4343 crossover modifications

[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...

[B&KMan]

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

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

Sorry for dust ....

[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 ( 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 ) 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 ( 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/vbullet...=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...




Jean.

[B&KMan]

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

[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

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


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 .

(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 .

(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 )


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.