DC-Biased collection of caps has an approximate value of just 15 uf ( 14.9 uf ).

- I've corrected the value / assuming one does want an @ 60uf value capacitor in that position / before the 2121/2 midbass unit .


:)

[B&KMan]

- I think I've answered my own question . Jean, according to the above posted Charge-Coupled™ schematic , your network is a take-off on Giskrads' evolved N3145 network .
- Is that correct ?


YES !!



- Anyways, in looking at the above ( referenced ) network created by Jean, there seems to be a typo ( at least I hope it is )

- Right now , that DC-Biased collection of caps has an approximate value of just 15 uf ( 14.9 uf ).

- I've corrected the value / assuming one does want an @ 60uf value capacitor in that position / before the 2121/2 midbass unit .

:)

Thanks Earl, it is effectively Typo mistake by many copy & paste action :o:

Your correction is good and correspond to reality ( see pict)

Thanks for relevant for other member...


Jean.

[B&KMan]

OK DOKAY other small news the project is still foward again a little bit :D


Here 2 pict of look test piece box wood for estimation ovelarp and smeel look after end jobs...

(of course any comments are welcome...)

Jean.

[Ian Mackenzie]

So Jean,

What is the Golden answer to making your network sound wonderful in your system system as one speaker even if a four way?

Think....about all the clues and what you have learned from JBL schematics and information on JBL monitors.

Its easy if you know how. And know don't even think about using the native ear.

A simple measurement tool as shown is quite sufficient as we are only concerned with measuring differential change in sbl.

Adjust the Pads up full on the slot, turn down the horn and mid to minimum.

Create a pure sine wave of 15 kherts and arrange immediate near field measurement on a tripod (triffod!) of slot.

Adjust signal generator for Plus + 4 dba on meter (on 4344 or see foil cal of 4343) with L Pad up full then carefully turn the L Pad back until 0 dba.

Place meter same distance and in front of horn with lens on, adjust frequency to 5kherts, place masking tape over slot to avoid cross modulation and set Horn L pad for 0 dba.

Repeat as above and for mid cone adjust for +3 db on meter and L Pad adjust for 0 dba. Several mid cone measurements maybe necessary due to longer wavelength. Pulse and FFT measurement is not necessary nor appropriate for this type of measurement. No you won't see this else where on the forums however a search of 4345 plans might illustrate the point.

I guarantee this will improve your listening pleasure beyond belief.

Ian

Edit. Jean the reason for highlighting this is I did these procedure last night and it works like a charm. I am only trying to help your cause..why spend so much effort and waste it all on a incorrectly calibrated system. Some will no doubt jump to the conclusion this a crude measurement and pull out their 1/3 octave analyser. This is a different kind of measurement and I refer to the above details.Kiss= Keep it simple stupid..works everythtime:)

[Earl K]

Hi Jean

So , is this semi-circular thingy that is stuck to the back of the cabinet supposed to hold all 3 of your 12" by 12" granite plates ( all at the same filled with sand ) ? I'm quite confused by the smallish size of it . :blink:

(A) A few observations FWIW ; ( since you seem to be past the point of no-return on my first point ).
- I would (have) like(d) to see :

(i) Your networks contructed so that you can "connect/disconect" the bypass caps and hookup wire .
- The reason ? this way you can get a "feel" for what their sonic contributions are to the final presentation .

- Right now whatever your end result turns out to be ; you ( & the rest of us ) won't know ( learn from ) what the individual parts "of your great effort" actually contribute . This leads to a lot of "eye-glazing" no matter what one says after the fact since there are a lot of extra variables added into the customization .

(ii) I would like to hear about your of views on the sonics' of your new networks before you take the final step of "damping-down" all possible sources of external "microphonics .

(iii) I also would like to see a study ( with your B&K gear ) of what the net results are of encasing Solens' Caps / Coils in ;
(a) Parafin Wax
(b) Epoxy
(c) Sand
(d) Wood ( Solid Soft & Hardwood )
(e) Some other midweight "energy-sump"

- As you mentioned on the previous page / JBLs' older stock PRO networks had a better dampening approach ( by encasing the large value pieces in epoxy ? ) You are one of the first around here to care about that . :D

(iv) Worth considering about possible induced microphonics into capacitors is the effect of DC biasing . One purported benefit ( from the K2 S5500 copy ) to biasing is that it pre-"swells" the capacitors' dielectric layers.
So; expanding on that thought :

- If the caps inside expands against a flimsy wall ( like Solens ) / the resonant microphonic signature should be somewhat lower than a hard wall. This will need a specific approach to "energy" dumping to avoid so-called "blooming".

- If the inside expands against a firm wall ( like some hard shell caps / or especially those already encased in epoxy ) / the resonant microphonic signature should be much higher ( perhaps supersonic ). This will need a different approach to "energy" dumping to avoid "UHF ringing".

Example - My main "go to" RC surplus capacitors have the metallized polypropylene core encased into hard epoxy . They start off very quiet & once "DC biased" they are even "quieter" , offering ( I believe) a much deeper sense of resolution .


Some Food For Thought <> ;)

[B&KMan]

A simple measurement tool as shown is quite sufficient as we are only concerned with measuring differential change in sbl.
Ian


Hi again Ian and thanks for many tips...

(p.s. I love your pict member :D )


:( AIIIIII :(

I really sorry to read your post for manys reason and I feeling to start a another long long bing bang controverse...


1--- First the SLM Radio Shack have a really poor response in high frecquency... ( I remember my friend I have one maybe it ispossible to create comparaison...).

The flat response is not up to 15K it is ( in my memory below 12K)
and the lower is not good below 100 or 80 Hz...

Because the feeling of pitch or balance tonal is critical (0.2 -0.4 dB) the precision is THE difference. ut of course it is possible to feel coarse set-up in this...

the SLM distord the field response and it big eratic alteration response mesurement...

2--- Your method is classical for mesure acoustic response with minimise background interferance create by room.

3--- Unfortunately the driver is fluctuate consistently in area of freq response... So 15K is maybe 2 dB higher of average power level of this driver of maybe below couple dB... In according .4dB variation is the best set-up the methos of only one freq is too aleatory... add imprecision tool ...

4--- Well your method is explain in rane electronic crossover... and explain many tecnic for phase and SPL calibration with simple SLM... but it is really coarse adjustement and not create magical set-up...

5 the magical set-up is very expensive in money and time...
a--- big high end instrumentation and ultra high level mic (free field or and pressure field) 2 positions according to ansi or ISO standart... in regard of sweet height spot... generate pink noise (with high accuracy) and average over a 30 secor more record mesurement. (slow = 2 seconds but it is not satisfying the stabilisation of low frequency for accuracy mesurement...)

b--- more records positions more accurate mesurements: mic is keep the node and modal response room, so the many different positions is cancel this effect by averaging result many positions...
Of course you catch the response room and it is important factor, distance and angle and plane surface modifying response curve and it is important to set-up Lpad in this reagrs for better fla response possible... not theorical driver response....

Of course the more sofistication analyse modal response room expose where specially bad or poor response area... If you have push the button , the treatment of echo and other modal parameters room influence the result...


Finnaly the level of finest FFT analyser or RTA expose 1/24 oct is really surprise response in regard of 1/3 octave... and discovery the value is in realy just a small peak frequency to boost result Consequence : more details more precise value...

But YES I go in your point of views, the subtil tune-up L-pad reveal the finest of nature of speakers and sound...

After works this set-up, I realise many news details in music, subtility sound, many high frq. details but L-pad is lower than before perform test !!!

Best regards,

Jean.

[B&KMan]

Hi Jean

So , is this semi-circular thingy that is stuck to the back of the cabinet supposed to hold all 3 of your 12" by 12" granite plates ( all at the same filled with sand ) ? I'm quite confused by the smallish size of it . :blink:



---Sorry I limit of my english this piece is not a final board : it is just gabarit for assume look and better easy to handle by chunky final board...

--- Yeah, I realise couple of error because too small experience electronic...

--- Of course I perform many tests and adjustement (I hope to not change caps of resistance...) for tuning response before final sand operation...


--- I perform later comparison device exposure of vibration and attenuation resonse in epoxy and epoxy-sand and plain set-up but my experience is more is better...

For Wax, Solen is not recommend because the eat destroye the caps...
epoxy is limited because if one conponent is broken or fail, I'm not interested to put entire network in garbage... For my coil I paid extra charge for 3 dipp in emanel for solid core coil ( yuk yuk yuk )

But you help my strongly in this project and I'm weeling to perform free a set of vibration response : one big day for work perform so later but try soon )...









(iv) Worth considering about possible induced microphonics into capacitors is the effect of DC biasing . One purported benefit ( from the K2 S5500 copy ) to biasing is that it pre-"swells" the capacitors' dielectric layers.
So; expanding on that thought :

- If the caps inside expands against a flimsy wall ( like Solens ) / the resonant microphonic signature should be somewhat lower than a hard wall. This will need a specific approach to "energy" dumping to avoid so-called "blooming".

- If the inside expands against a firm wall ( like some hard shell caps / or especially those already encased in epoxy ) / the resonant microphonic signature should be much higher ( perhaps supersonic ). This will need a different approach to "energy" dumping to avoid "UHF ringing".

Example - My main "go to" RC surplus capacitors have the metallized polypropylene core encased into hard epoxy . They start off very quiet & once "DC biased" they are even "quieter" , offering ( I believe) a much deeper sense of resolution

Well for this points it is really theorical and long response... the fundamental point is any solid is vibrate... Well, what better tecnique for reduce this natural response ??? The damping fator is same of ration of echo response in room to higher , the sound is run in bell concert, to dead, the sound is dead too... so the ideal approach is found good time damping factor and faster drain power vibration appear around or inside caps... Yeak epoxy is really hard and drain fast energy but I have not mass energy to put down level response... Let me know it you undertand ....

Jean.


BTW the shape of back is ideal dispertion panel so the standing wave on back speaker is broken and dissipate... Better flat response and less noise floor: better imge stereo, and details infor... (I hope in theory)

:cheers:

[boputnam]

Anyways, in looking at the above ( referenced ) network created by Jean, there seems to be a typo ( at least I hope it is ) Hey, Earl...

Sharp eye. I've edited the .jpg. Can you take another close and careful lookie at this - it's important it honour the design.

Cheers...

[Earl K]

- Will Do ( this pm ).
- I need to first assemble/compare the 3 varieties/styles of upper filter topologies.
- This also means assembling a .jpg collage incorporating the two evolutions that Giskard did ( the type minus the autoformers ) .

:cheers:

[B&KMan]

Hey, Earl...

Sharp eye. I've edited the .jpg. Can you take another close and careful lookie at this - it's important it honour the design.

Cheers...

Yiah realy sharp...

bravisimo...

Jean

[Earl K]

Well this is getting messy . :blink:

Jean

- Here's another mistake found in the UHF section .

- I haven't scoured your drawing for all others but this one stands out .

Explanation : Typically , in an Lpad , the inline ( series resistor or "buildout resistor" ) is placed on the "source" side ( ie; amplifier, drive side ) of the network because it is usually present to correct an impedance mismatch / usually caused by placment of the parallel ( or Shunt or Conjugate ) resistor across the line .

Jean ; please nuke ( remove ) the above drawing ,yes, the one you just placed with the correct 120uf value.
- I think it's best to wait for some more corrections before putting a new one in.

- For instance ( and this makes quite a difference to me ) . Each of your cascaded capacitor setups actually used just one resistor to deliver the 9 volt bias / instead of the usual 2 or 3 resistor that is shown on your schematic . So your reality is a different approach from the schematic.
- Now, the reason this divergence matters to me is that the 2 or 3 pairs of capacitors now have a central connection point or crossover pathway . This is the way you have shown pictures of your wired setup ( & FWIW ) the way I prefer to connect all cascaded & biased teams of capacitors.
- Why my preference : I believe that having that central connection ( the resistor "tie" point ) helps average out timing errors . ( Okay, This is truly geek-based tweaking stuff , only for those who care to listen this much to the effects of capacitors ). :p


:D

[Ian Mackenzie]

Hi again Ian and thanks for many tips...

(p.s. I love your pict member :D )


:( AIIIIII :(

I really sorry to read your post for manys reason and I feeling to start a another long long bing bang controverse...


1--- First the SLM Radio Shack have a really poor response in high frecquency... ( I remember my friend I have one maybe it ispossible to create comparaison...).

The flat response is not up to 15K it is ( in my memory below 12K)
and the lower is not good below 100 or 80 Hz...

Because the feeling of pitch or balance tonal is critical (0.2 -0.4 dB) the precision is THE difference. ut of course it is possible to feel coarse set-up in this...

the SLM distord the field response and it big eratic alteration response mesurement...

2--- Your method is classical for mesure acoustic response with minimise background interferance create by room.

3--- Unfortunately the driver is fluctuate consistently in area of freq response... So 15K is maybe 2 dB higher of average power level of this driver of maybe below couple dB... In according .4dB variation is the best set-up the methos of only one freq is too aleatory... add imprecision tool ...

4--- Well your method is explain in rane electronic crossover... and explain many tecnic for phase and SPL calibration with simple SLM... but it is really coarse adjustement and not create magical set-up...

5 the magical set-up is very expensive in money and time...
a--- big high end instrumentation and ultra high level mic (free field or and pressure field) 2 positions according to ansi or ISO standart... in regard of sweet height spot... generate pink noise (with high accuracy) and average over a 30 secor more record mesurement. (slow = 2 seconds but it is not satisfying the stabilisation of low frequency for accuracy mesurement...)

b--- more records positions more accurate mesurements: mic is keep the node and modal response room, so the many different positions is cancel this effect by averaging result many positions...
Of course you catch the response room and it is important factor, distance and angle and plane surface modifying response curve and it is important to set-up Lpad in this reagrs for better fla response possible... not theorical driver response....

Of course the more sofistication analyse modal response room expose where specially bad or poor response area... If you have push the button , the treatment of echo and other modal parameters room influence the result...


Finnaly the level of finest FFT analyser or RTA expose 1/24 oct is really surprise response in regard of 1/3 octave... and discovery the value is in realy just a small peak frequency to boost result Consequence : more details more precise value...

But YES I go in your point of views, the subtil tune-up L-pad reveal the finest of nature of speakers and sound...

After works this set-up, I realise many news details in music, subtility sound, many high frq. details but L-pad is lower than before perform test !!!

Best regards,

Jean.

Your remarks in point 1 & 3 are all irrelevant (wrong). As I said earlier you are only measuring a relative difference, not an absolute value of each discrete driver. If the crossover voltage attentuation is correct then all that is required is subtraction of -4 db for the slot and horn and -3 db for the midrange.

Surprised as you maybe , the Tandy measurement tool is quite capable is detecting a relative SBL difference and as it is affordable I would encourage other members to use it for this specific purpose.

A top engineer can get the job done with even basic tools. Don't over complicate what you are trying to achieve

Dalek

[B&KMan]

Well this is getting messy . :blink:

:D



Earl, you have not only sharp eye , you have a mind too... :applaud:

Your argumentation a resistance dc and constatation for resistance UHF is entire correct...
I fixe the schematic tonight

Let me know other mistake...

thanks a lot !!!!


Jean.

[B&KMan]

Your remarks in point 1 & 3 are all irrelevant (wrong). As I said earlier you are only measuring a relative difference, not an absolute value of each discrete driver. If the crossover voltage attentuation is correct then all that is required is subtraction of -4 db for the slot and horn and -3 db for the midrange.

Surprised as you maybe , the Tandy measurement tool is quite capable is detecting a relative SBL difference and as it is affordable I would encourage other members to use it for this specific purpose.
Dalek

Hi Ian, Maybe I'm not understand your procedure works...

if you put 15K in uhf the mesurement: the response is produre only by UHF so it is not possible to perform absolute comparison at same frequency with HF ??? No ???

My brother is keep mic in dolorama and the error is absolutely same in any mesurment so of course it is possible to produce high degree of precision for same conditions... But it is important for this forum all understand this is not a better approach but yes excellent start approach)... My point is not shocking anybody or put myself in higher than other...
I pain many many weeks for discover with reading and experiencing the difference and why one record response big differt to other by subtil displacement or scale of data...

I try to send later test for explain my point of view...

:cheers: and thanks for all.

Jean

[B&KMan]

Here again great members with , ( I hope ) stunning info pict...

Well Before put by big network box on back speaker, I mesure response modal and resonnance cabinet...

what difference ???

Cabinet is resonnace produce by run speaker in pink nois or sweep sive wave and keep what is fr is exice the cabinet and what amplitude... This is important for catch the part of adding or supressing fr in final playback mode

the modal property is extract motion and spectrum response of specific place on cabinet... In this I have info what exactly comportement backplat speakers without the rest of the response cabinet....

Of course this is permit to keep signature ( 1" med density presswood) material and exclude internal resonnance or other...

of course this is not a complete study and many aspect is neglected... Why ?? TIME bordel de merde the calibration and complete analyse is big big work and for this application a good average is OK for feell the better or worse modification...

Ok resere and theory is pass now look result...

the modal is multipoint mesures (equal distance and same line displaement point mesure)
the hammer impact instrument and accelerometer is produce and keep signal.. the dual analyser is compute the datas and expose result... and a hand with no face is works the process :D

[B&KMan]

Comments result:

Weel no surprise the presswood is specific signatue and is relatively altered by network fixed inside box...

the material is good fixing inside and brce to keep excess natural amplitude response... No surprise is a JBL original box ( :applaud: )

Ok the dark side now... The persistance of signature all in points expose the transfert flavor to the drivers... fortunately the signature is relatively round and not too sharp peak and node...

now part B pink noise run on speaker and listen the resonnace emmision on back panel... The track resonnace is the sum of panel resonnace and cabinet resonnace : no surprise the cabinet resonnace is more strong to backplate resonnance and produce peak signature.... note if you keep signature in upper pback panel the bass is dissapear for expose more med enrgy... again this is a good point and expose the resonnace cabinet is controled and relatively good dispertion response in all freqeuncy...

Note I keep the level mesure of little cardboar plate where the post is fixed... look the difference signature and amplitude... ( the tribo-electric effect is back !!! ) yuk yuk yuk


well I hope this info is ineresting for open put the phenomenon on your sharp eyes members.

:cheers:

[Earl K]

Hi Jean


Cabinet is resonnace produce by run speaker in pink nois or sweep sive wave and keep what is fr is exice the cabinet and what amplitude... This is important for catch the part of adding or supressing fr in final playback mode

***Okay, that part I understood .***

- With music ( or pink noise , swept sine wave ) as a source , one can excite the cabinet ( from the inside ) to find out its' modal resonant points . Then one can proceed to find a way to dampen them down or out .



the modal is multipoint mesures (equal distance and same line displaement point mesure)
the hammer impact instrument and accelerometer is produce and keep signal.. the dual analyser is compute the datas and expose result... and a hand with no face is works the process

***This last part I don't understand .***

- What is the relevance in exciting those modal resonances with a hammer from outside the box ? :blink:

- This approach ignores what is going on inside the box. That's where the energy generator is located :o:

FWIW : I agree that these stock boxes can use better bracing. Your "2in1" solution to firming up the back-panel with your "custom crossover sand-box" is one such approach.
- But, personally, I'd like to see these two functions split, between a dedicated holder for the crossover ( that keeps the plates horizontal ) & better external bracing on that back panel .

- Just My Opinion <> :D ( since I don't actually own any of these things to need to worry about them )

EDIT ( after the second post )

Okay , maybe I get it now . Please, tell me if I'm right or wrong.

(a) The hammer was used to map out & clearly "identify" the backpanels' modal resonances from the point of view of a "worst case scenario " . This was to create a reference point so that one can see the effects from internal dampening.

(b) The second test was with an "internal energy source". This revealed the damping effects occuring because of the cabinets fiberglass and the fixed location of the present crossover.

Jean , what is your conclusion to this test ?

[B&KMan]

Hi Jean

***This last part I don't understand .***

- What is the relevance in exciting those modal resonances with a hammer from outside the box ? :blink:

- This approach ignores what is going on inside the box. That's where the energy generator is located :o:

Okay , maybe I get it now . Please, tell me if I'm right or wrong.

(a) The hammer was used to "identify" the worst case scenario so that the resonances from the back panel were clearly delineated. This was to create a reference point .

(b) The second test was with an internal energy source. This revealed the damping effects occuring because of the cabinets fiberglass and the fixed location of the present crossover.

Jean , what is your conclusion to this test ?

Hi and thanks for interest,


---- the hammer keep signature and movement of back panel without consideration of tuned inside cabinet... in this you keep the real signature and effect produce by back panel.... the hammer produce a implct pluse inside material and the acc is keep this spectrum, amplitude, phase and distortion...

after you understand a specific backpanel is play, you evaluate the entire speaker in normal mode and check if the signature of the backpanel is relevant

if conclusion if yes = the back panel produre big effect in global signature playback !!! (same drum)

if no = back panel not signifiant effect on global signature...

I hope this second result (this is result of good speaker)

if I put big box network on panel, and modifying the signature of back panel if mesur is expose it is not important I will go in this way, otherhand if back panel is contribute stongly of the global cabinet signature, the alteration is tricky to detune cabinet...

And because the cabinet resonnace in a multiple additition signature, of course the even port and inside cabinet encluse is calculate for produce energy sounds (this is a concept of helmholz box resonnace...)

Of course you understant the criterium strong signifiant and non signifiant is relatif and the ideal is no response material and just pure helmholz renonator... but the real is not utopic...

the summary analyse expose -20 differencial in material and internal resonnance so this range is detectable... fortunately the boxNetwork keep energy backPanel and keep energy of flanking panel I hope to brop signature by 10 or (I dream... ) 20 db...

The result is the micro information is drastic more relevant and clear, more deep, and better subjective transient amplitude... (the law of the mass )

But maybe this higher resonnace material is balance the bass resonnance and if is desapear, the bass bloom is more more present...

yurkk it is nothing to easy modified studio monitor...

better comprehension phenomenon ???

:cheers:


Jean.

[B&KMan]

well look next page for other mistake netwoks diagram...

:(

jean.

[Earl K]

Hi Jean,


better comprehension phenomenon ???
Yes, Thanks ! for the above explanation(s) .


The result is the micro information is drastic more relevant and clear, more deep, and better subjective transient amplitude... (the law of the mass )

- What about the wires inside the box ?
- How does one disconnect them from induced microphonics ?

NOTE: I've just downloaded your latest "corrected" N3145 network .

- I'll print it out so that I can look at it outside of this office ( my listening chair ) .

- If I don't look at these type of things ( off-screen ) I miss too many details . :D

[B&KMan]

- What about the wires inside the box ?
- How does one disconnect them from induced microphonics ?



the wire is original come with the JBL network and I do not perform any treatment for vibration control.......

for the rest I expose my in naval war controverse argumentation :D

http://www.audioheritage.org/vbulletin/showthread.php?p=56119#post56119

:cheers:

[B&KMan]

Hey hey hey It me again (yuk yuk yuk )

Well I just finish a painful technical tips evaluation... Of CAPS vibration...
pffff....... I hope you keep this very very expensive ;)




(i hesitate to create a new thread fo this but many info is in regard of modification 4343 project spirit so I run this)

(And it is a cherry for any guys is arrival here... :applaud: )

=======================================


I perform a impac modal response to mesure the spectrum response (FRF) and amplitude response of vibration of 3 set-up caps... and expose the result at your eyes judgement...

1--- I tarap firmy (limit to broken tarap) a metal can caps phillips on thin 1/8 inch clawboard (I dont know what is tecnical term). I put the accelerometer on top of caps and I slamHammer impact instrument into board...

2--- Same on 1 but I put in thick clawboard (1/4 inch.)

3--- Same but mdf 3/4 inch...

4--- same but in my big Network 3/4 mdf plus rock (and eventually sand...)

I send this the pict of 4 piece to expose at my hammer....

[B&KMan]

OKAY

For interpration correct it is important to keep in your mind the dB scale is log Scale so 3 dB is look cheap but it is big real difference so I have shap shot result in lin and dB presention so look carrefully the appelation on pict...

Look the vertical scale too !!! Because I have not idea before perform test what is the max level , I expose a couple difference level so for cross link check carrefully the y scale...

finnally it is very tuff (in particular it is not my language) to expose simply a ton and complex information and interpretation relation... so i hope your help to decode my english...

:D :D :D


======================

first result is thin plate board fixing caps versus thick plate board...

one pict is lin result and other is same but in dB scale ...

[B&KMan]

???????????????????????????????????

The thick layer is induce more energy vibration than thin and appear worse than thin.... why ?????

My hypothesis is the thick is produce vibration but sensitive mass diff. so the energy vibration is not controled and more power than thin board...

This is a perfect exemple of the common feeling is run opposite scientific result.... So as this peak fr your caps is extremely exicting and produce tribo electric, or other distortion, additive, signature... Your caps is sound wood...


next is comparison of thickboard versus the big network response... No surprise the mass and composite layout is cut natural frq. of each material... and distribute energy more equally...

:D

[B&KMan]

ok next is more complicated info and suppose more background theorical...

The pict I expose the time disipation of impluse... This info is critical for articulation and deterministic working bkgnd noise generate by too long time slope... According one note musica come realy close to next and if your time dissipation is more than this time you listen new note and the echo of the first... If it is musical nature of harminic and musical alnguage it is not a language of reproduction... because you add echo material on electrical music... produce more floor level and confusion in electrical and acoustical signal ....

If you look next pct the thin board expose long time dissipation and push up the noise floor .... In big net the time dissipation is fast and expose lower noise floor when next note is come... so better articulation and extend subjectif dynamic... mmmmmmmmm


Jean

[B&KMan]

In regards of previous commentary,
just componend is abbility to broken natural response material

add thi point add the law of the mass and you put more fast time dissipation and more lower peak, more linear response...

one pict is lin and other is linear scale...

Jean..

[B&KMan]

After all you hare won to look the size of BIG NET work...

55 Lbs pound at this pict (more double with sand....)

:applaud: :applaud: :applaud:


Jean.

[B&KMan]

Of course any comments are welcome


:D :D :D :D

Jean.

[Earl K]

Jean ; like I said before ,,,

This is getting a bit messy . :blink:

- I've only really corrected the HF & UHF sections .

- I've inserted Giskards version2 jpg showing his "modified N3145 equivalent circuit" ( HF & UHF sections only ) .

- I've made the corrections to some capacitor values ( in red ) so that they more closely mirror Giskards "non-biased" version . I can't get exact values without totaly reworking your existing values . As a result , in one postion where a value of 5.6uf is called for , I have only gotten the values to 5.5uf .

- I'm not sure what you'll want to do with this information.

-You've increased the net capacitance at 3 points within the Horn circuit. - You might want to changeout what you bought for the proper values ( not really an expensive proposition with Solens of this size ) .

- On the other hand , these smallish differences might not be that "over the top" to the end result since they are within the typical stated tolerances for most capcitors .

- Someone with Crossover SIM software could tell you that info / unfortunately I don't have any ..

- I hope to study the midrange and woofer circuits tomorrow .

:cheers:

EDIT Oh,, the "G" = some value , refers to "Giskard" ( for those wondering ) .

[Earl K]

Hi Jean

- Here are my most recent corrections to your schematic .

- Maybe your substitution of the 20 ohm & 15 ohm resistors for the (2) 39 ohm resistors was just another "cut & paste error" ( divorced from what you actually bought ). I don't know, you'll have to tell us what you are actually wiring up .

- I've also divided out the (2) biased 20 uf caps in the woofer portion of the network . The original N3145 has the original 10uf after the Zoebel (10 uf equates to the 2, 20uf caps placed end to end, in the biased setup ). The 10uf was not expressly included as part of the large capacitance shunting located after the 5.4mH inductor so I wouldn't move its' location. I don't know if this truly makes much of a difference. Maybe someone with some Network SIM software might be able to lend a hand . ??
- Were you given explicit directions to combine these two areas of capacitive reactance into one ?

- I have also included the recommended polarity conventions for this N3145 network ( at least for the woofer & midrange driver ). The horn and the tweeter sections use the same polarity as the 2121 midrange. I suggest you stick to this official convention.

- I've included ( again ) Giskards' N3145 network workup since it was buried in the early pages of this thread .( As far as I'm concerned ) "it" is the master template for a 4-way "DC-biased" network using these components . You should print it out, review it, and then figure out "where & why" you have deviated from it ( in a few spots ) . You should do this before firing up your network ( or burying it into sand ) or measuring it's performance or anything else I can think of at the moment .

:cheers:

- Oh , once again, here is Giskards' parts' list ( specifically for Solen parts ).

[B&KMan]

Hi Jean

- Maybe your substitution of the 20 ohm & 15 ohm resistors for the (2) 39 ohm resistors was just another "cut & paste error" ( divorced from what you actually bought ). I don't know, you'll have to tell us what you are actually wiring up .

IT IS MY MISTAKE COPY PASTE (...) :banghead:

- I've also divided out the (2) biased 20 uf caps in the woofer portion of the network . The original N3145 has the original 10uf after the Zoebel (10 uf equates to the 2, 20uf caps placed end to end, in the biased setup ). The 10uf was not expressly included as part of the large capacitance shunting located after the 5.4mH inductor so I wouldn't move its' location. I don't know if this truly makes much of a difference. Maybe someone with some Network SIM software might be able to lend a hand . ??

I assuming the zobel correction is most close as possible in driver for best match inpedence, I keep in my hand the , futile maybe, position is inverse so i decide to put 20Uf before... but I recognize I play in apractice-sorcier

- Were you given explicit directions to combine these two areas of capacitive reactance into one ?

- I have also included the recommended polarity conventions for this N3145 network ( at least for the woofer & midrange driver ). The horn and the tweeter sections use the same polarity as the 2121 midrange. I suggest you stick to this official convention.

you absolutely right I just exposed the electrical polarity not post speaker... I realize the confusion here

- You should print it out, review it, and then figure out "where & why" you have deviated from it ( in a few spots ) . You should do this before firing up your network ( or burying it into sand ) or measuring it's performance or anything else I can think of at the moment .

I have a copy and check check and re-check

in the works 100 comeback.... (...)

MY DEEPER THANKS FOR FOUND THIS ERRATUM... (...)

:cheers:



:cheers:

[Ian Mackenzie]

Well it looks like this thread died:blink: .

Jean, I am putting a basic new equivalent network together soon .

Have you completed your charge coupled version yet?

The Doctor.

[Mr. Widget]

Well it looks like this thread died:blink: .

All good things must come to an end...:uhmmmm:

Widget

[Zilch]

Jean is testing, no doubt.

The process will require considerable knob twiddling and head scratching.

:D

[Ian Mackenzie]

Yes its absolutely annoying when nothing goes according to plan.

No doubt he will hear from Jean soon enough.

In the mean time I put this together...I do believe it works but no doubt someone around here will disagree...They always do.

I played Floyd " The Wall Live in Berlin" last night and it was quite satisfactory.

Now for the finishing touches. Some CNC routing for a nice front face plate perhaps!

The Doctor

[Ian Mackenzie]

The business of making audiophile passive crossovers is interesting and fun.

I am not suggesting the above is best practice but it is a means to an end and was designed around some very specific requirements to hot rod a 4343.

Aside from the whole ethereal and romantic aspect of soa parts there's the mechanical side of the job. That's the hard part. The layout was a logical expression of the schematic. The whole thing must be mechanically stable and endure delivery and a lifetime of use.

It must also be a turnkey solution for the user and so special attention has been given to interfacing terminations and installation instructions.

I made the terminals from pure copper sheet on nylon stand offs, silver solder, some teflon 16 gauge wiew and your on your way.

The material is 3mm plate aluminum and 1 inch square corner posts milled in the lathe. The whole thing is tapped and bolted with M4 screws.

Simple in concept but difficult to execute well.

It reminds me of a Valve radio, deceptively simple but rather a black hole for time.:eek: . Never mind its all fun. That's why you would never see this as a commercial product.

The Doctor

[B&KMan]

Well it looks like this thread died:blink: .


Hey hey Great members...


Sorry for long step silent time...

--- Actually In montreal is meteo is really pain, my computer is tilt a couple of time this week because the tempeture is arround 100 F or 32 C plus 95 degree humidex (so the relative temperature is arround 40 C !!!!...)

--- The ends of school and vacancy of kids and small vacancy of my wife is put out myself of my little construction of bigNet...

--- Professionnally I'm very busy, I run big set-up calibration of my B&K of national reaseach lab and couple of work job of insulation architectural design: investigation layout performance, so my instrumentations is request for the job (sic),
in couple days or week the situation is back to normal... and the passion is come to married of JBL 4343 and Bignet...


I keeped my small time to correction of the error in my electronic set-up in regards of faucon eyes of Earl ... (thanks again...)


Jean.

[B&KMan]

Jean is testing, no doubt.
The process will require considerable knob twiddling and head scratching.
:D

Yes my dear...

many small details keep big time for good integration andd looks sharp for eyes...

I work wood for round backpanel...

-I works gabarit for internal brace support, and look design for plug on top and botton for fill in and out the sand of network...

and because I have no experience on that the time for integrate all aspect is long... sorry for time delay suspense (yuk yuk yuk ) :D

exemple

---I hesitate to put the back panel in entire black, flat finist paint, or put a wood (same as speaker...)

--- I hesitate to plug look finish ( i hate if is look to plumber plug )

--- because the price of good L-pad I buy a cheap for test impedance and electrical correlation voltage spl ( maybe some poeple is feel what I'm talk here :D muhahahahahaha)

so dont desasperate I garantee you receive the final word of this adventure in world of passive network....


:cheers:



Jean.

[B&KMan]

Now for the finishing touches. Some CNC routing for a nice front face plate perhaps!


Hi and WOW ! metal bench !!! is not a joke :applaud:

Beautiful set-up ...


your network is look same a individual component electronic...

do you put in floor ?? on speaker ???

And what is intriguishing connector ??? speackon on neutrik ???

:cheers:


Jean.

[Ian Mackenzie]

Glad to see you are around Jean.

The external crossover will sit behind the speakers.

I will show more pics when complete. Just some pics here as a primer for a project.

NEWS FLASH

At the moment the recipient is doing some very interesting listening tests in full passive mode with original passive 4343 crossover without active crossover.

"To present an analogy, with the bi-amp setup it sounded like there was
someone or something in between me and the speaker. I could still hear the
speakers fine. But now, powered only with the Pass and going all passive
it's as if that obstruction has moved out of the way."

Less is more. At least with a Passlabs X250.

Further details of this evaluation will be discussed in a project. I must get Nelson Pass to view this thread!

This arrangement may seem a contradiction to convention for those biamping, particularly those convinced bi amping and bypass of the 52uf capacitor sounds better.

The introduction of the extenal passive crossover will be very interesting.

The Doctor

[B&KMan]

"To present an analogy, with the bi-amp setup it sounded like there was
someone or something in between me and the speaker. I could still hear the
speakers fine. But now, powered only with the Pass and going all passive
it's as if that obstruction has moved out of the way."

Less is more. At least with a Passlabs X250.

This arrangement may seem a contradiction to convention for those biamping, particularly those convinced bi amping and bypass of the 52uf capacitor sounds better.
[/i]


HOHOHOHOH

:applaud:

This is a big dip point of view and you start or re-start a old war... My brother is look multi-amp is better... For me is real only if you put same high-end quality and you cancel completely the passive network...

so accuphase (look same serious of passLab) electronics network for 4 driver + interconnect + 4 amp + regulation, + control vibration furnish = more more more money than high-end passive crossover !!!

So I put my head on the wood too and I assume entirely this point of view...

:cheers: , :cheers: and :cheers: again...


Jean :D

[Ian Mackenzie]

Jean,

I am not interested in a long winded debate on bi amping.

I merely quoted the impresssions of a listener who removed his active crossover.

I found the subjective remarks so profoundly accurate from my own experience with active crossovers I posted them here....there is perhaps one exception.:rotfl:

Audio reproduction by the way is not an exact science and as such each listener is entitled to his own opinion.



The Doctor

[Ian Mackenzie]

HOHOHOHOH

:applaud:

For me is real only if you put same high-end quality and you cancel completely the passive network...

so accuphase (look same serious of passLab) electronics network for 4 driver + interconnect + 4 amp + regulation, + control vibration furnish = more more more money than high-end passive crossover !!!

So I put my head on the wood too and I assume entirely this point of view...

:cheers: , :cheers: and :cheers: again...


Jean :D


Jean don't jump to conclusions,

Everyone has their own point of view and The Lansing Heritage Forums respects that.

When I was a young student of audio nirvana I would always look up to and respect Karen, the beautiful Passlabs Diva and the Master Nelson Pass.....and I continue to do so.

There are some things which you just can't put a number too but which guide you to perfection.

The Doctor

[B&KMan]

Everyone has their own point of view and The Lansing Heritage Forums respects that.

There are some things which you just can't put a number too but which guide you to perfection.


:blink:

I'm not insult or irrespect others members !
I just put my point of view.

(manys others expose a terrible scepticism of cable, high tech amp, etc..)


I expose the subjective remarks so profoundly accurate from my own experience with active-passive crossovers and I posted them here..


:D :D :D

Jean

[Ian Mackenzie]

:blink:

I'm not insult or irrespect others members !
I just put my point of view.

(manys others expose a terrible scepticism of cable, high tech amp, etc..)


I expose the subjective remarks so profoundly accurate from my own experience with active-passive crossovers and I posted them here..


:D :D :D

Jean



Jean,

Most sceptics are ignominious bastards, they don't ask the right questions so they never get the right answers!

The Doctor

[B&KMan]

Most sceptics are gnominious bastards, they don't ask the right questions so they never get the right answers!
The Doctor

Wow !!! I listen the voice of moderator in you...

:applaud:

Jean.

[B&KMan]

OK

(after this interlude philosophy...)) I'm back with new pict of construction...

The problem of postSpeakers is really complex, the connectors is not accomode this depp of wood and the pine is too soft for build hards fixation ( the bolt is loose by natural compression of wood (...) )

So I put a very HDF plate and work a couple of hours for create a 2 layouts swandwich fixation... My first problem is the connector never loose in the time because the inside of network is filled on sand... so imagine the party for the back nut of post is loose... :banghead:

So, I know the WBT is not the top on the top connector but I have...

:D

next news soon...

:cheers:


Jean

[Ian Mackenzie]

Noa's Ark anyone?

The Doctor

[B&KMan]

Noa's Ark anyone?

The Doctor

I'm quite shaking by the long silence of other member... :blink:

I put a vibration analyse of caps in regards on the nature of panel fixation and no nothing comment... :(

Well, in other hand, The look up thread number is up regulary so I suppose many silence is not equal to dead thread... :D

so, for this silent forum (at this time) I put my last evolution of BigNet...

I hope this is maybe broken silent (yuk yuk yuk )

see pict
---start 4343
--- without grill
--- open strip
--- and prelude of test (the set-up mesurement...)

The L-pad is cheap but for test is OK...

I not decide if I reclycling original L-Pad (dispendious ) or put resistance for equivalent...

I try this set-up

and if is good , I probably buy a original metal oxid resistance (high grade) exposed in listing of Giskard...

:cheers:


Jean.

p.s. the result is coming soon (Ah !! the suspense... :applaud: )

[johnaec]

so, for this silent forum (at this time) I put my last evolution of BigNet...

I hope this is maybe broken silent (yuk yuk yuk )Jean - I think the main reason there aren't a lot of members commenting in this thread is because what you're doing is so far advanced that most of us can only watch. Very few people here, (me included), have the comprehension to really understand all the nuances of what you're doing. I'm totally impressed, as I'm sure most of us are, and can only wait to see what your final impression is. Hopefully, some other members here will get the chance to listen to tham and add their opinions also.

John

[Zilch]

O.K., I like the tripod microphone dealie.... :p

[Ian Mackenzie]

Jean,

I saw your results and will try to interpret..


I have a good reference on such and but it refers to loudspeakers per say:

Basically it relates to the panel resonant modes due to torsion and panel flexure, their frequency and magnitude dependent on the Q of the material and also its thickness and density. With an increase panel thickness and density resonant modes are less likely to be excited.

Densities vary, lead being 11.3 x 10 3 kg/m 3, chipboard 0.81 x 10 3 kg/ m 3 and plywood depending on grade 0.67 x 10 3 kg/m 3..

Enclosure Resonance is extremely complex but essentially a clamped panel will have its own acoustic output when forming part of an energised loudspeaker, derived from sound within an enclosure.. The output consists of standing wave modes at higher frequencies and pressure modes at those frequencies where the wavelengths exceed the internal enclosure dimensions.

In theory, a clamped panel has a well defined vibrational series in both longtitudinal and bending modes. A further mode is due to panel mass resonating with its own and the enclosure's air volume stiffness.

Many have attempted to control panel resonaces with limited success. The results dissplace such resonances to more subjectively acceptable frequencies but not their magnitude. Sand can be effective but is awkward to use and mass soars dramatically. When damping is used to control resonance, the Q of the panel is reduced.

But where high mass was an advantage, it is detrimental with panel damping.

Resonance contol uses the principle of dissipating energy through friction and a given panel should be laminated with a layer of comparable mass. The mass equivalance ensuring that a good mechanical match is achieved b/n the layers to permit the effective transfer of vibrational energy from the panel to the damping material.

I hope this makes some sense.

I would also point out that an often overlooked fact is that adding fibreglass to cavity actively absorbs vibrational energy through frictional losses and therefore assists in reducing enclosure related energy transmission..

The Doctor

[gerard]

Jean - b&kman


I saw there is fiberglass all over the 4343 ; Do you think it is better .
My clone only have in the low part of the system ? .

Gerard

[Zilch]

I put a vibration analyse of caps in regards on the nature of panel fixation and no nothing comment... :( You were surprised (and somewhat dismayed) to find that solidly fixing the components to the substrate serves to directly transmit the substrate resonances (which are within the audible range) to those components.

Now you must damp the substrate. Am I gettin' this? ;)

[Ian Mackenzie]

You were surprised (and somewhat dismayed) to find that solidly fixing the components to the substrate serves to directly transmit the substrate resonances (which are within the audible range) to them.

Now you must damp the substrate. Am I gettin' this?

Medium density foam rubber and secured with cable ties seems to work.

I put everything in an enclosure, A) to survive the delivery B) to keep the noise out of the works!

The Doctor

[Robh3606]

You can use a buffer material. I use 3/4 inch fiberglass ceiling tile between the actual crossover substate and the speaker side panels. You can keep the wire leads off the panels or use a staking compound that absorbs vibration like a pliable RTV.

Rob:)

[B&KMan]

Jean - b&kman


I saw there is fiberglass all over the 4343 ; Do you think it is better .
My clone only have in the low part of the system ? .

Gerard


Hi dont know if you remark the x cross wood under fiberglass for support the HF and all fiberglass is flue in frame plus 3 sheet on top for hf and uhf section... Yes i is strongly recommended because it is destroyed the standing wave prisonner on top section... Remind the 4343 the event is just very low botton and med close box create a cage resonnator so for more control ( conform to engeneer I recommand to put in )

:cheers:

Jean.

[B&KMan]

You were surprised (and somewhat dismayed) to find that solidly fixing the components to the substrate serves to directly transmit the substrate resonances (which are within the audible range) to those components.

Now you must damp the substrate. Am I gettin' this? ;)

Yes and no.. (english transaltor maybe)

Shit back, keep big drink, chip, and play maybe a printer version and read slow this fundamental study of vibration oriented in acoustic :p

(i hope is not to hermeneutical exposition and my english support the very tuff synthese of what is what... :D )

(I dedicace this post in all members who help me strongly in my poor conception in electronic...

:cheers:


=====================================


All material have a natural vibration (right?)

more soft or massif = more down amplitude... (right?)

The thin massonite is light and is same material of thick massonite)
So I realise in this test the vibration is normally down with increase the mass

so the thick is normally create less vibration and normally absorb by mass the vibration of caps...

Bad surprise the thick massonite is vibrate same as thick.. but, because it is more mass, the excitation contain more energy and create a more resonnace transfert on caps than only natural resonnace caps only... so the total vibration is increase !!! in this case the thick massonite is create resonnace on caps than keep resonnace on caps... (double bad way)


keep same in other hand, I hope it is more clear in this..

the test modal expose the thin is not realy good but better than thick !!! because if the thin is not keep the energy on caps by mass but it is not create a double hit resonnace energy (create by thick.):the thick create double peak resonnace and appear in listener more flat but in reality is induce more distortion in caps and finnaly in signal so the total noise is higher and if not keep engergy vibration is give on caps extra energy...

interpretation of ear listening:

remember the first criterion of discerne coloration is the power of peak and restriction in spectrum... In fact if you have peak in entire spectrum, you not keep signature but just not discern the detail hidden by this noise floor.

Ex: is you play pink noise and induce a sinewave you ear is detect the coloration immediately but if I put narrow noise in pink noise, you ear keep difference but not discern what is what, etc, etc, etc,

I try in other exemple, if you put a diapason into body of guitar you listening clearly the note but more powerful...= natural amplifing...

So the thick and thin massonnite have natural frequency and at this resonnace is play same a diapason on caps... so less power is better and if you material have multi-discrete resonnace than a major single peak, you hit 2 function:

1--- you not produce resonnace energy in caps and
2--- you have a great opportunity to keep the energy of caps for restriction of natural frecquency of caps= less resonnace caps noise....


In theorycally the next point is realy specific on acoustic problematic...

Normaly, the engeneer is work to create a damping structure for limiting the amplitude of natural frequency... in case of audio, this damping is only valuable by approach of hysterical or structurally damped...

Why ?? reason: TIME DISSIPATION ENERGY !!! (see pict)

The engeneer, on standart structure, not consider if is create a secondary peak or more bkgnd noise ... For audio the time dissipation is critical... more than .2 ms the signal bkgnd noise is push up and blur details, space ambiance, etc... so if you put asphalt, rubber, neopren or any other absorber, the peak is converted in time dissipation so you reduce the peak amplitude but increase bkgnd noise floor...

so the best in the best is create a sandwitch with multi material : all material is hard and lower than .2m delay dissipation peak but each have big difference natural frequency... so each resonnace is cancelled by the other with out increase time dissipation (right ??)

Add at this principe the concept of mass for inertia and you result the best : big mass with composite of hard material for cancelling peak resonnace... and Voilà !!!... apply this concept of box enclosure, board electronic or electronic unit, stand speaker, or stand electronic and you produce the best of the best inertia as possible...

If you control each aspect of the vibration insertion on your sytem you control a big aspect of problem coloration. but if it is not sufficient for garantee the result: it is necessary for good way of flat and ultra realistic reproduction...

Ex: Mr hiraga, (world sommit of horn specialist) consider a enclosure of sand arround horn is garantee this horn is not generate a parasite natural resonnace in sound because the freq of sand is too high and the capacity of absorb fast with not extension in time dissipation...



Finnaly, for more specialists in read this, last aspect is not cover by this small experience is factor of air born sound transmition...

Same as romm, insulation is 2 magor factors: impact energy and air sound...

this is second aspect explain what it is important to create hard connection of caps to material... the air sound produce exictation on caps and result in resonnace free without never material to reduce this resonnace... So if you put caps in the air the energy of caps is not go a way to go out of caps and energy is just procure more resonnace...

but if you put damping material, you affect this rule in amplitude to distribute enrgy in time so you put up the bkgnd noise... plus you create a seond signature of the damping material into caps...

So it is strongly important to fix caps hard to higher mass as possible and this mass is more fast dissipation energy as possible...


I understand many guy Is create apparently good result with out this rules but if you keep sofisticated material analyser, you realize many fantasist apporch...

I hope this rules is help anybody sam this few member is help my in electronics... I knoe the material for analysing this is (extremely prohibitif).


I conclude by exemple in quest in my young age... I have a big rock sstand slab for support my electronic andunfortunatly the sound result is realy harch.... So I try to tune the slap with damping eleastomere... smal peice, med peice, full, lead, woood, etc, etc, etc, finnaly couple years later , after my lab, I check my slab tuned... The result is simply catastrophic, after applied the rules exposed here , the result is just incredible... but the moral of this story is my damping material is tune a peak for 2 db but the real power peak is 32 db !! so imagine the futile correction is here....

Jean

Weel, time for beer yuk yuk yuk !!!

:cheers:

Jean.

[B&KMan]

Here pict who I explain as possible the phenomenon of damping problem in time

Jean

[B&KMan]

You can use a buffer material. I use 3/4 inch fiberglass ceiling tile between the actual crossover substate and the speaker side panels. You can keep the wire leads off the panels or use a staking compound that absorbs vibration like a pliable RTV.

Rob:)

Euh sorry what is pliable RTV ?? :blink:

Jean

[B&KMan]

Hopefully, some other members here will get the chance to listen to tham and add their opinions also.

John


Well, thanks for gentlemen appreciation, unfortunalety, I leave in quebec canada, so the listening is tuff at this distance, I try to record with my B&K result with part tune but my old computer is too fast saturated...

So I compansate by pict and result analyser...

thanks again for good words...

:cheers:


Jean.

[B&KMan]

OK dokay, fresh new for tis project...

I send a couple of pict to speakers and the bignet install on back, most details of insides speakers for cloner guys... :D

:cheers:

[B&KMan]

Installation on bignet....on back of 4343

:cheers:

[B&KMan]

Finally replug speaker and most important, place exactly at same place and angle position...


and result of sound speaker left in grey bar (3143) and (in highlight line) the result of right with bigNet...

--- First constatation: my uhf is not good... (it is not my first modification caps it is a cone replacement :banghead: )

--- second constatation the response of each driver is really differ to original response...

:blink: :blink: :blink:

I feeling wrong problem network...

[B&KMan]

Why I have hole in mesure...

I mesure each driver individually out of box with network... I put pink noise and mesure with mic at 2 inch of outpput for keep clean response driver...

see pict for detail....

[B&KMan]

Ok now I compare each driver to next other....

[B&KMan]

same mesure comparison but in 1/3 octave slope...

the last pict is the difference level of peak and cut freq of hf.
close to 9 db down (...)

:blink:


it is my driver of I feeling is to much drop... ???

:dont-know

:help:

thanks for all is decorticate this info for confirm or unconfirm the response data coherence in regards of design network 3145....


Jean.

[Ian Mackenzie]

Jean,

Have you checked to see how much vibrations from the woofer make transmission through the rear baffle at various bass frequencies?

It just seems to me that the resonances and vibrations from the speaker cabinet would far exceed that from within the capacitors and this would be worth testing and I am sure a some simple test with Floyd Dark Side of the Moon would tell you this..LOL:D

For this reason would it not be better to completely de couple the crossover from the speaker cabinet?

Okay removing the crossover components from the interior of the bass enclosure is acknowledged as better for "audio" but if you are critical of vibrations why attach the crossover panel to the speaker box and rear panel the largest panel most susceptable flex and to vibrations?:hmm: :dont-know .

Sorry if am am a pain in the arse.!

The Doctor

Ps Be careful of the fibreglass, it can make you very itchy in all the wrong places..Muhhahahah

[B&KMan]

Jean,

Have you checked to see how much vibrations from the woofer make transmission through the rear baffle at various bass frequencies?

It just seems to me that the resonances and vibrations from the speaker cabinet would far exceed that from within the capacitors and this would be worth testing and I am sure a some simple test with Floyd Dark Side of the Moon would tell you this..LOL:D

For this reason would it not be better to completely de couple the crossover from the speaker cabinet?

Okay removing the crossover components from the interior of the bass enclosure is acknowledged as better for "audio" but if you are critical of vibrations why attach the crossover panel to the speaker box and rear panel the largest panel most susceptable flex and to vibrations?:hmm: :dont-know .

Sorry if am am a pain in the arse.!

The Doctor

Ps Be careful of the fibreglass, it can make you very itchy in all the wrong places..Muhhahahah



:blink: excelent remarks...


Ok Why I put network on back Panel ???


Yes it is the major panel vibration...

I have option to put into indepedant box but I try to fix multi-aspects

1--- load mass back speaker increase resolution of speaker ...

2--- Remember all in fill in sand so really dead for caps and really flat result on back panel...

3--- The form of back panel is a " natural " diffusor so is broken the standing wave load on back side speaker...

the mass of network adding of mass on speaker is more better than 2 mass separate...

Well I feel all this point is necessity of proof Ok I try a build experience for you...

(I hope you help me better in regards of the curve of my hf -uhf driver.... )

Jean.

[B&KMan]

Ps Be careful of the fibreglass, it can make you very itchy in all the wrong places..Muhhahahah

I keep pict fiberglass position for cloner ....

:D

Jean.

[B&KMan]

Well I feel all this point is necessity of proof Ok I try a build experience for you...
Jean.

I play dark side in sacd mmmm really good but I have better procedure for kill anybody doubth...

OK I perform another dispendious test for you great member... :D

the test protocol:

1--- I put mic in front of the low driver 2 inch precisely
2--- I put same power and same impluse at all mesure.
3--- I put sensor vibration on 3 positions:
A--- on the bignet rock plate (the big Net is fixed on back-panel 4343)
B--- On back panel on unaltered 4343
C--- On floor 2 inch back and half distance of each feet 4343.

The first pict is expose the set-up

Syncro signal with mic and sensor vibration...

The special function is check the pulse input to mic and compare of acc for compensation of variation of signal excitation and put algorith for expose real power excitation in regards of variation "shaking"

Finnaly look carrefully the picts for difference in DB or linear scale...
N"B" each notch scale on botton =200K (so 0,200,400,600,etc.) 1600 max Hz evaluate...

[B&KMan]

I put sensor vibration on 3 positions:
A--- on the bignet rock plate (the big Net is fixed on back-panel 4343)
B--- On back panel on unaltered 4343
C--- On floor 2 inch back and half distance of each feet 4343.

ok this 2 pict is compare level vibration in regards of spectrum 0-1600Hz

THe stored is the mesure of BigNet and input is the mesure on standart back panel 4343

same pulse power, same position mic reference, etc...

First look the difference of DB level vibration... more 30DB !!

in mode linear scale the difference is look more the Db scale is logarith :D

OK do doubth the bigNet composite and addition mass is create inertia and flateness response.. but this is not proof is better than put network on the floor...

[B&KMan]

So what is the noise vibration on floor ???

Attention this mesure is keep the result response of 2 ways

the mesure is the sum of direct impact vibration and acoustical exication...
So it is possible to produre big reactance on floor and put the speaker on supension with no contact of the floor...


ok this 2 pict is compare level vibration in regards of spectrum 0-1600Hz

THe stored is the mesure of BigNet and input is the mesure on floor locate at 1 feet back speaker...

same pulse power, same position mic reference, etc...

First look the difference of DB level vibration... less than plain back put realy higher than bignet...

in mode linear scale the difference is look more the Db scale is logarith :D

OK do doubth the bigNet composite and addition mass is create inertia and flateness response.. this inertia of this mass added to back panel help the back to drop level vibration... so, better reaction of cabinet and less general vibration... for power air sound the addition of mass create a better inertia vibration than is you expose a device free placement on back speaker or n top or arround speaker...

Now add the 80LBS sand in network and imagine the dead vibration result...

the mesures is surprising ???

I hope you appreciate thi 2K test !! :D

[B&KMan]

HUMMMMM

:blink:

Another long silent here...

but the visit is increase regulary...


Well well , I back just for write my surprising on Bignet, the time rodage appear maybe small or the dc charge is more stabilising. the unpleasant momment with sound go up and down, stiff and bloom, is really small...

after 48 hours the sound is quite stabilising... if the normal effect is reproduce the final result was very impressive because actually the speaker modified play same detail but little bit more stiff than old speaker not modified...


scotie entreprise end report !!! :p

Jean.

[Ian Mackenzie]

Jean,

Good to hear your project is progressing.

I here is an update of the cabinet for the 4343 external crossover.

As seen by the A4 laser copy they are about the size of a PC Mini tower.

I should have details of final assembly in a day or two.

Oh I forgot the other box, this one is for the active crossover.

It was the first fabrication of the style, I am starting to get better at it.

I might buy a CNC router next year, that would be nice.

Ian

[B&KMan]

Jean,

Good to hear your project is progressing.

I here is an update of the cabinet for the 4343 external crossover.

As seen by the A4 laser copy they are about the size of a PC Mini tower.

I should have details of final assembly in a day or two.

Ian



MMMMMM It look very professionnal....

:cheers:


For suite of ear report

The time rodage is continuing... and maybe modification...

I dont know if is the soldering silver or is the caps film and foil is suck or not ringning, but my high UHF is play more low than my first modification. After a test I realise a good power if i put out the 2 resistances of end circuit uhf nework...

Ian, (and other) do you a experienced this effect with cascade set-up ???

[Ian Mackenzie]

I am not quite sure what you are saying?

My interpretation of your question is that the parrellel effect of smaller and alternate dieletics will result in a different ESR and other electrical properties of the capacitors. It will therefore in all probability sound subjectively different and that's the whole point of the exercise.

Jean, there is still a lot that measurements don't record in terms of subjective impressions of sound quality and it is even more difficult to describe adequately in words what we often hear.

My own inclination is that superior capacitors tend to sound less like a cheese grater on the inner ear and may therefore sound lower level until you become more acquainted with the change. This is despite the fact that you are able to hear and differential more specific details of sounds from the louspeaker with improvements in audio fidelity.

Enjoy the rediscovery of your record collection.


Ian

[Robh3606]

"I dont know if is the soldering silver or is the caps film and foil is suck or not ringning, but my high UHF is play more low than my first modification. After a test I realise a good power if i put out the 2 resistances of end circuit uhf nework..."

Do you know if the driver levels have changed at all. Any baseline measurements before?? Could be the caps could be a 1-2dB driver level change as well??? I found that with the charge coupling on my 2344 horns that the levels were right where they were as best as I could measure so the difference had to be the caps. What I found was they sounded smoother and were not as harsh you I think you could hear that as a drop in level especially at higher listenning levels. The effects were subtle but there.

[rek50]

"Silver Solder Connections"... It must be my "Granite (Hard) brain that doesn't allow me to understand how a SOUND mechanical joint can/would be influenced by the top gas/air barrier (solder). To me, the "Sound Mechanical Joint" IS the connection (the cake), the solder is the "Frosting" on the cake. A helical twisted connection (with wire nuts...ala Henry Kloss) beats a poor weld/solder job, because the solder is the connection, rather than the "Sound Mechanical Joint". I've read about cascading caps and it is interesting but all those caps, instead of one or two for the needed value, might just cause its own problems; a lack of cohesion and focus.... Is it Live, or is it Memorex? At any rate, Nice Work guys! Any Unchallenged man remains a King of nothing. Trying/Working reveals who we really are....:)

[B&KMan]

The problem of the UHF is not a problem of subjective perception... If you look at the photograph, the UHF is definitely too low and although L-pad is to the
maximum....

The same phenomenon occurred at the time of my modification of the
3143 original... The wire were not twister etc,

I notice that by doubling the condensers the attenuation is stronger
but that frequency response is excellent.... Thus which is the problem? (#$#@$#$)

Jean.

Solen is explainthe higher power handling the less sensitivity caps...

My hypothesis is the place where the power is really low is UHF so maybe is pass the line of minimum ...

What your feeling ???

[B&KMan]

Do you know if the driver levels have changed at all. Any baseline measurements before?? Could be the caps could be a 1-2dB driver level change as well??? I found that with the charge coupling on my 2344 horns that the levels were right where they were as best as I could measure so the difference had to be the caps. What I found was they sounded smoother and were not as harsh you I think you could hear that as a drop in level especially at higher listenning levels. The effects were subtle but there.

Thanks Robh3606,

Not, I is not made of exaustive measurement on the power before the
condensers... I am obviously conscious that the distortion is a factor which
accentuates the UHF... In this precise case, one speaks about a lack of more than 10 decibels... It is more than one subtlety or of the distortion... I obtain has little close the good result by removing the 2
6.5 ohms to the end of the circuit of UHF...

Other hand if you look my result on prev page, The the crossNetwork on HF and UHF is really strong and deep... and create little hole.

I have in my mind this circuit is not for 2420 diaphram but for UHF it is a good layout...

strange.... other idea ???

Jean

[B&KMan]

"Silver Solder Connections"... It must be my "Granite (Hard) brain that doesn't allow me to understand how a SOUND mechanical joint can/would be influenced by the top gas/air barrier (solder). To me, the "Sound Mechanical Joint" IS the connection (the cake), the solder is the "Frosting" on the cake. A helical twisted connection (with wire nuts...ala Henry Kloss) beats a poor weld/solder job, because the solder is the connection, rather than the "Sound Mechanical Joint". I've read about cascading caps and it is interesting but all those caps, instead of one or two for the needed value, might just cause its own problems; a lack of cohesion and focus.... Is it Live, or is it Memorex? At any rate, Nice Work guys! Any Unchallenged man remains a King of nothing. Trying/Working reveals who we really are....:)

Thanks rek50,

But I tested conductivity of connection without soldering and the caps is completely fixed on board... Helicoidal twised connection or just put lead in hole of pcb and soldered, my feeling is the lead on pcb is more fixed by soldering no ??? my reserve is for the quantity of metal create a resistance but not bad electrical connection... but I'm not electronic man... :D

the cascade maybe broken cohesion or focus but it is possible to keep 10 Db plus 4 db (full max L-Pad) difference !!! hey is not a details here : it is major problem...


=================

but a fundamental question here:

16 smalls soldering it is better than 3 big soldering ???

=============

any specialise have a response ???
Jean.

[rek50]

Jean, I'm not an electrical master either. I do see the logic of a single, big (uF) cap being slower than the same value divided by two caps. To that end, I haven't "Cascaded" so to speak, but rather chose values that compliment all the values in the network. Say if I only needed 22 uF for the LF, 8 uF for MF, 3.3 uF for HF, and 2.2 uF for UHF. For the 22 I used (2) 10 and (2) 1, for 8, (1) 5 and (1) 3, for 3.3 (1) 3 and (1) .33., for 2.2 (1) 2 and (1) .22. My attempt is to keep the response close to equal, considering the duty of each section. I'll measure each cap and write down the measurement. Then I assemble (Sound Mechanical) the caps and measure the total. If the measurement equals the sum of my writen measurements, I'm ready to solder. After I solder, I measure again. The twisted together wire idea was to point out the need for a sound mechanical joint, as I don't view solder as a "Glue". :D

[Robh3606]

If you are getting that magnitude of difference something is obviously wrong. Are you sure nothings misswired and the values are correct?? Are the pots on the right legs?? Double check the phase on the drivers?? I am assuming the white plot is the new?? Why the roll off from 2K?? You should be flat out to about 8K from the 2420/2307?? Are you picking up the HF from behind the 2121 low pass filter???

Rob:)

[B&KMan]

If you are getting that magnitude of difference something is obviously wrong. Are you sure nothings misswired and the values are correct?? Are the pots on the right legs?? Double check the phase on the drivers?? I am assuming the white plot is the new?? Why the roll off from 2K?? You should be flat out to about 8K from the 2420/2307?? Are you picking up the HF from behind the 2121 low pass filter???

Rob:)

the thin (very white) is new result and grey is old result ...

yiah the HF is drop fast and the uhf is poor power... but I verified at close mic close to 2 inch impulse set-up (see : http://www.audioheritage.org/vbulletin/showpost.php?p=61791&postcount=343 ) the result of each driver appear ok but I dont have a classical response of 3145 of 4345 driver for verifiyng correct correspondance...

Thanks I run soon test of phase, maybe this is surprise is here...)

Jean.

[B&KMan]

Jean, I'm not an electrical master either. I do see the logic of a single, big (uF) cap being slower than the same value divided by two caps. To that end, I haven't "Cascaded" so to speak, but rather chose values that compliment all the values in the network. Say if I only needed 22 uF for the LF, 8 uF for MF, 3.3 uF for HF, and 2.2 uF for UHF. For the 22 I used (2) 10 and (2) 1, for 8, (1) 5 and (1) 3, for 3.3 (1) 3 and (1) .33., for 2.2 (1) 2 and (1) .22. My attempt is to keep the response close to equal, considering the duty of each section. I'll measure each cap and write down the measurement. Then I assemble (Sound Mechanical) the caps and measure the total. If the measurement equals the sum of my writen measurements, I'm ready to solder. After I solder, I measure again. The twisted together wire idea was to point out the need for a sound mechanical joint, as I don't view solder as a "Glue". :D


Thanks for details...

Do you have picts ???

[Mr. Widget]

Thanks I run soon test of phase, maybe this is surprise is here...)


Reversed polarity won't do that. Something else is very wrong. Reversing polarity (phase) will primarily give you a notch at the crossover region.

Widget

[Ian Mackenzie]

the thin (very white) is new result and grey is old result ...

yiah the HF is drop fast and the uhf is poor power... but I verified at close mic close to 2 inch impulse set-up (see : http://www.audioheritage.org/vbulletin/showpost.php?p=61791&postcount=343 ) the result of each driver appear ok but I dont have a classical response of 3145 of 4345 driver for verifiyng correct correspondance...

Thanks I run soon test of phase, maybe this is surprise is here...)

Jean.

Jean,

Check your wiring layout against the schematic then refer to the 3145 attenuation tables and use the JBL standard test jig to confirm attentuation at the crossover points.

On the topic of Solder, (Yanks pronounce it Sodda) I used silver Solder from Welbourne Labs and Teflon wiring also from WelbourneLabs.

As a sub set of the technical excellence = subjective perfection one of my aims was to ensure an almost pure copper signal path to ensure maximum conductivity. Poor quality or corroded teminations are the worst demons of loudspeakers, not solder as such.

In attempting to acheive this goal the wiring is either Teflon 14 or 12 gauge high purity copper, bindings post premium Cardas, heavy duty 8 pole Speakon Neutrik output socket, terminals are 1 mm pure copper sheet. Provision will also be made for Mills fixed resisters once correct Pad levels are determined.

As a matter of polarity convention I used Black for negative and Red for postive.
There is much talk about solder but nothing beats best practise, make sure everything is clean, heat the joint correctly and apply solder to the joint.

The Doctor

[B&KMan]

OK I'm back again...

But not with good news...

I re-read entire thread (it is a real photo-soap :D )

I try to found what is erratum to affect cut frecquency...

first I realise I dont have a electrical response of classical 3145 for 4 driver...

Anybody have this magical electrical slope cut frecquency ????


2---


I back again on a 2 schema presented before in thread...
(see pict)

I verified the value of airinductor coil and I realize the value look maybe inversed in one schema

(if you look serial value of one, it is appear in other schema in parrallele)

It is a mistake Or really equivalent ???

Giskard, Robh3606, Ian, other great members ???

:blink:

Jean.

[Ian Mackenzie]

Jean,

The voltage drives are posted earlier in the thread.

You can use either of the schematics, they appear correct.

The Doctor

[B&KMan]

Jean,

The voltage drives are posted earlier in the thread.

The Doctor

Where # post ???

I found 3143 and comparison of hf and uhf but not complety curve in regards of driver...

:blink:

thanks...

[4313B]

first I realise I dont have a electrical response of classical 3145 for 4 driver...

Anybody have this magical electrical slope cut frecquency ????
Here.

[4313B]

It is a mistake Or really equivalent ???It looked equivalent to me back when I did it. There could be issues though.
If I remember correctly this stuff was all posted somewhere else on the forum.
I think the green curves (A & B) were the second version of the equivalent while the orange curves (C & D) were the original version of the equivalent.
E, F, G, H were stock.

[4313B]

Changing loads to 16 ohms for HF and UHF.

If I remember correctly the LE85/2420/2421 was considered a "12 ohm" device and the 077/2405 was considered a "10 ohm" device so 16 ohms should be too radical. I think Mr. Widget recently measured a bunch of the ring radiators but I don't recall seeing any impedance curves. Someone should measure their LE85/2420/2421's and post those impedance curves too.

Second graphic is with UHF set to 10 ohms and HF set to 12 ohms.

Try tweeking the HF circuit by replacing the 0.27 mH choke with something like a 0.24 mH or 0.22 mH choke and then adjust the L-Pad. If you have a spice package and measurement gear you should be able to get it spot on with a couple of tries. If you want me to help you with the spice portion then provide me with LE85, 2420, or 2421 impedance curves with the HL91 or 2307/2308 (data files, NOT graphs). If you want help with the UHF too then impedance curves of the 077/2405 will also be required, although the slop in that L-pad is probably sufficient to make any tweeking there fairly pointless. I would have thought the slop in the 8-ohm L-pad in front of the HF would have been enough. Impedance curves of the CD, horn, 20 ohm shunt resistor across the terminals and the L-pad set at 100% on, 75% on and 50% on would be nice too. I don't have alot of this ancient gear around anymore or I'd do it myself.

What diaphragms are you using again?

[Ian Mackenzie]

Giskard,

Due to time constraints I have not yet verified the attenuation on the new equivalent network but I can confirm the original equivalent circuit functions correctly under test with the JBL test jig. A graph I posted earlier confirms this with an in room response run. (my driver is an 8 ohm version however but you suggest there is enough slop in the Pads to deal with this).

Ian

[Ian Mackenzie]

Giskard,

I am incline to think Jean has a wiring or parts error looking at his analyser read out.

But the 5.6 uf capacitor in (A) appears large relative to the original eqivalent schematic and stock design.(3uf)

Jean,

Have you tested for induced currents into the slab of material under your inductors?


Ian

[4313B]

Giskard,

I am incline to think Jean has a wiring or parts error looking at his analyser read out.Easy to do.

But the 5.6 uf capacitor in (A) appears large relative to the original eqivalent schematic and stock design.(3uf)It's fine for certain loads as shown by the voltage drives. Anyone can change the values as required/desired to get the performance they desire.

[Ian Mackenzie]

To assist with some clarity here I will arrange the new equivalent schematic and run some voltage drive tests using the std JBL test jig over the next day or so.

Those curves from GT's file look almost identical to the readings on Db scale of my analogue meter...by the way.

[4313B]

Good.

I must have done something right then.

Thanks.

[B&KMan]

If I remember correctly the LE85/2420/2421 was considered a "12 ohm" device and the 077/2405 was considered a "10 ohm" device so 16 ohms should be too radical. I think Mr. Widget recently measured a bunch of the ring radiators but I don't recall seeing any impedance curves. Someone should measure their LE85/2420/2421's and post those impedance curves too.




First thanks very lot Giskard for this precious information and I'm feeling you put a finger touch on good aspect...

I talk with my brother and the 2405 is originally 12 ohms impedance and substitution by final 16 ohms... is look same story for the original 2420...

actually I have changed recently the 2405 for new original 16 ohms

and the sceal of 2420 is not here. so maybe ???


----------
I keep a moment for understand this excellent flood information...
But I feeling, in regards on your information, the best it is mesure the real impedance of each driver...



Thanks again...


===================

Ian,

first thanks for reply,


You trouble my head :blink:

maybe my slab rock is contains a iron and create a induction !!! :banghead: I no idea what test good for verified this aspect... do you IDEA ???
(of course the simple probe tester in ohms is create 0 value....)

I have a slab and inductor of second network not mounted so maybe it is possible to perform test but I dont know the protocol....


Jean.

[Mr. Widget]

There is no such thing as a 16 ohm 2405!!!! How many times.....? Here are the impedance plots again.

http://www.audioheritage.org/vbulletin/showpost.php?p=59784&postcount=10

For the record, the red plot which has the lowest impedance was an original 2405 with 16 ohms proudly printed on it's undisturbed foilcal.

Widget

[4313B]

But I feeling, in regards on your information, the best it is mesure the real impedance of each driver...SOP if one is as serious as you appear to be.

There is no such thing as a 16 ohm 2405!!!!No there isn't! Not even close! :rotfl:

For your purposes I'd go under the assumption that they are 10 ohms.That would appear to be SOP.



Anyway, it had to happen sooner or later.
Someone was bound to start testing a bunch of these legacy drivers and discover the truth. :p
At least now you all know why some folks hand pick their components and carefully match them. Westlake comes to mind...
I wonder if this will start a flood of forumites getting all their drivers measured. :applaud:


thanks very lot Giskard for this precious information and I'm feeling you put a finger touch on good aspect...You're welcome. I'm outta here. I hope these posts were of help.

.

[Ian Mackenzie]

Ian,

first thanks for reply,


You trouble my head :blink:

maybe my slab rock is contains a iron and create a induction !!! :banghead: I no idea what test good for verified this aspect... do you IDEA ???
(of course the simple probe tester in ohms is create 0 value....)

I have a slab and inductor of second network not mounted so maybe it is possible to perform test but I dont know the protocol....


Jean.

Reposition the inductors at least three inches and measure gain.

Jean,

Have you verified with a meter (with a db voltage scale) with the filters connected on the 8 ohm load resisters the attentuation as per JBL spec sheets on the 3145 crossover?

This will help you determine where the fault is.

[B&KMan]

There is no such thing as a 16 ohm 2405!!!! How many times.....? Here are the impedance plots again.

http://www.audioheritage.org/vbulletin/showpost.php?p=59784&postcount=10

For the record, the red plot which has the lowest impedance was an original 2405 with 16 ohms proudly printed on it's undisturbed foilcal.

Widget


Thanks Mr Widget,

Do you have perform correlation into impedance curve and spl ???

According in basic principe the 8 ohms is play (in theory) 2 more loud than 16 ohms... Right ?? so the 16 to 10 ohms is contain potentially big difference ouput power... right ???

Jean.

[B&KMan]

Hi everybody,

I keep just 2 minutes in rush time for give a late news...

actually audio system is open 24H of the 24H, left speakers or old network and right is new Network...

After 1 week, the sound is change... yesterday surprise the speaker right is rebird with phenix...

Yesterday the bignet is pass the first step rodage... the sound is less squeese sound, more deep and wide...

I no time now but I very curious to check mesure again...

I'll be back !!!

:cheers:

[Zilch]

According in basic principe the 8 ohms is play (in theory) 2 more loud than 16 ohms... Right ?? so the 16 to 10 ohms is contain potentially big difference ouput power... right ??? Yes, if such actual impedances existed in 2405's, which they don't, no matter what it says on the foilcal label, which is the point here....

[B&KMan]

Hi everybody,

I keep just 2 minutes in rush time for give a late news...

actually audio system is open 24H of the 24H, left speakers or old network and right is new Network...

After 1 week, the sound is change... yesterday surprise the speaker right is rebird with phenix...

Yesterday the bignet is pass the first step burning... the sound is less squeese sound, more deep and wide...

I no time now but I very curious to check mesure again...

I'll be back !!!

:cheers:


Another 2 minutes for expose sound effect result in ....step 2..

Now the right speaker with bigNet is play clearly higher than old network on left speaker.... little more wide and fast but (here is come baby) the sound is metal harsch... because I'm not familiar with the russian mountain burning in this set-up but, I'm refer to burning of modification on 2143 in first post in this thread:
The metal harsch is long phase before final great opening sound and down signature...

actually I listenning more 50% details before modifications....

actually the sound is look same as first modification with film and foil Solen really improvement but a touch of metal on all aspect... this effect is completly dissapear after 5-6 week with system on 24H/24H...

I have no idea if the Dc charge is slow or not the time burning...


So the triller continue...

:cheers:

(p.s. I try to mesure entire speaker this week-end ( big time consuming))

Jean.

[Zilch]

In all of this, I have not yet figured out what "rodage" or "cie" mean.

Anybody have a clue? ;)

[B&KMan]

In all of this, I have not yet figured out what "rodage" or "cie" mean.

Anybody have a clue? ;)


sorry poor english => rodage = burning

=> cie = company


:o:


Jean

[Ian Mackenzie]

I have a long weekend so I should be able to key the system into absolute focus with the Hi end external crossover and Hi end active crossover.

Pictures to follow.

Here is the IN/OUT department.

The Cardas Bindings posts provide input from the large Spades. One Binding Post pair provides termination of the bass bi amped signal which is routed directly to the Output via the Heavy duty 8 Pole Speakon. The other Binding post pair handles the Mid/High biamped signal which is routed via the 3 passive crossover filters. The Speakon connector provide a convenient and effective means of routing 4 conductor pairs directly to the drivers while maintaining fail safe phasing of the discrete signals.

Earlier in the week I burnt my fingers solding the Cardas binding posts, they are an excellent conductor of heat!

The Doctor

[Ian Mackenzie]

WOW.

Okay,

Sometimes there are no words to describe what you hear, this is one of those moments.

What I can say is there appears to be a certain order about what is heard.

There is also the most believable life like expanse of sound such that you are projected to the venue of the recording.

[B&KMan]

Okay,
Sometimes there are no words to describe what you hear, this is one of those moments.
What I can say is there appears to be a certain order about what is heard.
There is also the most believable life like expanse of sound such that you are projected to the venue of the recording.


Oh yes .... WOW : better fine craftman than my project :applaud:

And in fact, I prefer the simplifying components appproach, less parts, less connection, better theory approarch !!!....

few question...

1--- time burning before exclamation ???
2--- what is network before ???
3--- better details of construction ???
4--- electrical slope ???
5--- This is not a 4343 what is this model ???
6--- Wath is schema of this network, I look not a Dc-Charge...
7--- More details wiring, solder, etc ???

:cheers:


Jean

[Ian Mackenzie]

Ah Jean this is not a competition.....I mean hello

few question...

1--- time burning before exclamation ??? 1 minute...:rotfl:
2--- what is network before ??? 3145 0riginal eqiivalent
3--- better details of construction ??? Are you kidding:eek:
4--- electrical slope ??? refer to (2)
5--- This is not a 4343 what is this model ??? Hello:snore: , refer to build 4345 thread
6--- Wath is schema of this network, I look not a Dc-Charge... Obviously
7--- More details wiring, solder, etc ??? A few pages back I post details:snore:
http://www.audioheritage.org/vbulletin/showpost.php?p=62157&postcount=366

Would you like me to try it full of sand? I have plenty at the local beach 300 metres away:rotfl:

[boputnam]

Sometimes there are no words to describe what you (see), this is one of those moments. Dammit! I have been lame on sending the JBL Blue swatch! Sheeeeeit!

Sorry dood - will get one posted, today!

Can you post any measurements of the response, before and after, or is it not something that will impress much in the measuring...?

[B&KMan]

Ah Jean this is not a competition.....I mean hello

Would you like me to try it full of sand? I have plenty at the local beach 300 metres away:rotfl:


:blink:
Sorry If I ask questions who responses is already posts... :o:

But this thread is realy long and it is better easy to resume little bit at few time for better easy compehension and appreciation works...

Well for your first comments... I try to understand where my posts suggest this attitute ??? My intention is just relevant the finest of craftwork: that is that all... Your soldering, your metal work, paint, screw, all is look very clean !!!

This is just a compliment... as same many peoples write "giskard is a king of schema" or "Earl is sharp eyes" ... this a compliments not a competition :D :D :D

anyway, bravo again and I subscribe to post of Boputnam... :)

Jean.

[B&KMan]

I have succeeds has to have a little times to make the electrical
measurement of BigNet for found the problem...

first :
To ensure me of the precision of measurement I compared the signal of
origin with that received by wire of BigNet. (measurement taken on wire of the internal connection of the terminals speakers)

The result shows that the addition of all wires, connectors and electronic, vary the response only of 0.25 dB...

x1= original signal
y1 = response signal

so the problem is the rest of system...

[B&KMan]

Each parts is individually verified with the reference instrumental at Solen before fixing ...

ex: in this the big caps is not a real 160uF it is 158 uF so if I add the second 18,2uF plus the bypass 2 Uf + ,01uF the total is

158.000
018.200
001.993
000.001
======
168.994 effective => 170.001 theorical so 0.5 % real deviation !!!
the inductor is specifiquely built for value and verified before and after: 1% variation (difficult because any metal arround affect precision...) anyway largely inside of tolerance design...

so new parts is systematically mesured and verified...

so no problem in this way...

[B&KMan]

The electrical measurement of the HP are not taken at the post
of the drivers, but rather with the last on the connection of the last wire of exit on Bignet. The negative one being always taken in the same place as reference measurement (on the interior wire of the terminal speakers)...

Ok ...

This is comparison of hf and uhf response in 1/3 octave and 1/24 octave.

the theorical crossover frequency is +- 8750 and I have a 8650.

(n.b. the 2 pict the curve is collapsed because I set-up set-up for go up to 40K for more separation of curve of uhf and the weight A & Line representation
As zilch relevant in later post ( and my eyes too) I have rool over at 2 k ... this is the reason I put low the L-Pad... this coincidance is look at peak impedance of this old original diaghram... maybe one response false is here...

[B&KMan]

After checking of the original response curves with those obtained(thanks again at giskard to give original pict), I notice 2 problems important.

1 --- the lack of being able of the UHF makes so that acoustic
energy, product by the addition of the two curves, is weakened... By restoring the capacity of the UHF one should obtain a more
horizontal response around 7-9k.

2--- the original driver with diaphram is not same impedance and probably the notch arround 2K is not present in newer diaphram...

[B&KMan]

verification on med and hf response in 1/3 octave and 1/24 octave for check cross curve...

the original is 1600 hz ( on paper curve) and my mesure expose a arround 1520hz curve.

the form of curve is not same fla because it not a same diaphram in med and in hf.

[B&KMan]

I try to push up the L-pad for boost signal (avg 2 dB) and of course the response is same defective.

See pict.



According to electrical response the drop is really high ...

Soldering problem ???

anybody have a hint to detect bad solder point ???

[B&KMan]

In agreement with the track of the impedance that several raised, and in glances with the great lack of being able of the UHF I took on
me to cancel resistances of 6.5 ohms in series and parrallèle at the
end of the circuit of the UHF....

the adjustment of L-pad is to +1 and produces the following result...

(...)

[Ian Mackenzie]

Dammit! I have been lame on sending the JBL Blue swatch! Sheeeeeit!

Sorry dood - will get one posted, today!

Can you post any measurements of the response, before and after, or is it not something that will impress much in the measuring...? Basically the latter


No worries,

Just get a ruler on some graph paper and draw a 12 inch line.

Frankly I find stuff that particularly boring, people read too much into measurements before they listen...I find it helps if you know what your doing in the first place. The measurement kit basicslly tells me if I have scewed up some where although I've seen a small hand held 1/3 unit by Ivie ? I really like the pretty bar graph.

Its all about your ears Dude..as you know too well.

Ian

[Ian Mackenzie]

:blink:
Sorry If I ask questions who responses is already posts... :o:

But this thread is realy long and it is better easy to resume little bit at few time for better easy compehension and appreciation works...

Well for your first comments... I try to understand where my posts suggest this attitute ??? My intention is just relevant the finest of craftwork: that is that all... Your soldering, your metal work, paint, screw, all is look very clean !!!

This is just a compliment... as same many peoples write "giskard is a king of schema" or "Earl is sharp eyes" ... this a compliments not a competition :D :D :D

anyway, bravo again and I subscribe to post of Boputnam... :)

Jean.

I understand....it was only a joke....and there are certain cultural barriers on the internet.This is why I love to call Ken Alphabet when he is in dire need of assistance...I love his laugh.......and then he knows I'm not being a smart arse.

This unit I built for another member in America for his 4343-4344 upgrade

I think he will like it.

Ian

[B&KMan]

Well, for complete this job I re-keep the mesure of the electrical driver with L-Pad placed at goos mic set-up...

the result is the curve of uhf is less flat, in particulary the 20k is higher, but overrall response appear more flat and normal....

(see pict)

n.b. the uhf is play more higher but at measurement has which go up in Hz, the frequencies are proportionally attenuated by the distance... and directivity. So you remark the 20K is more down than electrical value...

Anyway, in this it not perfect but it is play more correct... now I wait the another time for final placement....

I try to buy a mils spec resistances recommended by list parts of giskard... for improvement result... and after last this modification I go in beach and I put a ionized sand in network (yulk yuk yuk )

It is joke... I put the cover and fill dry sand of special quartkz sand...

anybody;

Do you feeling I hope to put little value resistance for help to flat impedance electrical ???

or this is just OK???

anybody have a experiment this problem with dc charge and uhf ???

thanks at all.


Jean.

[B&KMan]

I understand....it was only a joke....and there are certain cultural barriers on the internet.
Ian

add problem of barriers of language and read and write post easely ... out of phase....

yuk yuk yuk ... ;)


Jean.

[B&KMan]

No worries,

Just get a ruler on some graph paper and draw a 12 inch line.

Frankly I find stuff that particularly boring, people read too much into measurements before they listen...
Ian

My teacher of unniversity repeated unceasingly
"science is not the response to all! but
Between good hands, it is the surest means and rapid between the idea and the realization...

it also makes it possible to evaluate quickly and objectively all and
allowing the repeatability of the experiment.

It is the basement who start art...


For my part, Bright Engineer Greg Timber has in his hands right-hand side a
laboratory and genius in the left...


"B&KMan"

yukyukyuk.


seriously, it is hard to fit criterium on just WOW no ???

But maybe I missing again somes substilitys..

:cheers:


Jean.

[Zilch]

You must normalize the drive curves to determine the actual crossover frequencies....

[B&KMan]

You must normalize the drive curves to determine the actual crossover frequencies....


I dont understand...

it is possible to explain more what is normalize curve ???

thanks...

[B&KMan]

BTW Zilch,

I have maybe a response what is notch of 2K of 2420...

the first pict : upper expose the old electric response keeped on post of driver (all drivers connected on network...) the bottom expose the new electric response at same condition...)


the second pict : upper expose the new response electric on post of 2420 the bottom expose the impedance curve of driver...

Look the notch of impedance is corrolar to notch of electric response...

As your feeling it is possible the peak impedence is create this effect ???


Thanks again...

[Zilch]

I do not know the answer to your impedance peak question. What you suggest makes sense.

Regarding "normalizing," I do not know how your are determining your theoretical and actual crossover frequencies, but see post 398. You have your cursor positioned where the two response curves cross. BUT, since your upper midrange driver has about 8 dB more sensitivity than your lower driver, if you raise the upper response curve that 8 dB to approximate the same output from both drivers, the crossing point and cursor would be moved to a lower frequency.

Normalizing means adjusting so that the outputs are the same, which is the desired objective. As a designer, I believe you would characterize the frequency each filter according a particular point of frequency attenuation in the knee of the curve. In practice, of course, it's where the actual responses "cross."

A member more knowledgeable than myself may be able to clarify these points....

[B&KMan]

Ok now I undertand more what is normalized curve....

1 ---

The impedance of the 2420 selected in white corresponds enough
to the point which appears in the electric response of the 2420.
One would be easily to try to believe that this bump in the impedance
deteriorates to the top frequency response of the 2420....

2 ---

Indeed the answer is standardized and not normalized.
The electrical measurement is based on the calibration of L-Pad
according to best the acoustic answer than I then to obtain...
Thus the 2420 is not in the same way dynamic.
for example if one refers to the diagram of origin that Giskard A
provides it is seen that different the curves are not with the same
height... it is according to these curve and of these cut which I to look at
if, in-sutu, my answers are good...

am I in the error???

[Zilch]

am I in the error???No, you are fine. They are neither the design nor the actual frequencies, is all. As long as you are comparing under the same conditions, it is valid....

[B&KMan]

OK but for the impedence it is possible to corrected the electrical curve with addition of zobel ??



the 2421 is probably not same impedance step peak ???


anybody have impedance curve of 2421 for comparison of 2420 ???


what your feeling....

[Ian Mackenzie]

Jean,

Refer to the PDF of 3145 schematic and the voltage drive measurements.

Do those measurements to verify if your crossover is right/wrong (use an 8 ohm resisters load when measuring DBA of voltage).

If you do this it will resolve all your Questions.

Ian

[Ian Mackenzie]

Several senior members will be opening a new thread shortly "How to Upgrade your 4343 to 4344 Specs".

I can promise you it won't take 372 post to write.

Ian

[B&KMan]

Several senior members will be opening a new thread shortly "How to Upgrade your 4343 to 4344 Specs".
Ian

YOUPPI !!!



:dj-party:





I can promise you it won't take 372 post to write.

Ian


Still again a Australian humour??? :banana:



euh we are with 414 post!!! :duel:



and I do not be finish to post the results of the second network, nor of the final result of the first filled with sand of beach :smthsail:

in fact the first does not have to even finish its burn-in yet...

and testimonial of all members... :spin:


Moreover, yesterday, it passed with brio a next step of burn-in. :applaud:


The sound is clearly big and spacious, it contains definitely more details and of articulation. The holes that there was by the lack of integration is dissipated...
unfortunately, the metal sound persists. in fact it is not ringing or other distortion, but really a color metal in the sound... As same, I changed the caps in the 3143 by film and foil of Solen. This effect should be dissipated completely in 4 weeks...


to follow soon for all fan of this soap thread...


:beach:


:cheers:


Jean.

[Ian Mackenzie]

I am not sure if you saw my earlier post about asking you to measure using BL test Jig.

Have you done that yet? :no:

You appear oblivious to posts offering assistance that is why no one is here anymore.

Even if i made the new equivalent design (I am very short of time) that does not solve the issue of your network being validated as correct??:help: ...that is the issue at the moment so it would be best if you can prove that your nework is functioning correctly using the JBL test procedure before making any claims about the boutique aspects of filter performance.

You can't have any pie if you don't eat your pudding!

So go to your room and don't come out until you have done your homework...:deal:

[4343mod]

I just stumbled onto this thread. Is there any news to share? :)

[B&KMan]

I just stumbled onto this thread. Is there any news to share? :)

hey hey, hey,

well,

Last new , I check of course many aspect: cable, soldering, phase and ....

I have a couple idea what is wrong....


1----

I compare the network and response driver of 3143 and 3145 no surprise I have look the nature of the problem.

If you loof the pict , the hf original 2420 is poor response compare at 2421 the 2421 is more linear and flat, in this circunstance, the hf diaprharm is a issue of the problem of drop in high frec... Normally in original design the uhf is start really more below in fr and support bad response of 2420 so i I cut the uhf in 10 K the section of 6,7,8K is too much drop and create a ole sound in this area...

2---

because the 2121 is too unstable at high frec it is imposible to mach 2420 and 2121 with-out notch upper arround 800K --unfortunately the first notch peak impedance of 2420 create impossibility to cut more down the 2420 for compasation of limit upper of 2121.

So, the bignet amplified this effect and create more feeling of hole sound.

Actually My conclusion is : it is look not very possible to create this network with original diaghram because the limit and difference response of old diaphram react good and bad with new network.

of course the new network is more snaps, fast, details, and because the driver in hf-uhf is rapidly cut, the blur effect of overlaps is really decrease. result more signifiant coherence, less horn effect, and error in time phase...




my last only mysterious problem is the problem of uhf ... I dont know but I cancel the last set of resistor in end of uhf circuit and I pout pot in max for generate correct spl... ???

I know this is same effect with just change original caps on 3143 by caps Solen film and foil... Solen explain the power handling is invert proportionnal at the sensibility. so I feeling due to very power in this section the caps of 400V is too much for the small signal...


Anybody have experienced this effect ???


Anyway I hope you appreciate this part of conclusion... and dont hesitate to expose your point of views.... all is appreciates.

Jean.

[Ian Mackenzie]

Can you verify the voltage drives using 8 ohm resisters please per the table in the 3145 nework?

Ian

[B&KMan]

Can you verify the voltage drives using 8 ohm resisters please per the table in the 3145 nework?

Ian


bonjour,

To afflict for this long deadlines and thank you at all for your patience, But work first !!! And this test takes a few hours to dismount and go up
and measure and adjust and print and compare and comments and...


-------------

Here the complete graph of the electric response of the 3145. I put all L-pads at the max and I substituted for turn of role the HP
measured by a resistance of 8ohms.
(note for this test I reconnect the resistances in UHF section: so all conform to schematic exposed )

( I hope that it is the good way of making... )

-------

The point of crossing of the curves is to indicate as follows with respectively
variation with the maximum:


LF - MF : is 294 Hz variation: LF is 5,82 dB, MF is 3,24

MF - HF : is 1560 Hz variation: MF is 9 dB, HF is 1.58

HF - UHF : is 8540 Hz variation: HF is 8.35 dB, UHF is 8.35

[B&KMan]

This next picture compare the electrical schematic
(est 1225 , p 1 of 2 3145 voltage control full on , 8 Ohms load " internal position" )
versus the response electric of bignet in same situation...

I suspect the variation is my error, I supose one at time, I replace driver by 8 Ohms but I'm feeling in original approach; all drivers is replaced by 8 Ohms (all the time) ...

but 2 schema of voltage is posted By Mr widget and I dont know what expression internal position and external position is....

:blink:

or

the L_PAD is scrap and loss to much power by age
( I recycling the original L_PAD (the cost is realy higher...)


anybody have idea, please post it....

thanks for all

Jean.

[Ian Mackenzie]

Compare

[Ian Mackenzie]

Small adjustment of the L pads 'Levels should compensate for the variations.

Ian

[B&KMan]

Compare

thanks !!!

It is possible to fill explanation of legend what is goud curve of hf uhf ???

:cheers:

jean.

[B&KMan]

Small adjustment of the L pads 'Levels should compensate for the variations.

Ian

So I suppose right, if your conclusion of the elecrical set-up is correct ???

Jean

[Ian Mackenzie]

Note you must use either 2421 or 2426 drive specification with the 3145 crossover. The 4343 uses 2420.

Ian

[B&KMan]

Note you must use either 2421 or 2426 drive specification with the 3145 crossover. The 4343 uses 2420.

Ian

Thanks for reply,

yes the drivers is not same and it is my hypothesis for the nature of the problem

http://www.audioheritage.org/vbulletin/showpost.php?p=65212&postcount=418

i have a feeling the 2420 have problem to go more higher and is have a big notch instability arround 2 k. So impossible situation here in this set-up : big upper notch in start and too drop hole at finish. I'm shure the 2405 is pushed more lower in version of 4343 for patch this problem. But creat a notch response on start.


--- the first conclusion of this upgrade is not fully success if you keep the original diaphram of the 2420...

Of course the soundresult is, for many aspects, better . But, because more transparent: the notch and hole is.... more exposed...



if any members have idea, please post-it

:cheers:


Jean.

[Ian Mackenzie]

Well buy a 2426 diaphram!

[B&KMan]

Well buy a 2426 diaphram!


Hello Mr.. Mackenzie,

I hope not to appear too critical to you with all your good councils, but I read several point of sights which regard the titanium as too
metalic and which the best choice would be the 2421...

Can more expose your opignion please.

Thnaks

Jean.


In other way I suspect maybe the L-Pad is part of problem so I tried to work benchtest of pad befare running big money for diaghragm.


:cheers:

[B&KMan]

I have no idea why you decided to upgrade to a new equivalent crossover without upgrading your drivers...you appear hopelessly confused:dont-know and while your predisposition with vibrations is interesting its not the point of the exercise.
My expertise in vibrations extends well into the sub octaves:moon:, Pooo!:wasnt-me:

The 2121 diaphrams are very expensive and less robust then the 2426 diaphrams. The latter is still much better than the 2420 dias, smoother and sweeter. Its the best improvement along with a 2122 recone for your system.

I have used the 2426 diaphrams from some years now and you never hear me complain.

I dont know what is the problem with you and me but actually the level of insult is higher and only on your way...

I write with difficulty in english, and I started a modification of my 4343,

I sended over 10 posts for verified the compatibility with drivers....
for your great expertise is maybe appear stupid to not upgrade at same time but...
this is exactly the interest of the forum... share information.

it is easy to shit on the guy is not know..
and you have not only to posess the truth...



because I have tool and expertise into vibration, I have decided to exchange information with members. it is my gift action for recognize the helps of members in my quest. This aspect is big aspect of 4343 modification and it is include in many serious design...

nobody is perfect,

(look as your own post : you wrote 2121 instead of 2421 ).


I appreciate your reply information and thanks for that...

I hope your "bad vibration" on me is not too low and power.


Peace Mr. Mackenzie !!!!

[Tom Loizeaux]

Jean,
Your work on your 4343s has been of some interest to me as I too have a vintage pair of these. Your zest for rebuilding these and efforts in trying to improve these is all a bit over my head. Your interest in re-designing these crossovers in more ambitious then anything I would attempt.
When I rebuilt mine, I tried to restore them to original specs. I had to repair the cabinets, do major clean up on everything, replace the fiberglass, repair the grille frames, find replacement nameplates (thanks Guido), and re-foam or re-cone drivers. I also tried to upgrade the caps in the original crossovers. I installed some quality caps and did some bypassing, especially on the HF and UHF circuits.
Over the past year I've carefully adjusted my active crossover, set levels and adjusted the balance with my power amps and L pads to get what I think is nearing the perfect balance.
I did buy a pair of JBL 2421 diaphragms, but haven't installed them yet because the original ones still sound really great.
I feel that one needs to make small changes, and make them one at a time, in order to know what difference they're actually hearing. Too much too fast only leads to uncertainty.
This fall or winter I plan to open my 4343s up again. Then I'll install the new diaphragms and I might put in a few more high end caps. I consider this part of my slow, fine-tuning process.

I will read about your efforts and percieved results, but I'll take your experiences and advice very cautiously.
Thanks again for your enthusiasm!

Tom

[B&KMan]

THANKS AGAIN AT ALL MEMBERS IS KEEP TIME TO PERMIT REALISATION OF THIS PROJECT, SORRY IF I PUSH YOU TO REPEAT AND STANDY... BUT I SEND TON OF PICT AND TEST RECOMMANDED BY YOU, I FEEL HONEST IN MY WAY BUT FEW MEMBERS IS NOT REALIZE THE DIFFICULTY TO PLAY IN FUNDAMENTAL THEORY WITH A POOR LANGUAGE AND NOT FORMATION IN THIS FIELD...

I'M KEEP SUPER FEELING FOR SUPPORT OF THIS FORUM AT ME WITH THIS LONG THREAD BUT MANY FUNDAMENTAL QUESTION IS TALKING AND I HOPE TRACE OF THIS QUEST HELP MANY MANY GUYS. THIS PROCESS EXPOSE THE DIFFICULTY TO SURPASS THE ENGENEERING CONCEPT. AND MANY CONTROVERSE EXPOSE THE WAY OF EACH AND COMPROMISE...

OF COURSE BETTER SIMPLE NETWORK BETTER OUTPUT BUT IN FEW CASE MANY TWEAK IS BETTER THAN SIMPE APPROACH.

IN FACT YOU HAVE OVER 8000 VISITS IN THIS THREAD EXPOSE THE INTEREST OF FEW POEPLES.

I SEND THE FINAL PICT ON THE COVER OF THE BACK PLATE OF "BIGNET" (DEFAULT NAME).

THANKS PARTICULARY AT
:
MR MACKENZIE,
MR WIDGET,
MR GISKARD,
MR ZILCH,
MR EARL K,
MR GUIDO,
MR ROBH3606,
MR TOM LOIZEAUX,


AND FEW OTHER IS INSPIRE SUPPORT AND TRACE THE WAY OF THIS PROJECT BY EXCELLENT REMARKS, SUPPORT SPEC AND MATERIAL AND EXPERIENCES.

I HOPE MY LITTLE EXPERIENCE IN VIBRATION IS EXPOSE CERTAINS UNUSUAL ASPECT OF CONCEPTION AND FEEL FREE TO HOPE ANY QUESTIONS IN THIS REGARDS. ( THIS IS MY CONTRIBUTION)



NOW I LIVE FOR MONTH IN VACANCY AND IT IS TIME TO PLAY LOUD BEFORE !!!!



:dj-party:


THANKS JBL GANG


Jean.

[B&KMan]

I SEND THE FINAL PICT ON THE COVER OF THE BACK PLATE.



heres.

put hand soap for fill in the plywood : it is really tight set-up anf tuff the speaker is go away... shure it is not fill on sand...

by the sand this sans is silice sans (same as sand blaw of pool filter it is more weight of sand and it is more density... for abrasif fume (be shure you run this with adaquate instruction)

:cheers:

[B&KMan]

by the sand this sans is silice sans (same as sand blaw of pool filter it is more weight of sand and it is more density... for abrasif fume (be shure you run this with adaquate instruction)

:cheers:


OHHH BTW I put 1 and 1/2 bag of this sand so the speaker is increase weight by another 120 LBS (50 Kilos) add 50 lbs for network + 180 lbs of original weight of speaker the final is 250 LBS

Why the bass is snap realy tight and fast ???

anyway this is a first 4343 beach JBL. ans the sound is the sea


AHHHH paradise ???


:cheers:

[B&KMan]

Jean,
Your work on your 4343s has been of some interest to me as I too have a vintage pair of these. Your zest for rebuilding these and efforts in trying to improve these is all a bit over my head. Your interest in re-designing these crossovers in more ambitious then anything I would attempt.
When I rebuilt mine, I tried to restore them to original specs. I had to repair the cabinets, do major clean up on everything, replace the fiberglass, repair the grille frames, find replacement nameplates (thanks Guido), and re-foam or re-cone drivers. I also tried to upgrade the caps in the original crossovers. I installed some quality caps and did some bypassing, especially on the HF and UHF circuits.
Over the past year I've carefully adjusted my active crossover, set levels and adjusted the balance with my power amps and L pads to get what I think is nearing the perfect balance.
I did buy a pair of JBL 2421 diaphragms, but haven't installed them yet because the original ones still sound really great.
I feel that one needs to make small changes, and make them one at a time, in order to know what difference they're actually hearing. Too much too fast only leads to uncertainty.
This fall or winter I plan to open my 4343s up again. Then I'll install the new diaphragms and I might put in a few more high end caps. I consider this part of my slow, fine-tuning process.

I will read about your efforts and percieved results, but I'll take your experiences and advice very cautiously.
Thanks again for your enthusiasm!

Tom


thanks

I understand you process, but I burn step in regards of big expertise on the great members and because support by my first grade instrumental lab... otherwise be shure: one step at time is excellent process for keep the control. each room contribute to 50% result (sic) the rest of system is affect the final result => this is a chain !!! so control all parameters and assume one link is default and not other it is really tuff project....

:cheers:

and dont hesitate if you have futur question.

Jean.


:cheers:

[B&KMan]

I have a few minutes again for hint and tips...

The back panel is put on speaker and I listen the difference before fill sand....

No surprise but surprise I dont mesure the level of effect of diffuse sound is push very higher background realism!! all is more real, wide and deep. arround 200% better. the speaker before is really good integration in room and dissapear... but now is the room entirely arround the speaker is disappear !!!

I'm understand many high end speakers is build with pattern of back dissipation effect

( it is part of secret of k2 9800 ???) ...


:Cheers: again.

[rek50]

Jean, I like your "Experiment". Your exposure to vibration and instruments (B&K MAN) opens many doors, that others may have not entered, or cared to. After all, how many people have the testing equipment that you are fond of. To each his own, is an old saying, as well as, beauty is in the eye of the beholder (Beer-Holder, for some)... Enjoy the magic sounds, there are so many to choose from.

[Ian Mackenzie]

External crossover response curves, bi mode using hi end active crossover.

Here are some graphs taken with my very crude equipment by your standards.

The 1st graph is taken 1.5 metre on axis in the listening room . All controls at o position, no smoothing. The purpose og this curve is to proof overall system flatness. Compared to JBL data, the overall curve is almost identical.

The levels were set as previously descibed using a test tone and a Tandy spl meter. There appears no other way of accurately determining the 0 db position.

The other graphs show the individual response of the mid, horn and slot devices with the external crossover. The last graph is the right hand speaker.The slot devices appear to vary from unit to unit in response characteristics, therefore one should attempt to obtain matched pairs of devices if upgrading a system.

Ian

[Ian Mackenzie]

The original JBL test data of the 4345 system, controls set at zero.

While the above curves represent the performance of the 4345 18 inch system , those with 4344's or considering upgrading a 4343 to a 4344 should obtain similar if not identifcal performance. I do think however the offset drivers on the baffle in the 4344/45 accounts for an improvement in the system smoothness over the 4343, apart from the more advance mid and HF drivers. Perhap the imaging is also improved..no flys on the 4345 in that regard..well only one way to finds out.

Perhaps that is the ultimate upgrade for a 4343 owner, make a new replaceable baffle. The drawings are around somewhere deep in the forum.

I remember doing the measurements and the drawings...quite painful actually when you have a hangover like the one I have at the moment..time for quite one at the Pint on Punt.....LOL

Ian

[4343mod]

*snip*...Perhaps that is the ultimate upgrade for a 4343 owner, make a new replaceable baffle. The drawings are around somewhere deep in the forum...
Ian

Agreed, but make the new bolt-on baffle to bring the 8" up to a 12", maybe a D130? with Tatrix Horns! ..Leave the 15" alone, just add another cab and another 15" per side. Anybody here against (2) 15's per side for the extra 3dB?

How about this? Now add (4) 2-ch. amps, (1) DriveRack 480 in 2x4 stereo?

[Ian Mackenzie]

Its actually a 10 inch driver specifically designed for the system.

Other drivers may work but the network may require considerable re design.

Ian

[Ian Mackenzie]

Here are the voltage drives for the respective acoustic graphs above.

Ian

[Ian Mackenzie]

So with the above information and an inductance meter its almost a bomb proof exercise to update/modify the 4343 to a 4344 crossover spec.

You must however invest in updated diaphrams and recones in order that the network will work in a complementry manner to produce the required response.

Otherwise you've got you head in the sand thinking it will be a 4344.

For those content with their vintage 4343 components consider building a charge coupled version of the 3143 network and enjoy the improvements of an updated network.

Ian

[B&KMan]

http://www.audioheritage.org/vbulletin/showpost.php?p=60751&postcount=308


I again, I read this thread for check, resume and verified if all is OK... before build the second DC charge 3145 sanded network :D

And I realize I not expose my result in regard of phase .... and maybe I have a question for you great members...

============

the original 3145 expose the 15" inch is non inverted polarity post-current and I put positif (+) electrical amp signal to neg (+) post driver post. and all other driver is invert so all positif signal amp is connected into red (-) post driver of 2121, 2420 & 2405...

Well this is a mistake of my interpretation of schema...

but after check for invert problem of signal of cable and connection into L-Pad
I perform phase match signal and acoustic response

It is appear the 3 upper driver is inverted signal connection. Yup the 2121, 2420 & 2405 is inverted phase response

after test and test and test and verified cable schema...

the final set-up is all black post driver (+) (old convention JBL)
is connected into positif signal amp.

final;

all (+) signal is connected on black post driver (+) of all

:blink:


But For just shure it is not a weird mistake of me

please Earl and Giskard (specialised of 3145 schema) it is possible to just confirm for other reader this configuration connection is good ...

Otherwise other mistery is on cover ....

:cheers:

[Ian Mackenzie]

The output of the network schematic is drawn correctly.

The positive and negative of each leg of the network goes to red and black speaker connection respectively.

The phase is inverted within the schematics to the mid/hf andf uhf devices.

[B&KMan]

Agreed, but make the new bolt-on baffle to bring the 8" up to a 12", maybe a D130? with Tatrix Horns! ..Leave the 15" alone, just add another cab and another 15" per side. Anybody here against (2) 15's per side for the extra 3dB?

How about this? Now add (4) 2-ch. amps, (1) DriveRack 480 in 2x4 stereo?

it is look of 4355 project no ???

:cheers:

[B&KMan]

The output of the network schematic is drawn correctly.

The positive and negative of each leg of the network goes to red and black speaker connection respectively.

The phase is inverted within the schematics to the mid/hf andf uhf devices.

Well, just for shure if I'm realy understand:

the final result is :

if you send (+) impulse signal electric into a a speaker, normally if network is build correctly, all driver is go foward motion phase. the cone is go outward motion right ??

[Ian Mackenzie]

Well, just for shure if I'm realy understand:

the final result is :

if you send (+) impulse signal electric into a a speaker, normally if network is build correctly, all driver is go foward motion phase. the cone is go outward motion right ??

WRONG.:banghead: :crying: :help: :no: :idea:

The woofer should move inward with a positive voltage applied to the RED input terminal of the crossover.

The midrange/horn and uhf should move outward with a positive voltage applied to the RED terminal of the crossover. (this is not easy to confirm without bypassing the filters).

[B&KMan]

WRONG.:banghead: :crying: :help: :no: :idea:

The woofer should move inward with a positive voltage applied to the RED input terminal of the crossover.

The midrange/horn and uhf should move outward with a positive voltage applied to the RED terminal of the crossover. (this is not easy to confirm without bypassing the filters).


Thanks for reply but I just try without doubt your response.


1--- in theory, dont matter what code connections or nature of network... the final phase result is all driver have a outward motion right ??


2--- in regard of the theory, I understand your response is according to connection of network : and because the phase of 2121,2420 & 2405 is inverted by circuit network it is important to connect inverted the MF, HF & UHF.


3--- All this old drivers have a (+) post = black right??


4--- Manys folks is expose the JBL monitor is load the LF inverted phase for preload the inside speaker by inward motion. so independently the connection what is final motion of each driver ??? all outward ??? LF inward and all other outward ???


thanks again for light ...


Jean.





p.s. (this is not easy to confirm without bypassing the filters).

You have right... this question is delicate set-up and I study hard for control this parameter with my analyser... but it is easy now to perform this test with filters network load... after response , if you wish , I perform pict of this test.

:cheers:

[Ian Mackenzie]

Thanks for reply but I just try without doubt your response.


1--- in theory, dont matter what code connections or nature of network... the final phase result is all driver have a outward motion right ?? Wrong


2--- in regard of the theory, I understand your response is according to connection of network : and because the phase of 2121,2420 & 2405 is inverted by circuit network it is important to connect inverted the MF, HF & UHF. See schematic


3--- All this old drivers have a (+) post = black right?? see JBLPRO Tech Library pahse conventions.


4--- Manys folks is expose the JBL monitor is load the LF inverted phase for preload the inside speaker by inward motion. so independently the connection what is final motion of each driver ??? all outward ???

LF inward and all other outward ??? Y


thanks again for light ...of day


Jean.





You have right... this question is delicate set-up and I study hard for control this parameter with my analyser... but it is easy now to perform this test with filters network load... after response , if you wish , I perform pict of this test.

:cheers:

Ian

[B&KMan]

:blink: :blink: :blink: :blink: :blink:

WELL I COMPLETELY CONFUSE....


actually My old 4343 with original network 3143 is connected exactly same as the shematic representation.

all drivers of this speaker have a outward motion

I perform test phase mesure of speaker with new network.

the LF is outward but all other driver have inward motion if I connect with schematic... the sound is really different and appear more dark than old 4343.

After connect all in outward motion same acoustic phase , I realise immediately the sound on new is look better same at old speakers but more refined and details.

Normally all speaker produce final outward motion and because the network processe is invert phase, it is important to re-adjustement polarity in regard of this. But I konw the shema is good and Mr mackensie is write right but real acoustic invert phase is appear really not good in all my books. In JBL site it is really important to final connection is syncro to the input mic... (contractor litterature and all others installation recommandation.) So the final result it is LF inward and other outward ??

see pict for my test phase:

1---- set-up direct with big driver (thiel) for expose pulse motion easy perceived on eyes.
(the driver move half inch motion pulse Yiah baby real pulse test here without doubt... )

2--- pict expose the result of outward motion : the phase response in botton

3--- pict expose the result of inward motion :the phase response is upper

realy easy interpretation and not confusion here...

So what is what??

all connection polarity is verified all time... maybe the bigger inductor and caps double the rotation phase and re-align the phase ???

JBL is perform 4345 speaker with LF Inward and the 4343 is all outward ???

[Zilch]

"JBL networks and transducers are designed so that if the red terminal of each transcucer is connected to the appropriate high terminal on the network, the system will be electrically in phase and diaphragm movements will be consistent. However, this arrangement may not result in proper acoustical phasing."


"The connection resulting in proper acoustical phasing of any combination of components is difficut to determine in advance of assembly. Proper acoustical phasing of low and high frequency tansducers will be that electrical phasing which produces maximum acoustic outpout through the crossover region. In general, if the diaphragms of the various transducers are in the same physical plane or arc, in-phase electrical connection will probably result in proper acoustic phasing. If the diaphragms are not in the same plane or arc, proper phasing should be determined experimentally based on analysis of pink noise or subjective evaluation of program material."

JBL varies the electrical phasing of drivers in their products to achieve their design acoustical performance from specific combinations of drivers, physical alignments, and network topologies. It is not uncommon to find one or more drivers wired reverse-phase with respect to others within standard systems. JBL's standard wiring is generally considered as comprising the "right" connection. Mileage may vary....

[B&KMan]

"JBL networks and transducers are designed so that if the red terminal of each transcucer is connected to the appropriate high terminal on the network, the system will be electrically in phase and diaphragm movements will be consistent. However, this arrangement may not result in proper acoustical phasing."


"The connection resulting in proper acoustical phasing of any combination of components is difficut to determine in advance of assembly. Proper acoustical phasing of low and high frequency tansducers will be that electrical phasing which produces maximum acoustic outpout through the crossover region. In general, if the diaphragms of the various transducers are in the same physical plane or arc, in-phase electrical connection will probably result in proper acoustic phasing. If the diaphragms are not in the same plane or arc, proper phasing should be determined experimentally based on analysis of pink noise or subjective evaluation of program material."

JBL varies the electrical phasing of drivers in their products to achieve their design acoustical performance from specific combinations of drivers, physical layouts, and network topologies. It is not uncommon to find one or more drivers wired reverse-phase with respect to others within standard systems. JBL's standard wiring is generally considered as comprising the "right" connection. Mileage may vary....


Thanks Zilch for reply but I'm perplex ... :blink:

1--- my 4343 is " same " patern as 4345 ... so the phase is normally same .

2--- My test phase is realy magnifying the phase response driver but if I put mic on 1 meter and produce pulse and verified phase....

the result is really evident if one driver is out of phase because the segment corresponding of the response driver is inverted !!! so this type of test is relatively absolute phase response evaluation.

3--- JBL expose clearly in technote :
" JBL_TechNoteN1V12C_v5.pdf " and
" tn_v1n12b.pdf "

the importance of final result is outward motion to correspond at the real pressure response of mic.... I have doubt is make a opposite way in monitor studio and in same time is expose the importance for contractor, installator to respect the original phase... and expose the problem in technote for respect of multi amp system and active network...

4--- It is not make sense to invert phase and destroyed the phase response for correction of curve... the unnatural inverted phase expose more trouble than 3 db correction... in fact any good room affect more balance response than 3 db.... pls it is easy to corrected the drop of boomer in start of network by just little displacement....

strange realy strange...

[Zilch]

LF driver is wired negative polarity in 3143 network, and the other three are reversed, assuming JBL convention of solid color wire connecting to red terminals of drivers.

You are reading sound reinforcement technical information, where it is important that everything be in phase with the live instruments. Even then, however, recognize that within individual speakers, the phase of drivers is optimized for performance of that individual box, and the "everything in phase" rule does not apply.

It matters even less in studio and control monitors, and least in home systems, where the phase of the live instruments is not an issue. No speaker designer is going to compensate for phase cancellation in the crossover region due to driver alignment for the sake of preserving electrical phase integrity.

It's obvious that JBL didn't, in any case, and that is nothing in comparison to the fact that JBL phase convention itself is the inverse of industry standard, and this extended to its amplifiers as well, up until a few years ago.

If you believe your system sounds better with the LF driver phase reversed from the schematic, swap the wires and forget about it. At least you didn't decide that all of the others are "wrong." Alternatively, if you want to switch everything to Chebychev filters, have AT it. A "little displacement" is not going to work at 350 Hz....

Read and weep: http://www.jblpro.com/pub/technote/tn_v1n12b.pdf

See also: http://audioheritage.org/vbulletin/showthread.php?t=487 where we danced this number nine months ago:

"This is not sense??? http://audioheritage.org/vbulletin/images/smilies/banghead.gif (...)"

[Same tune....]

[B&KMan]

1---- LF driver is wired negative polarity in 3143 network, and the other three are reversed, assuming JBL convention of solid color wire connecting to red terminals of drivers.

2---- If you believe your system sounds better with the LF driver phase reversed from the schematic, swap the wires and forget about it. At least you didn't decide that all of the others are "wrong." Alternatively, if you want to switch everything to Chebychev filters, have AT it....

See also: http://audioheritage.org/vbulletin/showthread.php?t=487 where we danced this number nine months ago:



Thanks again Zilch,

Yes the LF on 4343 & 3143Network is invert but the final result is good phase all the way.

http://audioheritage.org/vbulletin/showpost.php?p=33180&postcount=82


I expose the difference to connect wrong and true polarity and result of phase match... in accord of the schema.

and I expose if you plug LF in invert (in corrolation of schematic) the final result is foward motion !!! not inward...

--------
You have right again for the question of taste : the final word is my ear ( for copy the celebrity word of Mr Mackensie).

-------

But for many other members and for the objective question, I stay in my question...


Do you have feeling because I keep more biggest wire and bigger caps the phase shift 180°

the problem phase here is weird :

1--- same box and driver

2--- I change network;

3---- result change phase ???

original design : all driver is push outward

well, I go to check again my connection :banghead:

[Robh3606]

Well if you charge couple and have 2 caps in series what is the phase change?? 90 or 180?? Intuition says to add the changes just like 3 pole 18db network. Figure with a charge coupled 18dB network you could get an additional 180 phase change. So from 270 to 450 if I am correct in this assumption. Depending on how many caps are in series to the drivers it could change things a bit. That puts you in the wrong quadrant as far as phasing compared to the original. To end up in the same quad you would have to inverse phase on the driver relative to the original well, I think, phase makes my head hurt:banghead:


Rob:)

[B&KMan]

Thanks Robh3606 for this hurt experiment answer... :D

1 --- firstly, it is certain that the phase acoustic measured is obtained
by comparison with the origin signal. In this direction that requires an apparatus bi-channel with integration of phase. What is the case here. the microphone is 1/4; thus extremely faithful to the signature of impusion of the diaghragm and very faithful to the response in phase.

with the result that independent of all rotations of phase created
by the filters, the real final acoustic result of the phase can be evaluated clearly.

it seems to to me under these conditions, you hold the explanation, it would be interesting to see whether one 4344 mkII with a network in
DC charges, and by keep in mind the drivers which function according to the
new convention of polarity reacts is if according to this explanation. For my part, I have difficulty with discern the schema of the 4344 MK II to
find an answer there.

In addition, it should be said that in more of producing a circuit in DC charges, I also the concept of cascade introduces.

In this mess electronic the final result it is a outward motion coherence for all drive in according of original 4343 & 3143 network no ??

thanks again.

[Ian Mackenzie]

Thanks Robh3606 for this hurt experiment answer... :D

1 --- firstly, it is certain that the phase acoustic measured is obtained
by comparison with the origin signal. In this direction that requires an apparatus bi-channel with integration of phase. What is the case here. the microphone is 1/4; thus extremely faithful to the signature of impusion of the diaghragm and very faithful to the response in phase.

with the result that independent of all rotations of phase created
by the filters, the real final acoustic result of the phase can be evaluated clearly.

it seems to to me under these conditions, you hold the explanation, it would be interesting to see whether one 4344 mkII with a network in
DC charges, and by keep in mind the drivers which function according to the
new convention of polarity reacts is if according to this explanation. For my part, I have difficulty with discern the schema of the 4344 MK II to
find an answer there.

In addition, it should be said that in more of producing a circuit in DC charges, I also the concept of cascade introduces.

In this mess electronic the final result it is a outward motion coherence for all drive in according of original 4343 & 3143 network no ??

thanks again.

You appear to be looking too deep that this issue.

Provided all the drivers have "like"" phase convention the schematics for both the 4343 and 4344 remain true.

Refer to basic 2nd order crossover theory at Rod Elliotts Sound pages for a thorough explanation of filter phase shifts.

Ian

[B&KMan]

Well for manys interesting by more result of phase.

I send pict of mesure of impulse test and result keeping my mic realy close of driver.

the first pict is the result of original form of impulse in nyquist presentation.

the second is the result produce by driver and keeped by mic.

[B&KMan]

Well for manys interesting by more result of phase.

I send pict of mesure of impulse test and result keeping my mic realy close of driver.

the first pict is the result of original form of impulse in nyquist presentation.

the second is the result produce by driver and keeped by mic.


Next is the result of pict A comared by result of pict B and integration for final difference result Normally called H1

the first pict is result in in phase and the second is out of phase in look of nyquist plot.

look the mirror effect ...

[B&KMan]

Next is the result of pict A compared by result of pict B and integration for final difference result Normally called H1


the next pict is expose maybe the better differential in response in phase and out of phase with nichol diagram.

the time delay appear in out of phase (pict 2 ) expose the problem of bad phase.

pict 1 is correct phase time.

[Earl K]

Hi


Well if you charge couple and have 2 caps in series what is the phase change?? 90 or 180??

- 2 caps wired end to end ( in series ), retain the expected 90° phase shift of a single cap .

- Also, a DC bias applied to the junction of those 2 caps , is really beside the point / though ,,,,, see below.

- I'll let someone else bench test my above assertions if you wish to prove me wrong . :)
( testing for a 180° phase shift "or polarity flip" is easy enough to be labelled simplistic )

Intuition says to add the changes just like 3 pole 18db network. Figure with a charge coupled 18dB network you could get an additional 180 phase change. So from 270 to 450 if I am correct in this assumption. Depending on how many caps are in series to the drivers it could change things a bit. That puts you in the wrong quadrant as far as phasing compared to the original. To end up in the same quad you would have to inverse phase on the driver relative to the original well, I think, phase makes my head hurt

- The rest of this hypothesis ( exra phase rotation within the poles ) falls apart if one can't establish the primary contention found in the first quote ( ie ; 2 caps wired in series = 180° phase shift ). Once again, FTR, the cumulative phase shift is 90° .

- I'm sure that Jeans' N3145 phase challenges result from some other overlooked area ( though right now, I don't know what that might be / reviewing 12 pages of fractured communications is not my idea of a good time ).

- Anyone around here with a stock 4345 to test ? If so and if you have the ability to measure relative phases/polarities ( ie with that "Cricket thingy that Bo has previously mentioned ), does the woofer ( 2245h ) move outwards with a positive pulse applied ? Now, what about the 2122h ?

- Now, how about those of you with your proprietary 4344(s) or 4345(s) . With a positive pulse applied, what is the relative direction of cone travel for the 2235/2245 and 2122h transducers ?

Thanks <> EarlK

[Ian Mackenzie]

Earl,

Hi,


In short the woofers move inward, the 2122H outward....yawn.

I'm going back to bed.....:snore:

Ian

[B&KMan]

Thanks Earl for try explanation.

well, in definitive, Only stock owner have a response and JBL service, so I try to contact service center for this question... but it is twicky because I dont find the color code connection just final acoustic phase response of 4344 or 4345..

Anyway I perform other test for dissipate question of integration phase in two drivers...

I put my mic in 1 meter 0° and in between LF and MF for mesure impulse response and analysing phase acoustic response.
(of course at 1 meter echo is contaminated the original phase so I cut echo response but it is result to cut the original end signal too. so do not put importance to detail amplitude phase just the side is important here...

I make this set-up for my original stock 4343 and my 4343 with 3145 DC +cascade set-up.

the first pict expose the phase result (perfect coherence in low (2231 + 2235 diaghram) and mid (2121 +2121 diaghram).

all driver it outward motion !!! original stock with respect of color code connection 4343 + 3143 network...

now 3145 DC +cascade set-up plugged no according to schematic expose same result if I plug invert config of schema 3145 mid !!!!

[B&KMan]

#!@$#$%#$%?%?&

Sorry for all readers...

I perform few time ago a test for my brother who not accept the LF is run Original inward motion and I omit to reconnect my LF in good original sense of my 4343 stock...

so I keep in mind a confusion and run my all test in this wrong assertion.

because I confuse too all in wrong test I keep a couple time for reset all parameter and I will come back with good value...

Sorry again.


Jean.

[B&KMan]

Refer to basic 2nd order crossover theory at Rod Elliotts Sound pages for a thorough explanation of filter phase shifts.
Ian


Thanks Ian for info,

After long process read, I realize - if this guy is expose all datas arround phase passive network- the question is not finally if outward motion or inward motion but the cancellation arround cross network area in two driver. specially in LF-MF.

so I perform pink noise, mic a 1 meter at half distance in LF driver and Mid driver
I do not modified any other parameter than the side connection of LF on dc 3145 (bigNet).

the result is maybe not surprise you but the LF is better integration if is outphase (inward motion) (black wire with black post driver) and the 2121 is outward motion.

see pict.
(dashe line LF inward motion + MF outward motion )
( plain line LF outward motion + MF outward motion )


Ok for that ... this is standart connection of 4343


but, actually the schematic (courtesy of Giskard ) expose exactly inverse connection ==> LF in phase (outward motion ) and all other driver outphase (inward motion).


:blink: :blink: :blink:

+ with + and - with -

but old convention polarity expose black post driver is + and red is neg..

I know the driver is not same but is appear to work on old convention polarity JBL ... what is final word of this connection set-up ???

Giskard ?? others members who work on 3145 shematic ???

help on your light please...


Jean.

[B&KMan]

same test but with original stock 4343 with 3143 network.

see pict.

( plain line LF inward motion + MF outward motion )
( dashe line LF outward motion + MF outward motion )


Ok for that ... this is standart connection of 4343

I remark the difference of out phase and in-phase is more important here than 3145 DC charge....

[Ian Mackenzie]

Okay,

I was wondering about the bass/mid intergration?

What value choke and capacitors is on the woofer filter also also the 2121 midrange. Are you using 2121 mid range or 2122 midrange ?

I went back to your earlier posts and saw the charge coupled schematic with the 3145 network values scaled for charge coulping.

If you are using the 2121 midrange the net woofer values should be 5.4mh and 72uf. The midrange should be 2.8 mh and 52 uf as the 3145 network is designed around the 2235H/2245H and the 2122H.

Again I only raise this points as you suggested earlier it sounded strange either way of polarity.

Ian

[Ian Mackenzie]

Also i saw in your pictures the internal JBL rotary switch .:eek:

I would be horrified if you have not bypassed that switch!:D

Also, can you advise how long the sand take to settle? Do you need to compact the sand down or allow it to settle over time then top up.

I read in UK speaker manual sand is very good but needs time to settle.

Ian

[B&KMan]

What value choke and capacitors is on the woofer filter also also the 2121 midrange. Are you using 2121 mid range or 2122 midrange ?

I went back to your earlier posts and saw the charge coupled schematic with the 3145 network values scaled for charge coulping.

If you are using the 2121 midrange the net woofer values should be 5.4mh and 72uf. The midrange should be 2.8 mh and 52 uf as the 3145 network is designed around the 2235H/2245H and the 2122H.

Again I only raise this points as you suggested earlier it sounded strange either way of polarity.

Ian




:blink:


Hi,

yes This is original 2121 with new old stock 2121 diaghram. (around 3 years )

but the 2231 is recone with only now 2235 kit... so I have hybrid

I build my network in according of schema so the LF is 5.4 mH and mf is 1.8 in series and 4.8mH in parrallele.


I have no time this week for recalibrate sytem in particular 2121 for evaluate taste of integration. in first impression I loose bass :barf: but it is normal in fact the high end spectrum of LF and low of mid is higher of 3-6 db.... change global ptch of speaker response....

but next week I'm back with correction L-Pad and feeling of taste.

but this is not explain why 3145 is plug LF in outward motion and all the rest of driver is inward ???

[Ian Mackenzie]

:blink:


Hi,

yes This is original 2121 with new old stock 2121 diaghram. (around 3 years )

but the 2231 is recone with only now 2235 kit... so I have hybrid

I build my network in according of schema so the LF is 5.4 mH and mf is 1.8 in series and 4.8mH in parrallele.


I have no time this week for recalibrate sytem in particular 2121 for evaluate taste of integration. in first impression I loose bass :barf: but it is normal in fact the high end spectrum of LF and low of mid is higher of 3-6 db.... change global ptch of speaker response....

but next week I'm back with correction L-Pad and feeling of taste.

but this is not explain why 3145 is plug LF in outward motion and all the rest of driver is inward ???

OOPS.
The coil should be 2.8 mh in parrellel as I recall!! ....hope this is not too disturbing.

By the way Do you watch Doctor Who? Sarah Jane Smith (on the motor bike) says hello and she admires your persistence.

[B&KMan]

Also i saw in your pictures the internal JBL rotary switch .:eek:

I would be horrified if you have not bypassed that switch!:D

Also, can you advise how long the sand take to settle? Do you need to compact the sand down or allow it to settle over time then top up.

I read in UK speaker manual sand is very good but needs time to settle.

Ian

1---- the switch is cancelled in first project modification with cap solen on original network 3143 the cancellation switch have improved really really more sound than new caps. it is a responsable to correction of curve of LF and more power in mid. but i not pull-up switch, just cut wire and put soldering...

maybe one guy in futur is wanted vintage speaker.... :D all pieces is keeped and stored...


2--- the nature of sand is associated of the level of gap settle.

this quarzt sand is more sharp and regular grid so the level of settle is not a big variation... it is same sand for sandblast or for filter of pool. just warning it is a abbrasive sand and fume is corrosif.... but the density is more than standart sand and no imperfection or surprise humidity. so better .

well, if you shake firmly the speaker for put sand all in small intertice, and you put a baf on back panel in couple of minutes the big sette is accomplish, maybe max 1 litre is next couple of time you arrival to fill step by step.

but if if you have a small gap it is not critical because the sand is recover completely the caps and others parts.... and other hand the majority of energy is at the level of LF so the inertia is garantee ad sand is not really create a vibration. finnaly at each fill the next is maybe 5 % so after couples of week you put a couple of spoon for fill up.


:cheers:

[B&KMan]

OOPS.
The coil should be 2.8 mh in parrellel as I recall!! ....hope this is not too disturbing.


http://www.audioheritage.org/vbulletin/showpost.php?p=62216&postcount=371



By the way Do you watch Doctor Who? Sarah Jane Smith (on the motor bike) says hello and she admires your persistence.

Unfortunaly my only series or soap is here.

I just look one time only the previous in english TV and I recognize your prev pict avatar. (I loved)

----------------------------------------
Remercier donc pour moi cette charmante créature qui me fait penser à une étrange version de harold et maude...

Ma devise est " tant qu'à le faire pourquoi ne pas bien le faire"

d'autre part, Le plaisirs de trouver des réponses valables à des questions de fonds sont la véritable essence de ces forums... et parfois de l'échange et du choc des idées nait la lumière qui nous élève presque au rang des chercheur qui bâtissent pour nous ces merveilles de technologie.

par chez nous on m'appele le totaliste car dans ce type de R & D il n'y pas de place pour les mettons, si ou peut-être.

----------------------------------------

and thanks for this compliments.

:cheers: :cheers:

[B&KMan]

ahhhhhhhhhhh

the life is never easy or simple....


do you remeber of few post I send pict of thiel driver for expose cleary the response motion and phase ????

at this time I have not keep information for you and my little head too full is not realize a phenomenon.

If I inverted polarity and put one driver diect of amp with out network, caps or other electronics the inward motion on impulse signal create same power of low but more power in high...

Maybe my intrumentation is not too higher for this phenomenon but the result is : if you plug LF in back motion you increase the upper end response....

so this difference phase according of pink noise excitation (in just prev post ) is maybe just result of reaction of inward motion diaghram not in phase reaction....

My feeling is go in this direction because it is strongly weird: the analyser expose good phase... and integration of 2 drivers: this result is a cancellation ???

well, be shure I go deep question next week.


:cheers:

[Zilch]

But this is not explain why 3145 is plug LF in outward motion and all the rest of driver is inward ???Did you build according to Post 124 (Page 9), or according to Posts 170 and 174 (Page 12)? Because, they are possibly NOT the same with respect to polarity. You do not show the driver connection polarities in 170 and 174, so we don't know HOW you connected the drivers.

Also, it's 120 uF and 120 uF, not 16 uF, as shown in MF of 174. I assume you or someone else caught that....

As drawn in 124, the (-) terminal (black) of the LF driver connects to the (-) input terminal, whereas it is the (+) terminals (red) of the three other drivers that connect to the (-) input terminal. Is that how you built and wired it?

The inversion is the same in the 3143 schematic. In both cases, when a positive voltage is applied to the (+) input terminal, the LF driver will move IN, and the others will move out.

Also, as an aside, when voltage drives were prepared analyzing the role and placement of the 52 uF capacitor, the switching out of the first (2.9 mH) inductor in the biamp mode was not taken into account. The 52 uF capacitor is also shorted out in biamp, which is not consistant with its alleged "protective" role.

Yes, I did see the assertion that the factory schematic is possibly wrong.

Factory 3145 schematic bypasses both its 60 uF and 4.8 mH in the biamp mode, as I read it.

Don't smack me on that point; I'll look at it again, and I've only reviewed the first 16 pages of the thread, and it is very late here.... :p

[Ian Mackenzie]

1--- For UHF section
a--- what is " .10mH + 9V " in uhf section ????

A cut and paste vestige - I fixed it.

b--- Because I'm a 2405 (16 ohms) and the DCR is 8 ohms and the 2405H
is 4 ohms DCR Right??
it is necessary to modifie circuit in according impedance change ???

No. All 077's, 2405's and 2405H's are 10 ohm transducers regardless of what is stamped on the foilcal.

2--- For LF section
a--- the 5.4 mH air coil is .63 Ohms....
but I found a air core of .30 Ohms...
It is a better but the difference in ohms it is problem ???

No. Use the 0.30 ohm 5.4 mH coil.

It is necessary to compensate by a resistance in serial ???

No.

b--- Heuu.... Giskard it is possible to other cut and paste error:
In your second listing parts you match 33 mfd and 51 mfd for
according of not 2 X 40 Mfd is not available on solen but this caps
is obligation to match value in opposite (according to DC Charge) ???

The 33 uF and 51 uF in series are closer to 20 uF than the two 39 uF.
Use whichever you want.

3--- For MF section
a--- You write put down mH in coil for 2121 => 1.8 to 1.7 but It is necessary to adap the 2420 too

I'm not going to modify the 3144/3145 circuit for a 2121. It has undesireable breakup modes and I would never use it. Recone a 2121 as a 2122.

4--- For HF section (according to modification) :
a--- 3 resistance in parralle in 62 ohms but one of .42 Ohms in serial with
.68 mH ?? It is value of coil resistance or it really other resistance
for compensation of transfo coil???

5--- The PB is polypropylene caps 400V...
it is better to build with a bypass at low value in paper and foil ???
( I know is realy higher price but I dont care at this state... http://audioheritage.org/vbulletin/images/smilies/biggrin.gif )

I don't understand what you are trying to tell me.

Here is a the clue.

Jean has not reconed the 2121 yet....Harararararararara:blink:

In case this Jean the whole mid filter not going to work as this filter is designed only for the 2122 re cone only...Giskard's response would be.:hate-pc: :crying: :crying: :help:


Ian:uhmmmm:

[Zilch]

Here is a the clue.That wouldn't make the LF transducer move out instead of in, tho.

I'm beginning to think the "Why" is a question intended for and perhaps best addressed by the forum design philosophy division.... :p

[B&KMan]

Did you build according to Post 124 (Page 9), or according to Posts 170 and 174 (Page 12)? Because, they are possibly NOT the same with respect to polarity. You do not show the driver connection polarities in 170 and 174, so we don't know HOW you connected the drivers.


well, according it is my first experience in electronic domain , I have many errors in the path of this quest... this scematic is not final schematic and many error is relevant after see ex: post 308..

finally my electrical response expose all is work in conform to 3145 schema response. the polarity is in regards of post schema of Giskard in post 308 or the original 3145 PDF. according of this the LF is connected in outward motion. and the 3143 is connected in inward motion. !!!

so yes maybe philosophy question here. if really the set-up of 3145 : why JBL put 15 outward motion ??




Also, it's 120 uF and 120 uF, not 16 uF, as shown in MF of 174. I assume you or someone else caught that....



This is error copy & paste and it is corrected in few next posts...





Don't smack me on that point; I'll look at it again, and I've only reviewed the first 16 pages of the thread, and it is very late here.... :p



yes I understand, yesterday I posted this in 4 am (I'm new York time)
:p

[Zilch]

Finally my electrical response expose all is work in conform to 3145 schema response. the polarity is in regards of post schema of Giskard in post 308 or the original 3145 PDF. according of this the LF is connected in outward motion. and the 3143 is connected in inward motion. !!!In post 308, Earl inverted the connection of the LF driver. See Giskard's schematic there. The (-) network INPUT connection goes straight through the network to the (-) output connection to the LF driver, which would be the black driver terminal. Earl shows it going to the red terminal of the driver, which incorrectly puts all of the drivers in phase.

http://audioheritage.org/vbulletin/showthread.php?t=1388&page=21&p=60751.

See the pdf schematics for 3143 and 3145:

http://www.jblproservice.com/pdf/Network%20Schematics/3143%20Network.pdf

http://www.jblproservice.com/pdf/Network%20Schematics/3145%20Network.pdf

In both of these JBL published cases, the black (-) network INPUT terminal connects to the black (-) terminal of the LF driver but to the RED (+) terminals of the other three drivers, as I described above. The wire and driver terminal colors are not shown on the pdf 3145 schematic, but, per JBL convention, (-) output designation goes to black driver terminal.

Earl's post 308 erroneously inverts the LF driver connection, putting it in phase with the other drivers (only MF is shown), and that is likely the source of the difference in driver motion between your factory 3143 and "Bignet" crossovers, assuming your drivers are connected per Earl's drawing in post 308.

In ALL cases, wired per JBL or per Giskard specification, the LF moves IN when (+) voltage is applied to the network input terminal....

[Ian Mackenzie]

Well Dr Watson, or should I say Zilchmeister you have solved another case.

Well done!

Ps. The 2121 must be reconed to a 2122 c/- Giskard.

Ian

[Zilch]

Well Dr Watson, or should I say Zilchmeister you have solved another case.
Jeez, I HOPE so! I'd like to get outta these hip waders.... ;)

[Earl K]

Earl's post 308 erroneously inverts the LF driver connection, putting it in phase with the other drivers (only MF is shown), and that is likely the source of the difference in driver motion between your factory 3143 and "Bignet" crossovers, assuming your drivers are connected per Earl's drawing in post 308.

Oh oh ! I felt my ears burning :blink:

- Sorry for the mess-up guys ! :o:
- Look at all these extra postings I've caused / shameful ! :( ;)
- & I promise that in the future I'll excuse myself from polarity discussions. :bouncy:

[Mr. Widget]

- & I promise that in the future I'll excuse myself from polarity discussions. :bouncy:

Earl, you should leave that topic to Phase Man... aka Bo.:D


Widget

[Earl K]

Earl, you should leave that topic to Phase Man... aka Bo.:D Widget
Yep ! ,,, :yes:

:)

[Ian Mackenzie]

Earl, you should leave that topic to Phase Man... aka Bo.:D


Widget

OL Tin Ear has been rather quiet lately.....I do miss the noise I must say.:wasnt-me:

Ian

[Ian Mackenzie]

So what is happening here NOT.

Jean,

Send me a PM, I have a surprise for you.:bouncy:

Ian

[Earl K]

So what is happening here ,,,,,
,,,,,

Hi Ian,

- Because of what Jean has stated within Post 432 (http://audioheritage.csdco.com/vbulletin/showpost.php?p=65803&postcount=432) / I think he's offline for a while , enjoying a holiday . :)

[B&KMan]

Hi Ian,

- Because of what Jean has stated within Post 432 (http://audioheritage.csdco.com/vbulletin/showpost.php?p=65803&postcount=432) / I think he's offline for a while , enjoying a holiday . :)


Hey hey ,

Your right Earl !!!

but the vacancy is finish :(

and what is a mess posts here... :scold:
( yuk yuk yuk )


-------------------------------------
well seriously,

I analyse, compute, and keep litterature on my vacancy of this funky phase connection question...

1--- JBL expose clearly what is (+) and ( - ) in the schematic but never expose what side connection of driver.

2--- if I suppose inverted polarity convention is applied the black (neg electric network wire is never connected in black post LF 15" because the neg post is (+) post of driver !!!!

many thread expose the convention color of JBL and many time I read
all the time black cable with black post of LF driver ...



WOW WOW WOW


:cooked:

I read the 4343 connection polarity and manys other old speakers vintage and

I just not understand

(maybe Earl is same of me )

but the original 3143 schema, the schema do not expose the connection with drivers color terminal post

=> just color code of wire with out polarity

ok ok ok .

I'm back in 3145. polarity is exposed on original schema

LF = signal + => + (post driver supposed ) ( + = black post in this case )
mF = signal + => - (post driver supposed ) ( - = red post in this case )
hF = signal + => - (post driver supposed ) ( - = red post in this case )
uhF = signal + => - (post driver supposed ) ( - = red post in this case )

=======================

the black code wire is by convention for the black post LF and LF go inward.

but the positif wire is clearly indicate for positif post not color a specific color post terminal driver (black)

=======================

the interpretation of Zilch of the driver polarity is inverted value is delicate because, at the time of the schema is produced, all driver of JBL is negative convention so.................


now with my poor english I feel really ShakespeareMan...


positif is positif or it is not positif :blink:


THIS IS A QUESTION !!



where is my philosophy study :p


1---- the invertion phase in schema 3145 is clear : Lf is inverted of the rest of the drivers. or it is inverse ???

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-


:hyp: the question is LF is move inward or outward. :hyp:


=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-


JBL ressource guys please go in yours email phone and kidnapp the guy who have a information....



2------ the invertion is according of 1 degree of butterworth filter but keep in this mind normally the phase shift is due to "slowly " phase shift....

so the question is slow the Lf or Slow the mid.... :dont-know

my litterature expose the LF is all the time in phase (outward motion and the rest is adapted to filter selection..

:help:


Jean.

[B&KMan]

In both of these JBL published cases, the black (-) network INPUT terminal connects to the black (-) terminal of the LF driver but to the RED (+) terminals of the other three drivers, as I described above. The wire and driver terminal colors are not shown on the pdf 3145 schematic, but, per JBL convention, (-) output designation goes to black driver terminal.



Her ZilchMeister,

where you read this " published cases" ???

I finded in JBL site and audio Heritage and I'm not found " published cases " who JBL explain this convention...

the negative polarity convention explain the black is in reality the ( + ) post !!
(normally it is red = + post)

thanks to put links in my " blind eyes " :D


Jean.

[Zilch]

"These published cases" refers to the 3143 and 3145 schematics linked directly above there.... :)

[B&KMan]

"These published cases" refers to the 3143 and 3145 schematics linked directly above there.... :)

well it too osbcur for me...


I call JBL Pro !!!!



ok dokay this is the final answer of The tech support of JBL pro California.

The vintage driver is wounded inverse motion so if you put ( + ) signal into a ( red ) post the system is inward motion.

all in there have this information... but the red is the (+) post of the driver !!!! AHHHHHHHHHHHHHHHHHH

the schematic convention is the (+) post is RED!!!!!!!


so if you back on schematic it is easy to understant the LF CONNECTION:

+ signal to (+) => RED post !!!

final result of 3145

-----------------------------------------------------

LF IS INWARD MOTION
MF, HF & UHF IS OUTWARD MOTION.....

-----------------------------------------------------

:applaud:

thanks for all


jean.

[Zilch]

The wire and driver terminal colors are not shown on the pdf 3145 schematic, but, per JBL convention, (-) output designation goes to black driver terminal.

In ALL cases, wired per JBL or per Giskard specification, the LF moves IN when (+) voltage is applied to the network input terminal.... :)

[Ian Mackenzie]

Apparently:) .

Just remember what we said about the crossover values being only applicable for specific driver upgrades:D .

[B&KMan]

Hi again...

after 2 smalls weeks for stabilise my speaker with final good phase connection, I perform mesure and fine tune...


no surprise the response is not really change than last posted here...

so I decided to wait build of my second network and stock little money for

probably refresh my 2420 original diaphrams for more recent diaphrams.

I know it is exist a controverse in alu VS titane and so any person have perform upgrade or chance to compare is welcome to post your experience.

:cheers:


Jean.

[B&KMan]

BTW:

I wait with by interest the threat of 4343 to 4344 upgrade and in this time I try to run test and expose at you great members the distortion create by vibration in caps.

next soon...

[B&KMan]

Yes Jean ,

- You should start a new thread about this topic of "micro-vibrations" within capacitors ( or inductors ) . It certainly warrants a lot of scruntiny. You should then see if you can enlist the efforts of a moderator to remove all the posts located in this thread that pertain to this unique topic . You'll need to do the homework and identify the posts that warrant removing and those that may need a duplicate left in place. Once armed with this blueprint you should then pass it on to the moderator. Then maybe, just maybe, this thread could get back on track. ;)

My gripe :

- By consistently locating your micro-vibration" observations in this thread , you're making it almost impossible for others to participate in the original topic ( that of reworking the network for better performance ).

- This thread has now become like "Moses wandering the Sinai desert" / and you should know, that he never did make it into the promised land .

:o:


Thank you still for this Earl follow-up,

You are right well on practically all the table...

In fact my problem is:
how to install problems considered to be by manys like ridiculous
seeing ésotéric in this forum??

and of makes since the question of the vibration was elements major in
the heart of this too long thread, I have lengthily to hesitate to build a new thread simply to expose this 1 post bracket.

It is certain that this thread if it were cleaned of these slags would
be certainements more attractive and easy.

I believe that Mr. Mackensie had clearly to express the idea that is to
condense that and of another prospect in a project and not in a forum.

alas, I made my first weapons with this thread and I practically learned my
English at the same time (sic)

other share, when I look at several thread complex like that of zilch "quick and
dirty..."one also falls into a disproportion.

In fact all the forum is a maze with only those which make go the
search engines finds the information which wishes.

I have personally to pass the hours and the hours to be sought and
liras of the infos which without preliminary research would be like a
sand grain in the world of the fabric of the www.

I your frustration includes/understands, that I also divide...

but which is best the step?? to produce full with small threads and to synthesize them in a final project???


this experience vibration is exposed just the fact why sand or other inertia material is a real benefit on condensers on passive network... in regards of my entire project network.


finally, I more produced the electric proof which missed with my vibratory
approach that one read in this thread...


hummmm manys of deep questions which resound here...

thanks again for cue and I open my pm for any sugestion.

:cheers:

Jean.

[BooBoo Magoo]

My gripe :

- By consistently locating your "micro-vibration" observations in this thread , you're making it almost impossible for others to participate in the original topic ( that of reworking the network for better performance ).

- This thread has now become like "Moses wandering the Sinai desert" / and you should know, that he never did make it into the promised land .

:o:Yeah, I read this thread and ended up parting out my 4343's. :rotfl:

[B&KMan]

- Yes, I feel that topics like "phasing/polarities" , "capacitor types", "inductor types", "wiring type", "crossover isolation ( from vibrations )" are all separate topics and need to be treated as such in separate / smaller threads .

For Example : - Remember when IanMac started a thread on how to align a 4343 with a simple Radio Shack SLM ? Well , you didn't agree with his method and you guys debated back & forth in a separate thread. The value to that is that the rest of us can either read it or not. If we are interested in that one single aspect of getting better performance out of a 4 way we can follow along and try out the ideas presented. That thread in its' own way contributes something useful to those with 4343/4 systems. Thankfully , the two viewpoints are confined to a separate thread.




- Well, I happen find the topic very fascinating and deserving of separate scruntiny within an identifiable thread. The topic may seem esoteric or ridiculas to some, but that's their problem. A new thread with its own name will allow others in the future to access this information and learn from it. Right now with that information scattered throughout this thread , no one will ever find it, even using the search engine .

:)



Well your right again...

but the mess is done, what correctif your sugest ??

restarts a few news threads for recapitulating each thematic ??

:blink:

[Zilch]

Other share, when I look at several thread complex like that of zilch "quick and dirty..."one also falls into a disproportion.I think of it as "comprehensive." :p

I'm nearly done, and putting together a summary/index of it. The key to finding stuff there is to use the search engine within the thread, if that works for you. If "Search" leads me to a large thread, I'm not typically reading the entire thing to find the desired info, nope....


In fact all the forum is a maze with only those which make go the search engines finds the information which wishes.

That's the dilemma, of course. There's so much information here you can spend the rest of your life tracking it down. That's the advantage of the extended threads. If someone's interested in Sub1500, 95% of the info is in the one thread.

Ian's approach is a good one, but it should be implemented in a project forum, with subthreads on each topic. It takes extreme discipline to make even that work, though; previous attempts have been only marginally successful at it.

As Jean points out, it's difficult if not impossible to organize a project "in progress." It goes where it goes, and there is value in the exposition of the process and issues that arise, and advice from others along the way is especially important.

That said, I agree with Earl that a separate synthesis of the crossover vibration issue would be helpful. Link to posts in the existing thread(s) to illustrate the points, as required.