stock 4345 crossover 3145 schematic

[santashooter]

Hi everone, i am gathering information in order to build a pair of 4345's in the next year (or so) and I have realized that most of my hobby of restoring pieces, is my work to reach their factory setting - not better, not worse.
Before people post answers like "The charged coupled networks are better" I would like to say that the idea for me is (generally with my hobby of vintage Hifi) that if i want to have something "better" i could just buy something else - and leave the old pieces as they are, representations of a time and a way of production that i admire - this is a big part of my hobby, and i hope you acknowledge this.
I am aspiring to build the pair as exact replicas of the originals. that means that everything will be done in the same scale and materials, with the same treatment to parts and enclosures as the original JBL.
That being said I would also like to build two crossover networks and these should as the rest be new productions of the original crossovers - with new quality components of course.
The problem that i seek to answer, exist that in spite of 5 days of searching (on google and this page) i haven't been able to find a readable schematic.
I have seen the schematic on the JBLpro site, but i can't seem to read the values of components, due to low resolution.
I have found several charged coupled schematics, but since i would like to remake the stock network i was wondering if someone could point me to a better-resolution schematic or scan of this?

best regards
Soren from denmark.

[BMWCCA]

Do you intend to re-create the enclosure for the crossover, too, including the clunky multi-pole switch for bi-amp function? As a car restorer, while I admire your purist philosophy, I'd think you'd be better off trying to find a used original pair and restoring those. Unless you're able to actually re-create the plastic housing, decals, labeling, metal mountings, and connectors exactly, I don't see the point. I'm not discouraging you from trying and I'm not even going to discuss the resulting improvement in audio quality available in a modern equivalent since that doesn't seem to be your goal. More power to you if you can actually pull-off what you're suggesting! Have you seen photos of what the original looks like? The schematic would seem to be only one of your worries if the goal is really to re-create the original in appearance as well as function.

My first Google search on the topic netted this: http://www.jblproservice.com/pdf/Network Schematics/3145 Network.pdf
It seems fairly clear to me. Maybe what you found was different?

You might also check out this thread, if you haven't already: http://www.audioheritage.org/vbulletin/showthread.php?13930-JBL-3145-stock-crossover.-A-pictorial

Best of luck and be sure to post your progress with plenty of photos.

[santashooter]

Thank you for you're concern, I have seen the pictorial thread and found it very informative - And yes I intend to build everything as the original, about the plastic housing, it might be the only thing i still need to find a way to deice upon - The decals will be custom-made to my project with another subtext to the foil cal: instead of "James b. Lansing Sound INC..." it will read, with the same font; "Copy by Søren Iversen, Denmark"
the same will happen to tags and decals- everything will be made to replicate the original in its manufacture and it will resemble the 4345 perfectly, but I will not make the speakers being mistaken as an original pair since they are copies made by me.
The problem about buying and why i have chosen not to.
I live in denmark, firstly they are hard to come by over here in their original condition, and they are extremely expensive to have shipped. (maybe not more expensive than the project, but:) I need a project and it is to prove that if you want to, and do the research probably, you can build an exact replica in any way when cost is no object.
It also allows me to make one investment at a time (buy the midrange first, then the woofers... so on)
It is a project i really look forward to start and I do have a little team of project members, which include a professional furniture restoration guy, who is assisting me in applying the veneer, alongside a couple of JBL 43XX, Marantz and JBL amp collectors with a huge resource of documents and speakers I can compare to when trying to match colors and finishes.

best Regards
Soren from Denmark

[santashooter]

about the clunky switch and the enclosure - I haven't yet decided the degree of exactness, in this function - could think of making the enclosure, but designing the cosmetic appearance and arranging of the crossover differently, also adding switch that would be less prone to filth and bad connection than the original switch.
My goal is, that the electrical (meaning the crossover being a passive design) and the cosmetic appearance outward are exact matching the original speaker.
About caps - I intend to buy "quality" caps - meaning factory standards in this field, not High-end caps, since i am quite sure that JBL used good quality Industrial standards in their crossover - not fancy gold oil caps.
I am yet to decide brand and make, but these are my thoughts.

[Robh3606]

You cannot build "stock" networks. You cannot get the tapped Inductors in the schematic. Use a Charge Coupled equivalent for best results

Rob:)

[santashooter]

I am not going to build stock networks, I am going to build a passive clone of the network using the design - this means i will have to source either tapped inductors or find some way to get the same electrical parameters with a substitution inductor.
How i will find them I still don't know, since i can't read the schematics i have found, and thus i haven't got a parts list with the components.:confused:
I will not Build CC networks as they are too far from the original design for me.:o:

If it shows that i can get the exact same electrical properties from a CC network compared to the stock crossover i will build a pair - but otherwise i am aiming for a clone using available parts.

best regards
Soren

[Robh3606]

The Inductors are custom wound for JBL using their drawings. They are NLA. The CC networks are correct equivalents without the Inductors. If you have the CC network schematics that's all you need.

Rob:)

[Mannermusic]

I am not going to build stock networks, I am going to build a passive clone of the network using the design - this means i will have to source either tapped inductors or find some way to get the same electrical parameters with a substitution inductor.
How i will find them I still don't know, since i can't read the schematics i have found, and thus i haven't got a parts list with the components.:confused:
I will not Build CC networks as they are too far from the original design for me.:o:

If it shows that i can get the exact same electrical properties from a CC network compared to the stock crossover i will build a pair - but otherwise i am aiming for a clone using available parts.

best regards
Soren

I would simply remove the double caps and battery wiring from one of the DC biased schematics. i.e. 20+20=10. About as close as you're going to get, I'd guess. I did this with a 3155 crossover and it worked just fine; the total capacitance values are unchanged. Mike

[4313B]

It sounds to me like you have to wait for a pair of the networks to show up on eBay or Audiogon. You might want to spend most of your time searching in Japan since that is where most of them ended up. The 3144 and 3145 are the same except for wire lengths to components and L-Pad foilcals.

Here's the stock network parts list. The tapped autotransformers are set to approximately -4 dB for the 3155 and approximately -8 dB for the 3144/3145 (they all use the same tapped cores).

The shunt resistors in front of the tapped autotransformers attempt to mitigate their rising impedance thereby presenting a flatter load to the preceeding series capacitors.

5.4 mH
20 uF
20 uF
20 uF
20 uF
20 uF
7.5 ohm
10 uF

20 uF
20 uF
20 uF
4.8 mH
1.8 mH
14 uF
5.1 ohm
5.1 ohm
39 ohm
39 ohm
8 ohm L-Pad

8 uF
62 ohm
62 ohm
62 ohm
tapped autotransformer 1.0 mH primary winding 0.143 mH secondary winding
12 uF
0.24 mH
3 uF
8 ohm L-pad
20 ohm

1 uF
39 ohm
39 ohm
tapped autotransformer 0.16 mH primary winding 0.026 mH secondary winding
1.5 uF
8 ohm L-Pad

[4313B]

Here are the drives of the stock network versus charge coupled equivalent:

There is more variation between individual components.

The stock networks sound two-dimensional and dead by comparison. I wouldn't waste my time with them. There seems to be a resurgence in interest in tapped autotransformers for some bizzare reason. I've seen the new "stock networks" costing upwards of $4k a pair, chalk it up to ignorance and personally couldn't care less.

[santashooter]

Here are the drives of the stock network versus charge coupled equivalent:

There is more variation between individual components.

The stock networks sound two-dimensional and dead by comparison. I wouldn't waste my time with them. There seems to be a resurgence in interest in tapped autotransformers for some bizzare reason. I've seen the new "stock networks" costing upwards of $4k a pair, chalk it up to ignorance and personally couldn't care less.
Thanks very much for your post - both of them!
I actually didn't realize the two were so close to each other, I will have to re-think whether i am going to use CC or a stock clone.
If i am able to build a CC network (i have the schematic for that) into the standard sized enclosure I will be happy.
I might consider talking to a friend of mine, who are custom winding coils and hear if he could make exact replicas for the stock network since i still prefer keeping it all original.
But the CC network looks promising, i must admit.
And also thanks for the parts list, that was all i needed - great forum with great people, i can't stress how much i appreciate this resource!
I will make a thread about the project as it goes along, and keep everyone updated, but it is going to take some time and thought - 4345 doesn't come quick!

best regards
Soren Iversen

[santashooter]

The Inductors are custom wound for JBL using their drawings. They are NLA. The CC networks are correct equivalents without the Inductors. If you have the CC network schematics that's all you need.

Rob:)
Thank you - i didn't realize the CC schematics were close equivalents to the stock network.

can anyone really say how long the 9v battery lasts in the networks?

[BMWCCA]

can anyone really say how long the 9v battery lasts in the networks?Three-years so far in mine. I check them on their anniversary and I'm using the "professional" 9V batteries we use in wireless mics.

[ivica]

Thanks very much for your post - both of them!
I actually didn't realize the two were so close to each other, I will have to re-think whether i am going to use CC or a stock clone.
If i am able to build a CC network (i have the schematic for that) into the standard sized enclosure I will be happy.
I might consider talking to a friend of mine, who are custom winding coils and hear if he could make exact replicas for the stock network since i still prefer keeping it all original.
But the CC network looks promising, i must admit.
And also thanks for the parts list, that was all i needed - great forum with great people, i can't stress how much i appreciate this resource!
I will make a thread about the project as it goes along, and keep everyone updated, but it is going to take some time and thought - 4345 doesn't come quick!

best regards
Soren Iversen

concerning the original 3145 and later comments from Mr. G Timber
http://www.audioheritage.org/vbulletin/showthread.php?10490-The-JBL-4345-Club&p=110075&viewfull=1#post110075

"....The second problem is the use of a passive crossover between the top of the woofer and the bottom of the 10". Passive crossovers set to frequencies below about 500 Hz react badly with the motional impedance of the woofer/box combination and give substantial gain around 100 Hz. Gain out of a passive system is generally a bad thing. In the case of the High Pass, we have to work the passive network through a really large motional impedance peak resulting from the 10" fundamental resonance in the sub enclosure. This means that the actual voltage drive that occurs at the terminals of the 10" is less than ideal....."

I have tried to get some improvements in 2245/2122 cross-point. From the simulation I have get that using 'Impedance compensation network, and kind of resonance -peak compensation", upper mentioned "lack" of 3145 can be improved, reducing mentioned "100Hz-peak and resonance interference " better then 3dB, but the complexity of the network has rised.

[santashooter]

I would simply remove the double caps and battery wiring from one of the DC biased schematics. i.e. 20+20=10. About as close as you're going to get, I'd guess. I did this with a 3155 crossover and it worked just fine; the total capacitance values are unchanged. Mike
So you would be able to take the CC networks (as posted numerous places on this forum) and remove the double caps with a single of half value and then you wouldn't need the battery?
how are peoples thoughts about the sound of these.
I would like to go all passive, but if i could achieve the same SPL with a passive network instead of a CC i might consider that.

best regards
Soren

[4313B]

concerning the original 3145 and later comments from Mr. G Timber
http://www.audioheritage.org/vbulletin/showthread.php?10490-The-JBL-4345-Club&p=110075&viewfull=1#post110075

"....The second problem is the use of a passive crossover between the top of the woofer and the bottom of the 10". Passive crossovers set to frequencies below about 500 Hz react badly with the motional impedance of the woofer/box combination and give substantial gain around 100 Hz. Gain out of a passive system is generally a bad thing. In the case of the High Pass, we have to work the passive network through a really large motional impedance peak resulting from the 10" fundamental resonance in the sub enclosure. This means that the actual voltage drive that occurs at the terminals of the 10" is less than ideal....."

I have tried to get some improvements in 2245/2122 cross-point. From the simulation I have get that using 'Impedance compensation network, and kind of resonance -peak compensation", upper mentioned "lack" of 3145 can be improved, reducing mentioned "100Hz-peak and resonance interference " better then 3dB, but the complexity of the network has rised.A system of this caliber should have been bi-amp only like the 4340, 4350 and 4355...

[Mannermusic]

So you would be able to take the CC networks (as posted numerous places on this forum) and remove the double caps with a single of half value and then you wouldn't need the battery?
how are peoples thoughts about the sound of these.
I would like to go all passive, but if i could achieve the same SPL with a passive network instead of a CC i might consider that.

best regards
Soren

Yes, you end up with a circuit essentially the same as the original production circuit sans the tapped coil (thank you Dr. G!). I went this way because I wanted to try some circuit tweaks since my drivers are non-standard. A lot easier to modify without the cc double caps! Surprised how good it sounds as is (old ears!) Mike

[tuyen]

Yes, you end up with a circuit essentially the same as the original production circuit sans the tapped coil (thank you Dr. G!). I went this way because I wanted to try some circuit tweaks since my drivers are non-standard. A lot easier to modify without the cc double caps! Surprised how good it sounds as is (old ears!) Mike


Hi guys,

Please excuse my ignorance, but if not wanting to cc the passive crossover is as simple as removing the dc battery and halving the cap value (which is a good thing?), what is the point of designing the CC version in the first place?

Being quite a beginner in passive crossovers, I'm still trying to get my head around what I should be doing for my '4345 build' too. If possible, I'd prefer not to use batteries.. do want to run bi-amped though.

Cheers,
Tuyen

[4313B]

I'd prefer not to use batteries..Use the diode method instead.

Example:

[Chas]

Use the diode method instead.

A very elegant solution IMHO. :yes:

[tuyen]

what the?! that looks really complicated!? hrmmm.. okay now I'm really confused. lol :(

[4313B]

Let's say you have a 4 uF capacitor in series with an 044 in an L96.

Replace the 4 uF capacitor with two 8 uF capacitors in series.
Between the two 8 uF capacitors in series place a 10k resistor and a 1N4935 diode.
Connect the cathode of the diode to the red input of the network.



what is the point of designing the CC version in the first placeWhat's the point in spending more than a couple hundred bucks on an amp?



JBL has taken biased passive filters to the next level. The technology is employed in the new JBL M2.

[tuyen]

Thanks mate. I sort of get it now.

I will try follow and build the bi-amped CC passive design as outlined in previous threads.

Probably need to first figure out what changes are needed to run the combination of drivers that I have.

JBL 2245H 8ohm, JBL 2122H 8ohm, JBL 2470 with factory tit diaphragms + Radian 1225-16ohm alu drivers + 2307 horn, JBL 2405 with factory (8ohm?) blue jbl diaphragms.


I would like to eventually change (upgrade?) to 2445J with truextent diaphragms (with suitable 2311 horn) when the opportunately arises. Sometime after when I have things running okay.


Cheers,
Tuyen

[just4kinks]

I would like to eventually change (upgrade?) to 2445J with truextent diaphragms (with suitable 2311 horn) when the opportunately arises. Sometime after when I have things running okay.


Tuyen, are your cabs finished? Do they follow the original plans?

A 2445 will never fit in an original 4345. A 2440/2441 is borderline, it contacts the midbass box. If you want a safe bet for a 4" driver, look for a 245x.

[tuyen]

Hi Joe,

Not started the build yet mate. I would be okay distancing the midhigh driver position slightly out to cater for future 244x driver? Not quite sure how big the 244x is.

The reality is that my povo 2245h/2122h drivers aren't using original jbl recones anyway, so I'll never achieve the genuine '4345 sound', thus my thinking was that there is not huge point for me to try keep exact to original 4345 cab plans.. I'm favoring the general driver layout of this diy 4345 (bass ports along side the woofer).

58901

You guys have any thoughts on my approach?

[just4kinks]

2445 is 9 1/4" diameter, 4 5/8" radius.

In the original 4345 design, the horn axis is 3 3/8" from the top of the midbass box, and 4" from the top panel of the 4345 (not including the bracing). So you would probably end up making the entire cabinet taller to make everything fit.

2441 is 7" diameter, only 1/8" (in radius) too large. You could probably make it fit, but your 2311 might not seat properly.

TAD TD4001 is ~6 3/4" diameter. 2450 is ~6 1/2".

When I built my cabs, I redesigned for an 8" diameter driver by sloping the midbass box. 4313B did something similar in his build (I think that might be the picture you posted....?).

http://farm7.staticflickr.com/6213/6377823517_65d03be9f7_b.jpg

As far as baffle modification goes, my $.02 is to stick with the original. Baffle layout is more than cosmetic, and with a time-tested design like the 4345 it's a lot easier to screw it up than to improve upon it.

[tuyen]

Seems like I've hijacked the thread, so I might as well keep going.

Thanks Joe. Yep, that is a photo of 4313B's '4345 diy' project. I'll attempt to do the same thing when building the enclosure for the 2122H driver.

A change to the 2441 seems okay if I was factor in the slope.


Taken your advice on-board. Will stick to original baffle layout!


Back to the topic of crossovers for 4345s, any clues where I can find suitable l-pads needed for the bi-amped cc xo? finding the right caps, resistors and coils is easy enough.

Cheers guys,
Tuyen

[ivica]

Use the diode method instead.

Example:

I would suggest for D1, D2 some lower voltage drop diode such as MBR160, even there would be about 0.2V to 0.3V minimal signal amplitude to be applied to produced polarization voltage. Such level with 93dB/1W/1m sensitivity speakers would produce about 70dB SPL.

Regards
Ivica

[Chas]

L Pad source:

http://www.parts-express.com/

[Earl K]

I would suggest for D1, D2 some lower voltage drop diode such as MBR160, even there would be about 0.2V to 0.3V minimal signal amplitude to be applied to produced polarization voltage. Such level with 93dB/1W/1m sensitivity speakers would produce about 70dB SPL.

Regards
Ivica

An Intriguing Recommendation Ivica !

The Schottky diode (http://en.wikipedia.org/wiki/Schottky_diode) (in some respects ) is superior to the standard silicon-junction based types ( ie; most importantly in voltage drop & recovery speed ) .

Unfortunately, it's specs are inferior in the following areas ( the reverse leakage current spec might be the deal breaker for me ) ;


Limitations

The most evident limitations of Schottky diodes are the relatively low reverse voltage ratings for silicon-metal Schottky diodes, typically 50 V and below, and a relatively high reverse leakage current (http://en.wikipedia.org/wiki/Reverse_leakage_current). Some higher-voltage designs are available; 200V is considered a high reverse voltage. Reverse leakage current, because it increases with temperature, leads to a thermal instability (http://en.wikipedia.org/wiki/Thermal_runaway) issue. This often limits the useful reverse voltage to well below the actual rating.
While higher reverse voltages are achievable, they would be accompanied by higher forward voltage drops, comparable to other types; such a Schottky diode would have no advantage [3] (http://en.wikipedia.org/wiki/Schottky_diode#cite_note-3) unless great switching speed is required.

- Apparently I'm one of the few ( on this forum anyways ) that can hear the deleterious effects caused by the battery within the bias circuit ( it drove me nuts I must say / sounding like someone had lightly touched the diaphragm with a "phantom finger" ) . I ended up disconnecting the battery after the caps where fully "formed" ( & then would let the caps slowly float/deflate as the AC beat up on the DC, before eventually rebiasing ) .

- I'll need to try both ( types of diodes ) to see if there's an apparent sonic difference .

Thanks for the suggestion !

:)

[ivica]

An Intriguing Recommendation Ivica !

The Schottky diode (http://en.wikipedia.org/wiki/Schottky_diode) (in some respects ) is superior to the standard silicon-junction based types ( ie; most importantly in voltage drop & recovery speed ) .

Unfortunately, it's specs are inferior in the following areas ( the reverse leakage current spec might be the deal breaker for me ) ;



- Apparently I'm one of the few ( on this forum anyways ) that can hear the deleterious effects caused by the battery within the bias circuit ( it drove me nuts I must say / sounding like someone had lightly touched the diaphragm with a "phantom finger" ) . I ended up disconnecting the battery after the caps where fully "formed" ( & then would let the caps slowly float/deflate as the AC beat up on the DC, before eventually rebiasing ) .

- I'll need to try both ( types of diodes ) to see if there's an apparent sonic difference .

Thanks for the suggestion !

:)


Hi Earl K,


It would be nice to inform us about your experiments with the diodes types, and battery too.

Here diode current is limited by the large serial resistor ( over 10 kOhms), but the problem is that in the most of the time signal amplitude is small because of high efficiency speakers ( 43xx, 44xx SPL is over 93 dB/1W/1m ) so only several Volts are present. Some time amplifier 'cross-over distortion' (owing to small signal present) can be present.



Regards

Ivica