4343 crossover modifications

[Robh3606]

I have the 3145 networks in a pair of 4344 clones. The networks I have were originally built by Ian Mackenzie and used in his 4345's. They use Holovand Caps and the Tapped Inductor solution Giskard posted. I have them biamped with the passive network up top. That network sounds great! It blends the drivers very well but still gives you a lot of control to voice the speakers to your room. You can also get them very flat on axis with just the network alone. I am not sure if you would require any tweeks with the 2121 vs. 2122 but the rest should be fine,

Rob

[B&KMan]

- I generally DC bias all the caps that I can access within any circuit. I would always chose this appoach over just blending caps .

- Blending is my attempt to try to cook up something that I like by combining different inexpensive dielectrics .
...

I've never heard an inexpensive dry film cap that didn't improve with some DC biasing .

MMMMMMMMMM well inexpensive caps all the way ???

OK I just tried to explain if I good understand this problem..

Th DC charge is created a pre-charge load dc signal. In this the caps is preload and the AC transient is modulate the value. So because the caps is already charge , the signal is more fast transient response because the enrgy to load cap is zero !!! If is that this type of circuit ir realy more fast and you have a better corrolar peak. Of course in RMS is relatively the same but in impulsion , the pre-load charged cap is of course realy fast...

But I'm just a little bit crazy and listen this point...

Of course the time load is one of the big influance of Signature caps. So probably you cancel a big part of signature. But the time response is better if the caps is extra fast in comparison of other... But not more linear in regards of the dc response caps....

So the best is a good cheap blend caps build in DC charge coupled ????

In this the price is same a esoteric caps but more fast and linear with pre-loaded charge and a extra better linear for maybe less money... but little more time for selection match...





I'm to crazy, heretic , barbarian ???

[Ian Mackenzie]

Okay,

Yes its better, I even tried the latter 43444Mk11 design and I got better results with the 3145 design.

First decide if you want to bi amp like Robert.

With the expense of charge coupling the midrange I would weigh up bi amping, particularly if your amp isn't up to driving a tough load.

If not my take on this is to charge couple the 52uf (double the values 104+104 UF..ouch $$ also the other shunt capacitor whatever value that is double it as before ) for the 2121 and bypass the larger caps like Daytons with Theta 0.01s or polystryrenes. Alternatively use a Solen like I have and bypass with 0.1uf Solen film/foil...its not bad.

Use some nice film caps(Sonocaps or North Creek ) and charge couple again for the HF and UHF if you can afford and bypass again (with double the values shown 16+16, 24+24, 3+3, 2+2 UF...ouch again $$ ) otherwise grab some Sonocaps and bypass them with 0.01 uf polystyrenes o Theta's like we did with the prototype project May crossovers.

For clarity base the network on the RC values shown in the 3143 network for the mid passband and the hf and uhf passband circuits Robert posted above for the rest (3145 tapped inductor alternative), wind and measure the split coils as above and volia.

There is a bit less sensitivity with the 2235 (the 3145 is designed around the 2245) but there should be lattitude in the attenuation on the pads for the HF and UHF for this not to be an issue.

If you needa a heads up on other details there are plenty of Dab hands around here to walk you through it. Just pay up the Visa and go for it!

Ian


The other no compromise route is here. Unless you have a turbo'd front end and slik amps forget it!

[Ian Mackenzie]

I can see I am going to have to come out with a Basic Hi End 2 way kit for the 4343 Guru's.

Let me have a cold one and figure it out.

Ian

[4313B]

The DX-1 is also an alternative for bi-amping the JBL Studio Monitors. I too will champion the bi-amp route. I bi-amped my 4343B's and never looked back (went so far as to hardwire the bi-amp mode and cut that switch out of the circuit).

[Ian Mackenzie]

I'm going to play and burn in some Auricaps tomorrow and I will post my impressions.

Ian

[B&KMan]

First decide if you want to bi amp like Robert.

I know the Bi-amp is better and the full 4 amp with electronic crossover is ultimate way...



but Check the bill !!! Platinum VISA
OK reference interconnect 3000$
Cable speakers 5000$
Other amplifier (USED) 5000$
But clean AC power Supply not Sufficient
so other 5000$ (used)
Plus space, Plus system Mechanical decoupling for all extra componend 1000$

Plus if go the total best way 3 other amp and interconnect and cable and regulator.

I'M just dead before Work for this money...

so the realy high end passif crossover IT is crazy??? yes but in sacrifice the better way for run the passif crossover but if I build realy good maybe close astronomical multi-amp amplification...

In other way this crazy big money is cohenrent by the age and quality of the driver and cabinet ???


Well yes a supercrossover, same as Giskard buid for MR Widget is for my point of way the best less onerous way...

The chains is measured by its weakest link...

but for buid a idea of price what your feeling this type of crossover cost ???

For the rest Very thanks Mr Makenzie for hint

[johnaec]

OK reference interconnect 3000$
Cable speakers 5000$

I really think you could save here...

John

[Earl K]

Hi B&K Man,

MMMMMMMMMM well inexpensive caps all the way ???

- to clarify ; I consider the Solen line of Metallized PolyPropylene caps to be inexpensive capacitors .
- Some of my other "cheap" MPP caps are actually a little bit more expensive than the Solens.

OK I just tried to explain if I good understand this problem..

Please Go Ahead ,,,,,,

Th DC charge is created a pre-charge load dc signal. In this the caps is preload and the AC transient is modulate the value. So because the caps is already charge , the signal is more fast transient response because the enrgy to load cap is zero !!! If is that this type of circuit ir realy more fast and you have a better corrolar peak. Of course in RMS is relatively the same but in impulsion , the pre-load charged cap is of course realy fast...

- Actually 2 caps "back to back" are likely going to be slightly slower than a single cap from one perspective ,,, .

Dc Biasing - An Analogy ; prebiasing a capacitor is somewhat similar to taking an amplifier that uses the "B" class of operating mode / and then / biasing all its transitors so the complimentary pairs are never "turning-off" ( as in true "B" mode ) . This bit of constant class "A" conductance (at very small voltages) helps the amp designer minimize the typical "B" modes "crossover distortion".
Read pages 10 to 12 on the K2 S5500 product Sheet to get a much more coherent description of the philisophy behind this design .

But I'm just a little bit crazy and listen this point...

- I'm listening ,,,,,

Of course the time load is one of the big influance of Signature caps. So probably you cancel a big part of signature.

- To some degree this is true. But it's not just the DC biasing that is at play here.
-Putting any type of cap back to back with a duplicate will alter the signature of the single cap type .
-Why ??? Well, there is now much more dielectric involved to get the same capacitance / this increase in plate area has an effect .
- Also , the length of the dieletric versus the width of the dielectric has dramatically changed / this has an effect .
- I'd say, IME , these 2 physical changes are large contributors to the sonic changes that one hears when the "signature" sound of a single cap is altered .
- The one thing that DC biasing is unparalleled in accomplishing is taming resonances inside the cap. Until one gets rid of these resonances, one is usually not even aware of their very existence. So if a cap has a "signature" sound that is partially dominated by resonances / well / DC biasing will alter that signature quite dramatically .

But the time response is better if the caps is extra fast in comparison of other... But not more linear in regards of the dc response caps....

- Some people think "faster is better" and use that as a mantra/dogma . I don't . Sometimes I'm after "slower" .
- I say it depends on what are the goals that must be achieved .
- For instance ; when trying to blend the dynamic signatures of a compression driver with a baffle mounted coned woofer / I always want to dynamically "slow" down the compression driver a bit / to help it better match the dynamic signature of the woofer. Cone midranges and cone woofers are all ( IMO ) quite dynamically lazy when compared to horn drivers. It's a constant struggle to blend well / but capacitor choice can help make the marriage work.
- If I was dealing with an all cone system , my approach would likely get altered. I could easily understand using all "fast caps" on all those "loofing domes" .

So the best is a good cheap blend caps build in DC charge coupled ????

- That's what I know & that's what makes up my universe at this time so I'm very biased ( ha ha ).

- I do know that considering what I just stated about altering the signature sound of "HiEnd Boutique" caps / I'd be really reluctant to take the chance that I could blindly alter them for the better. Each cap type has to be auditioned in pairs before any credible guess can even be articulated.
- Right now that pursuit is just too rich for my blood .

In this the price is same a esoteric caps but more fast and linear with pre-loaded charge and a extra better linear for maybe less money... but little more time for selection match...

- actually " maybe a lot more time in the selection match " . Caps also need to complement the rest of any audio system . You were recently talking about "silver cable" and different inductor types being "warmer or brighter" . Well , I'd venture that if you purchased all those types that were "bright" and fast / then the last thing you'd want would be an abundance of Solen caps in your system. I'd venture you would "need" Paper in Oil or MPP in Oil or anything that is noted for warmth / just to offset the choices made in those other areas. It's all "Yin & Yang" as far as I'm concerned .

I'm to crazy, heretic , barbarian ???

Who ? You or Me ?

[Robh3606]

Go Biamp first then decide what you want to do. The biamp is worth it by itself. I use a M552 now and have a 5235 in the wings if I get lucky and the cards show up. I run a full active with different drivers and the biamp set-up with the 4344's and the only difference I can see,not counting the horns used and I am not really sure if this is why, is the full active has better "jump factor" to steal a phrase from Stereophille. Just seems a bit quicker for lack of a better word and has more headroom built in. Maybe a bit more clarity? It could be the drivers. I am also not using any real "High End" equipment at all. Just good basic stuff all studio quality from say the 80's-90's. Got most of it used. Can't see getting a second mortgage to "relax" in front of.

Rob

[B&KMan]

WOW !!!

realy articulated explaination and many serious background experimentation is relevant...


Thanks VERY MUCH.

I take a couple of time to integrate information...

My first problem is: I have not formation in electronic so " my caps is fully charged"

yuk yuk yuk

And my english is really NEW second language

[B&KMan]

Jean,



- From my perspective ( & specifically with capacitors ) ,,, the only worthwhile measurement is what your ears will tell you . The rest is largely bunk and/or marketing jargon wrapped up in some quasi scientific techno babble.

Well I'M not specialist on the caps but for my long experience in audio YES many speech is a voodoo information who many grand is insert in reality.

ex " pur Copper" what is this !!!

99.99 % is pure
other company is 99,9999 is ultra pure and
99,999999 is just pure in other company.

The decimal number is called N purety. so is real good pure copper is minimum 4 N and eventually 6 or 7 N that is a real pure copper but astronoshing price.

the Goertz is if my memory is good 6N

But the silver is realy more fast metal so 4 N is better of 6 N copper.. AH that is real information but many many ass(...) cie is abuse to label pure and discredit the reality associated.

Exactly in same story of : oil vegetable is free cholesterol !!!

completely crazy !! all vegetable do not contains a animal fat so no cholesterol in any vegetable!!!

I'M shure I exist a difference but the level of precision is exponential price value. the RTA type 0 precision laboratory is sale for around 40.000 and mic fover a 2-3 thousand with preamp... just a fu(..) precision tone calibration is sale for over 2000$ so it is a fraud ??? No !!!

manys is called but little is receive

and at this jungle many many jumps into a grey zone for abuse the client and charge overprice...



But You righ in other way too the design is more critic by each grade of componend.

And many upgrade design is push obsolete a sofisticated old solution. But is not always for the best...

Old condenser ic is very more price but actually is represente the best type of mic... More dynamic amplitude , less coloration, wide band, etc. but really more price...

- Regarding your observations about the "electrical" response shapes from the testing of your 3143 network : any electrical response study only tells a partial story / you'll now need to look at the "net" acoustical responses obtained by the actual drivers with those filters attached.
- Everything you've posted above is really quite interesting ( especially the 2420 circuit ) / but taken alone, those electrical responses don't conclude much of anything .
- an "in room" RTA for each of the 4 sections would be a good next step .

Well, You have Absolutely right again, the electrical response is absolutely compred with the final acoustic response in near field (the near field is .5 -1 inch of the front end driver or little more long for compression driver.

And I have the result so by your interest and interest of this forum I send pict of mic response . ( I just hesitate because many pict is maybe to tired the discussion)

So sorry for guy is look this pict with too sterile technical approach.


Jean

[B&KMan]

Hi again !!

Heres is blend of fft and DFT mesured in pink and white noise FRF H1 response
Please chack the resolution in DBand scale X , Y for right comprehension.

first I mesure acoustic near field

the port event
2231A
2121A
2420
2405 (this UHF the response is bad and I verified the diaphragme and it scrap... So I order a now set for matching response as possible)

I hope anybody I analyse the electrical sygnal posted on page 1 and the result on this page... interesting correlation...

I back soon with more tecnical tips for rebuilt network 3145 for the 4343 spec driver...

jean

[B&KMan]

I selected (in white) the frequecy range of driver programme network response

any big horizontal is 5 DB scale in according 40 DB full scale



Jean.

[B&KMan]

The other no compromise route is here. Unless you have a turbo'd front end and slik amps forget it!

WOW I recognize this is a High-end jod... I want !!!!


for my turbo'd front end I feel just colibri !!!

and for a new old worlds I have a SCD 1 OF Sony.
and accuphase electronic and power conditionner...
All Van den hull silver cable...
I hope I'M qualified for a serious job yuk yuk