Jim Williams 4K VCA Caps Mod

[marcmozart]

I've collected tons of information over the years on all things SSL 4000. Gonna post some stuff that I think is worth discussing.

Audiotech Guru Jim Williams posted this on another forum:

"Jim Williams:

Run that AP with a frequency/phase plot. See that phase lead shift in the low end? Not too pretty, is it? That's a result of simple 2pi /rc/sq of selecting too small of a coupling cap. A 10 uf in front of the VCA is one sore spot with me. I would increase that a decade and call it a day. Remember, each coupling cap stage is a 1 pole high pass filter. A few stages, no problems, run hundreds and you will find a 10 or 12 pole filter with rather poor phase response and some non linear phase shift from the fact that all the roll-offs are not tuned to each other but are set somewhat randomly.

Fix them and then run that phase plot and see 1~2 degrees at 20 hz. Don't forget the polyprop film bypasses across them. A nicely fixed up 4k will shake your world if that low end is fixed.

Warning! You will find the high pass filter in regular use afterwards. A couple of nice poly film caps there will allow it to become your friend.

Greg:

Thanks 84K and specially Jim,
Sometime the answer is just dead ahead and we forget the simples rules...
You are definitely right when you are talking about the high pass filter effect and the phase shift!
Especially for us because we always use the VCA automation!

Have you any recommendation for the new cap? (value and brand)

Could you please explain me your warning attention because i'm not sure too really understood what you are talking about...

Do you have also another tricks in your pocket?
I've saw that you are talking about the change of some transistors in the master section in another post of this forum. Could you please also confirm this?

I give a quick look to the schematic and just to confirm your advise, the 10 uF cap (C9) is located after the PFL/AFL switch and before the R37 10K on the VCA card, is it the correct one?

And again, what about the transistor and the voltage feedback amplifier in the master section?

Jim Williams:

Yes, change that 10 uf cap . It should be a brown Roederstein bipolar, a really dark sounding cap. I would first measure the dc on both ends first and then replace it with a 100 uf Panaonic FM 25 v with the + pin facing the more positive side of the circuit. Next, slip on a Wima MKP-2 .01 uf 250 v polyprop film cap on the back of the pcb for the air.

For the master,replace the LM394 super matched transistor pair with a AD MAT 02, get the premium can package, it's about $15 from digi key. These have a lower frequency of hiss and have more top end details than the rather dark sounding National part. ( This also appiles to mic pre's, 990 opamps, MCI consoles and tape machines, Ashley SC-50 comps and other circuits that use the LM394CH).
For the opamp, the National LM6171 works pretty good there if you like bandwidth (100 mhz), speed (3600 v/us slew rate) and low distortion.

Greg:

Just to clarify your thought, if i had really understood i need to mount this two caps in a parallel design, am i right?

Jim Williams:

Bypassing caps is just laying another on top, those thin Wimas usually fit across the electrolytic and still clear the slot it fits into.

Jim Williams on another occasion:
http://recforums.prosoundweb.com/index.php/m/381986/0/
The largest improvement I have heard is on the master mix card. Replace the National LM394 transistor pair with an AD MAT-02. The 5534 opamp needs to go, I had great results with a National LM6171 at 3600v/us slew rate and 100 mhz gain bandwidth. The Roederstein bipolar caps are too soft sounding, replace with low impedance polarized caps once you determine the correct polatity via measuring offsets first. If you replace the 5534's with LME49710's you can sometimes remove the output coupling cap as the offsets are very low. That also applies to the 5532's, use either LME49720 or LME49860.
The quad tlo74 opamps on the monitor card should also go, use either a Linear Tech LT1359 or a National LME49740. Increase the size of the output coupling caps, especially the small 10 uf Roederstein's used in front of the VCA's, that's a low end phase shifter."

[Showcase]

Has anyone tried these mods?

Phase issues at 20 Hz, is that a problem to anyone?

[Matt Sartori]

Has anyone tried these mods?

Phase issues at 20 Hz, is that a problem to anyone?

well I have done some really quick tests on a 4064 I am working on using my neutrik A2 but I have not noticed much difference.
i have been discussing this with some people and there are mixed feeling about this. certanly there is a TINY improvement, does it then build up on a full scale mix, meaning the more channels you add with phase correlation correct the beter your bottom end?

Mattia.

[marcmozart]

Note: the replacement of the LM394 transistor pair with an AD MAT-02 in the 026 card is a change done by SSL (it's in the G+ schematic), probably between 1990 and 1992. The 82E291 Channel Amplifier Cards of my 1991 G-Series had also the AD MAT-02s instead of the LM392, though I can't say for sure if this has been done by the factory.

Have a look at your G-Series console. Quite obvious on the 82E291 Amp Card, as the "can package" (instead of an IC) of the AD MAT-02 has been used and inserted using an adapter (on my console).

Jim Williams on another occasion:
http://recforums.prosoundweb.com/index.php/m/381986/0/
The largest improvement I have heard is on the master mix card. Replace the National LM394 transistor pair with an AD MAT-02.

[brewery]

marc, i think you might be confusing a few things here.
jim is talking about the 26 card in the center section. there are two LM395, and afaik even the latest G+ consoles still had them. they certainly are in the G+ schematic that is available here.
also, like mattia said, the improvements are a lot smaller than you might think, and it's all part of the SSL sound, whether that looks good on paper or not.

all g series line amps that i encountered have MAT02s, but the schematics have a LM394.

[marcmozart]

Thanks for confirming that the G-Series line amps have MAT02s.

Also, the schematic that is available here has two versions of the 26 card, the later of which (from 13 Aug 90 on) also has the MAT02H instead of the LM394H.

marc, i think you might be confusing a few things here.
jim is talking about the 26 card in the center section. there are two LM395, and afaik even the latest G+ consoles still had them. they certainly are in the G+ schematic that is available here.
also, like mattia said, the improvements are a lot smaller than you might think, and it's all part of the SSL sound, whether that looks good on paper or not.

all g series line amps that i encountered have MAT02s, but the schematics have a LM394.

[Dan Korneff]

Might give these a try

[jwhmca]

I have done the VCA cap mod.... an absolute. Most for sure brings the bass in!

[marcmozart]

Did you also put the Wima MKP-2 on the back of the PCB? Did you do measurements?

I have done the VCA cap mod.... an absolute. Most for sure brings the bass in!