LA-4 Help Thread!

[helterbelter]

I see you still didn't put those tantalums in.

Damn, you're right ! I've totally forgotten that. Well, that's what happens if you put it aside for such a long time, haha !
I think I've got a bag with military grade tants...... somewhere. I'd better start searching, ha !

Anyway, thanks for the kind words regarding the frontpanel. And I wish you good luck with your fiendish LA4 channel !

 

[Rybow]

I see you still didn't put those tantalums in.

Damn, you're right ! I've totally forgotten that. Well, that's what happens if you put it aside for such a long time, haha !
I think I've got a bag with military grade tants...... somewhere. I'd better start searching, ha !

Anyway, thanks for the kind words regarding the frontpanel. And I wish you good luck with your fiendish LA4 channel !

Its all good man. I didn't use the tants in my current channels. I already had lytics, and I remember you saying they aren't in the audio path anyways. The blue LED's in your meters are very nice! I have to do that on my next project.

I am having one small issue with the fiend channel, and I am having an issue pinning it down. Its in the meter 0 adjust circuit. My GR meter tracking trimmer is affecting the 0 adjust, and the 0 adjust doesn't have enough range to bring the meter to 0. My brain is screaming short, but I can't seem to find it. Component values in the meter circuit are correct, the meter itself is fine, and my other channel does not do this. If anyone has any clues to what the cause could be, I would be very grateful.

Hope to post a pic soon. I redid the guts to put the boards side by side instead of stacked. Much better for me this way.

Anyways, back at it!

 

[helterbelter]

Hi Insomniaclown,

Well, in case I have said the tants aren't in the audiopath, then I was totally wrong. All 3 tants are in the audiopath, acting as DC blockers.

Regarding your faulty channel... (I'm just thinking out loud...)
-Have you checked the meter with the other channel ?
-Are the R50 and R52 connected to the voltage rails ? (maybe a link missing ?)
There aren't many components that affect the GR metering. It could be a faulty R50 or 52 (And by that I'm referring to a resistor with a faulty printed colorcode), the trimmer R51, the LDR ?

 

[herrmann]

Hi
Just finished a dual 1u version, I have the hairball 11148 for the output, and I just want to be sure of the wiring. I know it requires a DC blocking cap.

 

[herrmann]

I'm trying different wiring. This is the 11148 specs :

 

[herrmann]

And the responses, comp bypassed. Violet : prim in series / sec in series. Orange : prim in series / sec in parallel
Violet not good but lot of additionnal gain.

 

[herrmann]

Red=prim and sec in series
Blue=prim and sec in parallel
So parallel seems to be the winner ! But I'm really confused about impedance and loads things...

 

[edanderson]

herrmann,

primaries in parallel and secondaries in parallel is how the b11148 was used in the 1176, LA-3A, 1109, etc, so you are safe there.  i haven't looked at the LA-4 output stage in a while, but it should drive the transformer fine in this setup.  other configurations are possible but would require some adjustments to the circuit and loading for flat response.  be sure to use a DC blocking cap of at least 470uf.

ed

 

[Luny Tune]

@helterbelter & Insomniaclown

Here are pictures from original LA-4s. At least one is modded though which can be seen on the ICs. I'd like to comment on the tantalums. They're either replaced or were never there in all but one that I've seen. (I've seen more than these three. ;D) I don't think the tantalums are that big a deal.

And then I'd like to draw your attention to the optos. Check the new and updated pdf. I've found that the NSL32 actually doesn't create the correct LA-4 character and I recommend they be replaced. I don't hope everyone will hate me for this as it really is a learning process and experience for me and I suddenly learn new stuff about it that I didn't know before.

The compensation suggested earlier was either taking the character in a wrong direction (the parallel resistor over R83 which gave even worse times, it turns out) and my own suggestion about replacing R34 with a 2k5 trimmer which is compensating for the wrong LEDs alright but not helping the times to become right.

The solution I've tested and which works perfectly as a replacement for the LED/6910 combo I found to work as the original is a Vactec Vactrol VTL5C4/2, a 150kΩ resistor (R13) and a 10kΩ trimmer as R34. Actually a 180kΩ resistor will bring you closer to the 6910 in my machines with that LED/6910 combo but the 150kΩ gives you the same times as the original I've tested with.

Check the updated pdf for details.

I've tried to compensate this way with the NSL32 but it seems that we need more "workable" times from the opto itself. The NSL32 is too damn fast. The VTL5C4(/2) has times that let us "dial it in" better. There may very well be more compatible optos but we need to mind the times as well as the mentioned compesation.

 

[regularjohn]

Just finished this badboy...

CMOQ-2S output trafos, nsl 32 optos, LME49860ma input opamps.
I did not include the overload LED and it's related zeners, as I heard these reduce the fidelity.
1/2W metal film resistors throughout the whole thing.
I used tants for the 6u8 caps, and added psu decoupling to all the psu lytics on the back of the pcb.
This thing sounds killer!  I just used it to track some guitars and used it at mixdown on drumbuss - it worked fantastic!
I found the way to really get it to sound good was to slam the hell out of it.
Here's the song i used it on: http://www.yousendit.com/download/ZGJlRm85dENrUm14dnc9PQ  
The link is good til OCT 27 (yousendit link). It's a cover of my girlfriend's favorite Queens of the Stone Age song I recorded for her.
All the guitars were tracked with the comp at 8:1 and prob 8-10dB GR.  At mixdown on the drumbuss I had it at 12:1 averaging about 10dB GR while blasting the outputs.

I did find that it works better on stereo sources when not linked.  In Link mode, it created awful compression artifacts.
Also, I think it might be helpful to add more space between the psu filter caps.  The largest I could fit were 1200uF.  I would have preferred 1800 or 2200, but there was not space.

Lunytune...what difference do the other opto's make.  because i felt like once I had calibrated the comp, that the nsl32 sounded fantastic!
The only problem I seem to have is that the meter is not accurate in +4 mode.  It's about 7 or 8 dB too low.  I'm not sure how to fix this as there's no calibration for the +4 metering.

 

[Rybow]

Hey Luny!

In the 2 channels that I have built, my NSL 32 sounds really great. Its fast, but it works well for vocals and electric guitars. Also nice on bass. I have no experience with real LA 4's, but I like this one. I used a 1K trimmer for the ratio adjustment, and it calibrated perfectly. I also paired it with a cinemag CMOQ2S, and it is very clean sounding and transparent until you really lay it on that is, but it still sounds good.

My other channel has problems. I have tried the 6910/LED a few times, but I have yet to make it work for me. No matter what LED I put in there, I get no ratios except for 20:1.  So I moved to 2 x VTL 5C4 on that one. Works better, but  I think I will implement those mods you listed as the 1K trimmer I have in right now won't allow for full ratio calibration. Almost, but not quite. Don't know why I went for 2 x VTL 5C4 instead of the VTL 5C4/2, but it shouldn't be a big deal right?

Nice looking unit regularjohn!

 

[MrZpliff]

I did find that it works better on stereo sources when not linked. 

Hmmm... mine is exactly the same... when I link the channels one of them doesn't track very well but when in dual mono mode, if I set them about the same they track very well. I also use the NSL-32...

Regularjohn: Very nice job, by the way

Insomniaclown: I think you need the VTL 5C4/2 'cause it got two photoresistors driven by the same led... pls, correct me if I'm wrong...
Edit: oh, sorry, I see you used two of them, should work...

 

[dustbro]

I've found that the NSL32 actually doesn't create the correct LA-4 character and I recommend they be replaced.

I actually like the sound of these NSL32's better.  ;D
I think the originals are a bit slow, so these optos hit the spot for me!
Maybe a possible mod in the future? switch between optos for faster/slower attack?

 

[MrZpliff]

I've found that the NSL32 actually doesn't create the correct LA-4 character and I recommend they be replaced.

I actually like the sound of these NSL32's better.  ;D
I think the originals are a bit slow, so these optos hit the spot for me!
Maybe a possible mod in the future? switch between optos for faster/slower attack?

I like it too !

 

[Matthew Jacobs]

Me too, I like the NSL32... although I haven't tried anything else...

 

[ilfungo]

I prefer VTL 5C4/2.....

 

[ilfungo]

THANKS LUNY!!!
with the modification suggested by you (150k and 10k trimpot)
LA4  sound incredible!
Stereo Link also works best
ratio 2 - 4 - 8 are perfect with 5 db of compression,
10 db of compression out of 0.2

ratio from 12 to 20  one db off

I used them as a buss comp on  drum and I liked a lot more of my GSSL (turbo +2 SSC )!!!!!
on voice is fantastic
Tomorrow I will use them on acoustic guitars
THANKS ;D ;D ;D

 

[Rybow]

That settles it! I am putting those mods in place tonight when I get home. Then I'll finally be able to rack this bad boy.

 

[herrmann]

Yeah I'm a french guy who forgot the H in "tHreshold"... Anyway, it has B11148 output transf. Sound great, but this thing is compressing too much, well, the response of the threshold pot is very very sensitive, is this normal ?
I have to made the 150k + 10k trimpot mods, I remember the calibration was a bit difficult.

 

[Luny Tune]

Oh, I too think the circuit sounds really good with the NSL32. I just feel obligated to finish what I started and that's to make sure that we have documentation and the right parts to make a LA-4 clone that sounds exactly like the original LA-4. With the VTL5C4/2 and the necessary mods, it does. Most noticable difference from the NSL32 solution is the release time and the longer attack time gives it that low end punch that I love the original for having. But although it doesn't sound exactly like the original, it certainly works well with the NSL 32 too. It just misses that..."juicyness"... Looks like ilfungo has spotted that same character. Hehehe ;D

Great to see that you love as much as I do, ilfungo!! ;D

And keep in mind that you need to adjust the unit before you can really tell the difference in the sound. It needs to be moved to the correct part of the LDR's range and with the ratio trimmed too high it sounds too agressive.

Good job, Regular John! It looks beautiful.
Let me just think about the meter thing there for a minute... Will write again when I've had a look at the schematic. I have an idea, but let me check..

And great raww look you have on yours, Hermann! ;D

Uuuh, I get horny for some more Cinemag trannies...must...also...test...the...hairball...ones....argh...will...die...a...poor...man....