Am I building these correctly - JBL 4355 Crossovers

[dkalsi]

All,

Please forgive me, but I have never built crossovers before (I'm not even sure if I am reading the schematics correctly). I was hoping some one can verify whether I am building the crossovers correctly. I have included below the original schematics posted by Giskard, followed by my drawing of how I plan to lay the various components. Could someone please look this over for me?

Thanks in advance

Original Schematics:
http://www.audioheritage.org/vbulletin/attachment.php?attachmentid=46834&stc=1&d=1280446056

My Diagram:

http://i185.photobucket.com/albums/x205/dkalsi/4345Crossoverpicturev2.jpg

[dkalsi]

anyone?

[jerv]

anyone?

Your drawing looks good to me. (I guess L-pads are placed off the board somewhere else). Consider placing the inductors with 90 degrees orientation between each other to minimize interfecence. For tips on placement of inductors in crossovers, see http://www.troelsgravesen.dk/coils.htm

[dkalsi]

Your drawing looks good to me. (I guess L-pads are placed off the board somewhere else). Consider placing the inductors with 90 degrees orientation between each other to minimize interfecence. For tips on placement of inductors in crossovers, see http://www.troelsgravesen.dk/coils.htm

Thanks Jerv,

Yes - the L-pads have not been included in the diagram. Per the link you provided (thanks!), I moved/oriented the inductors a little bit differently. See revised diagram below. Additionally, I have also included a picture of how I was planning on charge coupling the capacitors - does this also look right?

Thanks Dhar

New Orientation:
http://i185.photobucket.com/albums/x205/dkalsi/4345Crossoverpicturev3-1.jpg

Charge-coupling the capacitors:
http://i185.photobucket.com/albums/x205/dkalsi/CCingCapacitors.jpg

[dkalsi]

142 views and only 1 reply? Could some one please verify whether I have the charged coupled methodology correct, and whether this symbols below represent positive input:

http://i185.photobucket.com/albums/x205/dkalsi/Input.jpg

[Robh3606]

Yes those are the positive feeds. The CC topology looks fine.

Rob:)

[RedCoat23]

I looked because I need to know too! :)

[dkalsi]

Yes those are the positive feeds. The CC topology looks fine.

Rob:)

Rob - You are the MAN!!!!!

Thanks - I can finally start building these things.

Cheers,
D

[jerv]

Thanks Jerv,

Yes - the L-pads have not been included in the diagram. Per the link you provided (thanks!), I moved/oriented the inductors a little bit differently. See revised diagram below. Additionally, I have also included a picture of how I was planning on charge coupling the capacitors - does this also look right?



Sorry about not being able to answer earlier - your layout looks good to me, and I also second Rob's comment on the CC topology. Good luck with the building!

Regards, Espen

[dkalsi]

Just realized that I had made an error when I revised the inductor placement. I incorrectly labeled the 18ohm resistor as 8.2 ohm and I had a positive lead coming out from 1.0mH inductor when it should have being coming from the 18 ohm resistor.

See revised diagram below (just in case someone is following this thread).
http://i185.photobucket.com/albums/x205/dkalsi/4345Crossoverpicturev3-2.jpg

[Chas]

Don't forget that your coils should be physically installed be at right angles to each other to minimise mutual coupling.:)

[4313B]

Here's an example of one that I recently completed (on the right in first photo). Nothing fancy since I needed to get the job done asap. The inductors are sufficiently far apart to negate mutual coupling. I've also included a photo of a fancier (external with case crafted by SteveH) version as well as the stock version.

469754697646977

[martin2395]

Sorry to dig up this old thread but I'm now making a plan for my 3155 networks and I'd like to know if what I do is right:

Pic:
http://puu.sh/c6K85/1a4f46d4dd.png

The network will be divided into 3 sections that will sit above each other, you could call it 'pseudo-modular'.
I still have to figure what happens after S6 and S8 (sorry, total noob in electronics! :) ) and then draw those parts.
The resistors are 10W Superesl caps are Jantzen Crosscaps and the coils are 14AWG Wax Coils, also from Jantzen.

By the way - what's the trick with 2.4 and .16 coild? I can't find them anywere in Europe...

[Odd]

This is easier to understand.

[macaroonie]

Sorry to dig up this old thread but I'm now making a plan for my 3155 networks and I'd like to know if what I do is right:

Pic:
http://puu.sh/c6K85/1a4f46d4dd.png

The network will be divided into 3 sections that will sit above each other, you could call it 'pseudo-modular'.
I still have to figure what happens after S6 and S8 (sorry, total noob in electronics! :) ) and then draw those parts.
The resistors are 10W Superesl caps are Jantzen Crosscaps and the coils are 14AWG Wax Coils, also from Jantzen.

By the way - what's the trick with 2.4 and .16 coild? I can't find them anywere in Europe...

This guy will wind them to order. Reasonable prices and in my experience quite fast turnaround

http://www.audio-components.co.uk/store/SubCategory.asp?SubCategoryID=6

[martin2395]

Thanks for the advice macaroonie!

Odd, the schematics you posted has more components, why is that? L4 and R16 for the mid and for L5/R6 HF.
They aren't noted on the layout posted in 2010 and on Giskard's 3155.

63375

I now also know that the drawing i posted before is wrong - the caps are all in series where some should be (2x paralell) + (2x paralell) in series.

[Odd]

They will flatten the impedance.
Giskard have explained this in a post, can not find it now.

[macaroonie]

Martin you should check whether the inductors should be wound to as low a DCR as possible or whether there should be an R value specified.
Some networks use the DCR as a component of the network and the wind intentionally is (6ohm) kept at a value that does influence the filter.

The guy at the link I sent you can do it whatever way you need , is very helpful also.
Tell him I sent you.

Mac

[martin2395]

AFAIK the DCR value of the original 3155 coils is unknown and in another thread here on LH someone said to use coils that have DCR not higher than 1/20th of the driver's impedance.:dont-know:

From the specsheet of the original 3155: http://puu.sh/catfk/1b4790e8d9.png

[Odd]

AFAIK the DCR value of the original 3155 coils is unknown and in another thread here on LH someone said to use coils that have DCR not higher than 1/20th of the driver's impedance.:dont-know:

From the specsheet of the original 3155: http://puu.sh/catfk/1b4790e8d9.png



I do not know, but I think "4313B" have the answer.

[martin2395]

I've sent him a PM.

[martin2395]

Ok, i asked some other people and been told to keep the DCR as low as possible. The drawing I posted before is wrong, I've reworked it on paper and what do you think, any errors?

The pics aren't resized so I post them as links:

MF (fixed an error on the go ;) ): http://nl.tinypic.com/r/2yk04mc/8
HF http://nl.tinypic.com/r/dy4mxh/8
UHF http://nl.tinypic.com/r/i4nzpw/8

[4313B]

Ok, i asked some other people and been told to keep the DCR as low as possible.G.T. usually designed around 0.5 ohms for DCR so figure around 0.4 to 0.6.

If you go higher or lower you will affect the slopes to some degree.

Just don't use ultra-low DCR coils like some of the Erse designs. They're great for use on the LF transducers or when designed into the network from the start but they should probably be avoided in legacy JBL bandpass filter designs. JBL did use some coils with 7 to 8 ohm DCR in some systems. The 4355 wasn't one of them.

This is easier to understand.Yes, I like that. Looking at my originally posted schematic, it made sense, to me (and within the context of LEAP), at the time. :p

[martin2395]

This makes it a bit tricky then, the 14AWG 2.4mH coil for the 2123H/2202H has around 0.36rdc.
The 1.0mH Jantzen 16AWG CrossCoil has about 0.34 and the cheaper 16AWG standard wire coils have even lower rdc.

[4313B]

Well, I guess I wouldn't sweat it too much because you can always go back and add in little resistors in series with the inductors if it is really necessary.
G.T. himself has added in a mere 0.2 ohms in some instances simply to get a certain circuit "spot-on".
I do know that I used 16 AWG and 18 AWG inductors rather than 14 AWG to get the "correct" DCR values in some networks.
I have a veritable ton of these parts in boxes, I should dig them out someday and see what AWG values they are.

And .36 DCR is really .4 DCR if you round up ;)

[martin2395]

I have a strange feeling, like I just learned something :D

What do you think about my drawings, are they correct? Never done it before but some day you have to start :o:

[macaroonie]

JBL did use some coils with 7 to 8 ohm DCR in some systems. The 4355 wasn't one of them.

Ah , Clarity. Thanks 4313b , that's what I was getting at.

carry on ....

[4313B]

JBL did use some coils with 7 to 8 ohm DCR in some systems. The 4355 wasn't one of them.

Ah , Clarity. Thanks 4313b , that's what I was getting at.

carry on ....It is a very good point!

I noticed with the beginning of the K2 type systems, dating back to the 90's, that G.T. used an obviously complimentary combination of coil DCR and series resistors to get the various notch and conjugate filters just the way he wanted them. When modeling the various networks in LEAP it became obvious that the coil DCR values really did matter in those particular circuits. Using "ideal" coils with no DCR resulted in some pretty messed up simulations.

Prior to that it was generally "0.4 to 0.6 ohms" in most of the legacy systems. That was before the days of all the notch filters.

In general though, especially with series coils on low frequency transducers, the lower the DCR the better.

Case in point - the dual 10 mH coils (~ 1.2 mH DCR) in the Everest II can be replaced with a very nice Erse 20 mH coil having significantly lower DCR in order to help mitigate the "gain" created by the larger DCR. This will reduce output slightly around 60-80 Hz and increase overall balance.

If one doesn't think that any of this matters one can reference the difference between the L250 and 250Ti. The low frequency performance of the 250Ti was markedly better. Same box, same tuning, same LF transducer. The only think that changed was the low pass circuit on the LF transducer. :) It turns out that the L250 versus 250Ti was a wonderful case study.

Another example that M.G. and I discussed way back in the late 70's with respect to the 124A - because that transducer had such a low Q it was discovered that a little more DCR in the low pass coil could actually improve performance in a passive system. Of course the best solution was to simply forget a passive circuit altogether and drive that transducer direct coupled to an amp using an active filter. :) It did require a completely different enclosure volume/tuning to make it all come together in that case though.

[macaroonie]

It is a very good point!

I noticed with the beginning of the K2 type systems, dating back to the 90's, that G.T. used an obviously complimentary combination of coil DCR and series resistors to get the various notch and conjugate filters just the way he wanted them. When modeling the various networks in LEAP it became obvious that the coil DCR values really did matter in those particular circuits. Using "ideal" coils with no DCR resulted in some pretty messed up simulations.

Prior to that it was generally "0.4 to 0.6 ohms" in most of the legacy systems. That was before the days of all the notch filters.

Thanks. It is probably worth noting that some circuit diagrams will show the coil DCR as what looks like a separate item ( resistor ) This will confuse if the nature of the diagram is not noted.
Poss snafu situation. :(

[4313B]

Thanks. It is probably worth noting that some circuit diagrams will show the coil DCR as what looks like a separate item ( resistor ) This will confuse if the nature of the diagram is not noted.
Poss snafu situation. :(And JBL is really good at leaving that information out of the schematics. :o:

[martin2395]

I will never forgive JBL the polairty madness! :D

By the way, can I simply replace the EQ section (= everything after after S6 and S8) with 8 and 16 ohm L-pads or am I missing something?

[Odd]

I will never forgive JBL the polairty madness! :D

By the way, can I simply replace the EQ section (= everything after after S6 and S8) with 8 and 16 ohm L-pads or am I missing something?

These are the two L-pads.
Replace this with two L-pads shown on the drawing.

63413

[martin2395]

Are you sure? This looks more like a 2-way switch for 2 different, fixed EQ options to me.

[Odd]

Are you sure? This looks more like a 2-way switch for 2 different, fixed EQ options to me.

Yes, use the drawing I posted earlier.

[martin2395]

Yes, use the drawing I posted earlier.

Now it's clear, it confirms that it's just an L-Pad, no hidden catch, no strings attached:D

Odd, do you have any address in Europe for 16 ohm l-pads? I can find 8 Ohm ones without any trouble but no 16's.

By the way, are there any requirements where a passive network has to be placed inside a speaker cabinet?
I was even thinking about simply screwing each section of the network to the back of the front panel of the cabinet.

[Odd]

Now it's clear, it confirms that it's just an L-Pad, no hidden catch, no strings attached:D

Odd, do you have any address in Europe for 16 ohm l-pads? I can find 8 Ohm ones without any trouble but no 16's.

By the way, are there any requirements where a passive network has to be placed inside a speaker cabinet?
I was even thinking about simply screwing each section of the network to the back of the front panel of the cabinet.

16 ohms is a bit difficult.
I buy mostly from US.
I have only found 16 ohms with short Shafts 8/3 "
Parts Express (https://www.parts-express.com/Search.aspx?keyword=l-pads&sitesearch=true)
You can build crossover in a separate box.
Look back to post 12

[martin2395]

I like the external one, like it a lot but I'm concerned if I don't get itchy fingers to tweak it since it's so easily accesible ;)

When I look at the schematics I might aswell replace it with resistors but risk a level mismatch due to room acoustics.
Does the 50% on the schematic say that the l-pad is set at 50% or is it just for reference?

[martin2395]

Guys, I wonder if this is an error or not but I see that on the original 4343 the L-pad for the midwoofer is rated for 16 ohm and the unit itself is 8 ohm, why?

[4313B]

Power handling.

[Odd]

The 50% on the schematic has no meaning.

[martin2395]

Slow but steady, getting there.
I'm still waiting for 2 coils - .16 and 2.4mH.

http://i61.tinypic.com/jzw2kx.jpg

[martin2395]

Major design failure :( While measuring the height and width of the crossover I totally forgot the depth and that the bass reflex pipes are so long on the 4343. It turns out that there is only 5.2cm (2.05 inch) clearance between the upper board and the b-r pipe.

The upper board will house the mid and high section and the bottom one the mid and the 9V CC module so I had to make the total area is a lot bigger than on the 3143 hence the Jantzen Cross Coils are simply massive and the amount of caps is sometimes double (4x1uF in place of 2x2uF for example).

Is it "doable" or maybe should I drill some holes around that place to let air pass through?

http://i60.tinypic.com/vovrc0.jpg

[martin2395]

I have finished 1 crossover but it doesn't sound good at all and I can hear hum when the amp is turned on. :banghead:
Back to the design phase...

[martin2395]

They are done, fixed some errors and they work.

I built my 3155's with Jantzen Cross Coils, Superes resistors, CrossCap caps and silver plated copper wire with teflon insulation. The L-pads are standard 3143 ones but cleaned.

First I wasn't so happy with them but then I found out that the l-pads were wired backwards :)

[Odd]

They are done, fixed some errors and they work.

I built my 3155's with Jantzen Cross Coils, Superes resistors, CrossCap caps and silver plated copper wire with teflon insulation. The L-pads are standard 3143 ones but cleaned.

First I wasn't so happy with them but then I found out that the l-pads were wired backwards :)

Do not drink and solder.:D

[martin2395]

:D
Still I can't really get used to it's sound atm, much more open but still somehow grainy, even with the "2 Radian phragm so it must be something else that causes this.
I decided to give my BSS FDS-360 another major overhaul, it already had it's input opamps replaced from NE5532's with LM4562NA's but now I took the plunge and will replace all of the remaining (LF353) opamps with OPA2134PA. :homer:

Hell, this will be the most expensive FDS360 ever :D
I have all the schematics and see if I can replace some electrolytics with fancy stuff like Nichocon Muse. They are all small values so it won't cost a fortune.

[martin2395]

By the way, has anyone tried bypass caps in JBL networks, for example the 0.01uF Theta caps?

[martin2395]

Just a small question regarding the impedance flattening part (coil+resistor), do they have any influence on sound quality or can I use cheap(er) parts for it?

[ivica]

:D
Still I can't really get used to it's sound atm, much more open but still somehow grainy, even with the "2 Radian phragm so it must be something else that causes this.
I decided to give my BSS FDS-360 another major overhaul, it already had it's input opamps replaced from NE5532's with LM4562NA's but now I took the plunge and will replace all of the remaining (LF353) opamps with OPA2134PA. :homer:

Hell, this will be the most expensive FDS360 ever :D
I have all the schematics and see if I can replace some electrolytics with fancy stuff like Nichocon Muse. They are all small values so it won't cost a fortune.
Hi martin2395,

how do you know what type of IC to replace with other type or brand?

regards
ivica

[martin2395]

I got tips from guys who know more than I do, I also have the complete schematics and documentation for the 360 ;)

Still, I fully tweaked the FDS with OPA's and Nichicon Muse KZ and then I bought the Bryston 10B, it trashed the BSS to no end :D

[.:hb:.]

By the way, has anyone tried bypass caps in JBL networks, for example the 0.01uF Theta caps?
The 3155 is factory equipped with 10nF bypass caps.

[.:hb:.]

Just a small question regarding the impedance flattening part (coil+resistor), do they have any influence on sound quality or can I use cheap(er) parts for it?
They are a part of the filtering network and have no clue that they are only impedance correctors :). If you are into boutique parts for your filter, there's no reason to stop here.

[martin2395]

There is a reason, my budget! :applaud:

Beucase there are no 2.0uF Jantzens I must use 4x1.0uF (2x1 + 2x1) and the cost is getting outrageous.
For two 3155's I'd need 12x 1.0uF, 4x0.33uF and 4x3.30uF - €220 euros including shipping :banghead:
I'm not even looking at the 15 and 18uF one's....

Also I'm really considering stopping with CC because of the price of good caps, if I won't use CC I could get away with a single 1uF cap in place of 4...

[4313B]

For further clarification:

http://www.audioheritage.org/vbulletin/attachment.php?attachmentid=46834&stc=1&d=1280446056

The LEAP schematic is correct. Ignore R10, R11, R12 and R13, R14, R15. They were used to simulate L-Pads at 50% power reduction. A person would use L-Pads instead. 16-ohm L-Pad on the 2441 and 8 ohm L-Pad on the 2405. The coils are standard Solen coils with DCR values lower then 0.5 ohm. The capacitors are Solen PB capacitors, the smaller sized poly caps. If I remember correctly, the PA caps were the larger, higher voltage caps.

Unfortunately this LEAP schematic doesn't show exactly how to bias capacitors, that knowledge is assumed based on many past discussions and posted schematics. One can use either the battery method as shown in many JBL schematics or the diode method as shown in the 4365 schematic.

[martin2395]

In a well dampened room they 'll do without the L-pads but YMMV.

Also - save money for good caps, I'm paying the price for buying the cheap crosscaps.
Something like Mundorf S/G/O or maybe some soviet-era PIO's should be miles better.

[martin2395]

Very, very big thanks to member JBLPRO for noticing and letting me know that my hand drawn schematic of the 2441 HF section has a major flaw - the .47mH coil should we wired in series, not in parallel :banghead:

Now I have to fix both networks :D

[JBLPRO]

In a well dampened room they 'll do without the L-pads but YMMV.

Also - save money for good caps, I'm paying the price for buying the cheap crosscaps.
Something like Mundorf S/G/O or maybe some soviet-era PIO's should be miles better.

What caps did you use? I ordered Solen's all the way around but I am not at the listening stage yet. G Timbers recommends Solen's so i figured that would be a good choice.

mk

[martin2395]

I started with Jantzen crosscap (apparantly Tony Gee rates these higher than all the SCR's) but they were disappointing - harsh and grainy.
Now I have the white Audyn Plus combined with Mundorf Supreme, only the 2123H is still running on 4x18uF Crosscaps because of cost and the sheer size of the Audyn caps.
I'm also considering adding MKP1837 on those Crossaps to see if they make any difference on the midwoofer.

Next step will be the resistors, I will be replacing the green Jantzens with Mundorf Mresist supremes.

[martin2395]

Still happy with the 3155's. I changed the HF amp to a tube OTL and never been happier. I've heard many systems so far but I'd never trade my JBL's for anything else, expect maybe the DD series or some high end custom TAD's.

However I still don't understand the need for a highpass for the 3155 (the need for lowpass for 2235H is obvious as it's directly coupled to the amp) as it makes you use an active crossover and a passive one behind it.:dont-know: or am I doing it wrong the whole time?

[Dr.db]

This is easier to understand.





63375



Is the overall resistance of the coil and resistor 7,5ohms together?
Or do I have to use a 7,5ohms resistor in addition to the coil? If so, what´s the resistance of these coils?