Info on 1974 API VCA's

[TubeMonkey]

Hey Everyone,

Looking for some information on mid 70's API VCA's. These are all out of my 1973 API 3224. I've removed all VCA's in my desk but wish to utilize 4 of them in a quad GSSL i'm building into the desk.  The part number for these is 2550, and they are similar looking to a 2520 package...  they were fitted onto the 915 routing modules.  I have board schematics which I will post tomorrow, as I need to get the prints re-scanned. If info can't be found... and  someone familiar with VCAs might help me identify the pin out/power requirements I would greatly appreciate it!

Thanks in advance!

EB

 

[JohnRoberts]

This probably doesn't help but legend has it Dave Blackmer (The DB from DBX) invented his VCA while working at API and they declined to commercialize it... He went on to form dbx corp.

JR 

 

[TubeMonkey]

Hey John,

That's actually very interesting. These would probably have been developed in 1972 or 1973. My desk left the factory in January 1974, and there are only a handful of other similar 3224 desks with the VCAs in them.  If what your saying is true these may have been designed by Dave Blackmer,  there is almost no information on them... and it seems API quickly discarded this design. I'm attaching a picture and schematic... hopefully someone more familiar with VCAs can help me identify the pin out maybe through MM measurement??

 

[TubeMonkey]

And partial schematic

 

[TubeMonkey]

Underside:

 

[JohnRoberts]

You may want to look for very early dbx VCA schematics. Unlikely to use same physical layout, but design may be similar.

JR

 

[TubeMonkey]

Thanks John,

Just realized I can reverse engineer the pinout from the PCB layout.  Still would like to know the functionality of X, Y, and Z on the schematic though....  open to any insight.

what's the purpose of "Symetry" functionality on the DBX 202c? 

EB

 

[PRR]

X Y Z are classic analog computer Multiplier input factors.

However that does not seem to be the classic Philbrick package or pinout. And it looks like it is not wired as a linear Multiplier. I'm guessing some "trim" pins are used to offset the pair and change gain, perhaps log-scale. But that part of the plan is cut-off your scan.

The audio path is clear and IS the same as dBx stuff: 22K in, opamp 22K I/V out.

 

[JohnRoberts]

Thanks John,

Just realized I can reverse engineer the pinout from the PCB layout.  Still would like to know the functionality of X, Y, and Z on the schematic though....  open to any insight.

what's the purpose of "Symetry" functionality on the DBX 202c? 

EB

Symmetry is a distortion trim.

Y looks like an internal operating voltage trim, perhaps for DC offset (guess).

I do not recognize  X and Z from later VCAs, looks like decoupling some internal nodes to reduce HF crosstalk.

If those VCA are not potted, you can reverse engineer the circuit. Not sure why for other than academic (historic) interest. There will be no magic from using very early VCA technology (IMO).  Of course some may have romantic notions about some legacy VCA "sound".

You can probably drop in some modern VCAs with appropriate control law scaling and realize a much cleaner path. Perhaps if one is broken, then you can compare for yourself ( use a null test between channels).

JR 

 

[fazer]

You removed them from your API I assume due to their grainy sound quality.  Why use them in a compressor now?  API quit using them because there were better sounding VCA's

 

[TubeMonkey]

Fazer,  that's a fair point.  I actually don't know what they sound like.  The VCA's had been removed and the archaic automation system bypassed for technical reasons. Although I had heard a few people mention they didn't sound great, I was unclear if they were talking about the 2550's or the Valley People TA101 that were also outfitted in the desk..... so maybe this will only confirm that.... but I guess it won't hurt to pop them in the circuit with a few alligator clips and compare.  I think it's worth a try.

John, the VCA's are unfortunately potted.... maybe i''l attempt a heat gun at some point, but outlook seems dismal from the one i've already attempted to take apart.

 

[fazer]

Hey TubeMonkey,  dont get me wrong.  If I had them I would want to know how they sound.  God knows I keep enough old gear alive.  I have an old Amber Analyzer.  All analog test gear.  I fix it every time I need it to fix something else.  Crazy!

 

[JohnRoberts]

Fazer,  that's a fair point.  I actually don't know what they sound like.  The VCA's had been removed and the archaic automation system bypassed for technical reasons. Although I had heard a few people mention they didn't sound great, I was unclear if they were talking about the 2550's or the Valley People TA101 that were also outfitted in the desk..... so maybe this will only confirm that.... but I guess it won't hurt to pop them in the circuit with a few alligator clips and compare.  I think it's worth a try.

I used the TA101 VCA designed by Paul Buff in a compressor I did back in the '80s. IMO it was SOTA for back then, but since then the new latest THAT corp VCA is vastly superior.

John, the VCA's are unfortunately potted.... maybe i''l attempt a heat gun at some point, but outlook seems dismal from the one i've already attempted to take apart.

Only of historical interest...IMO.  Modern VCAs are much improved, while a lot of that improvement is related to the IC process technology that allows superior device matching and good device performance for the current ratioing task.

JR

PS: For reverse engineering potted modules, a dremel tool with a cutting head can grind away the potting compound but you need to be careful doing it. (Don't ask me how I know this.  ).

 

[Walrus]

This thread galvanised me into action, and I've just traced out the schematic of the API 2551 from a module I de-potted some time back. It doesn't have all the values as some resistors broke and the solvent also removed the colour codes. Transistors 5,6,7 and 8 were in a small metal block to give them thermal similarity.
It looks incredibly similar to the DBX 200 schemo...... 

It shows where the X,Y and Z pins join the circuit, so if one of our esteemed luminaries could have a look and explain, that would be wonderful.

 

[JohnRoberts]

This thread galvanised me into action, and I've just traced out the schematic of the API 2551 from a module I de-potted some time back. It doesn't have all the values as some resistors broke and the solvent also removed the colour codes. Transistors 5,6,7 and 8 were in a small metal block to give them thermal similarity.
It looks incredibly similar to the DBX 200 schemo...... 

Yup pretty conventional... 5,6,7,8 are the actual VCA so they are thermally coupled to reduce thermal related Vbe errors.

Q3 is a zener connected transistor to create a low impedance voltage shunt between the top PNP pair and the bottom NPN pair.

The several devices at the top are creating a current source. That current flows from each top PNP through their collectors to the bottom NPNs "and" through the shunt zener in parallel.

node X shows the common base connection between the input side NPN and output side PNP.

Node Y and Z are effectively the EC or control voltage ports. The DC voltage of these nodes relative to X determines gain or attenuation.

The node you have labelled -ve does not look correct.  I would expect a resistor from that node to -V "and" another resistor from the output of the VCA external op amp to that node that responds to audio signals.

At unity gain and equilibrium X, Y, and Z are all the same voltage and current down the left pair is the same as down the right pair.  The op amp modulates current into and out of your node -ve to null the audio current coming into the op amp - input with current from the left side NPN/PNP.

Voltage gain or cut comes from shifting the voltage of Y and Z relative to voltage of X.  Only mV of offset voltage here will shift the current ratio between the left pair and the right pair. For VCA gain the ratio forces more current in the output (right side) pair, than the input (left side) pair. Since the op amp is nulling out the left side current with the input resistor (signal) current, the output current is stronger creating voltage gain. Alternately if the left side is made stronger than the right side, voltage attenuation occurs.

It shows where the X,Y and Z pins join the circuit, so if one of our esteemed luminaries could have a look and explain, that would be wonderful.

This is all pretty simple in concept but the devil is in the details . As I mentioned the modern THAT VCA takes advantage of modern IC technology to deliver excellent tracking and matching. The matching between input and output pairs as well as matching  NPN and PNP device characteristics is something that was difficult inside IC VCAs until this latest version.  Of course device noise matters too since that will get injected into the result.

Gary Hebert the VCA guy at THAT corp, sent me a copy of the AES paper he wrote about the new generation VCA a few years back. The paper doesn't discuss VCA basics, but more the advanced esoterica around their improvements.  Basic VCA technology has been around for several decades so relatively old hat and argued out. Back decades ago there were some hot public competitions between different VCA approach advocates. 

Soon to be yet another obsolete analog technology.  Gain change is just a one clock tick multiply inside the digital domain.  8)

JR

 

[Walrus]

Many thanks  JR, I'll read that all again in the morning when I'm fresh.

It's very possible I missed out a resistor on the -ve pin, I redrew that cct a few times from the pcb and tied myself in knots.
It was over a cup of tea that I had the bright idea of looking at the DBX schemos, and realised they were more or less identical, give or take the odd diode!

 

[PRR]

> possible I missed out a resistor on the -ve

Or they had not settled on a "perfect" current for the cell, and let that be set by an external resistor.

Also a secret trap for people who plugged early VCAs into later sockets....

 

[Walrus]

I'll have to remember that one PRR! 

However, in this case it was me forgetting the resisitor.  :
It got broken during the dissection and most of the colour codes were missing, looks like it might be Red Red something something something...

 

[JohnRoberts]

I'll have to remember that one PRR! 

However, in this case it was me forgetting the resisitor.  :
It got broken during the dissection and most of the colour codes were missing, looks like it might be Red Red something something something...

Yup a fixed resistor to -V is needed to approximately cancel the fixed current source dumping current into the top of the VCA.

Then another resistor (probably*** connected to the bottom of the VCA ) closes the NF loop for the op amp to modulate the current coming out of the left pair's collectors to match the audio input current coming into the op amp's virtual earth. 

**** in theory the op amp could control VCA current by modulating the current at the top or the bottom of VCA, bottom seems neater.

JR