Looking for ideas troubleshooting oscillation in console VCA circuit

[kakumei47]

Hi all,

I've been doing some changes to my console (MCI528c), and experimented with swapping out the DBX 202c VCAs with DBX202x and That 2002. I made the appropriate resistor changes per the That datasheets, but I'm getting oscillation issues with 5534 opamps. There was no compensation cap in the original design w/ 5534, but I think these VCAs have higher bandwidth so the problems happens with them. I added a 22pf cap as pictured here, but still get oscillation. This happens with all the types of 5534 I tried, but I get no oscillation with opamps like the OPA 134, or MA332, which both don't need compensation caps.

Any ideas of what to try? I know SSL used the 202x a lot I think with 5534s, but I'm sure the VCA card design is a little different.

In the schematic 2003=5534, and the stock VCA is dbx202c

Including docs here. I can just use the OPA or MA332, but it's bugging me why this is happening.

 

[kakumei47]

VCA card bottom

 

[kakumei47]

VCA card top

 

[kakumei47]

I referred to this doc for modifications needed for the new VCAs:

http://www.thatcorp.com/datashts/dn127.pdf

Copying the circuit for the 202x from that here too

 

[kakumei47]

Here's the SSL implementation for the 202x in the 4000 series. They use the 22pf cap (as I did in the pic, but I still get oscillation that stops if I use the caps that don't need it (ma332, opa134, etc)

I DO notice now that I didn't change the cap in the feedback loop, although I modified the resistors to do the appropriate gain staging. Could the problem be still using a 5pf cap instead of 100pf there? Not sure if that would cause this crazy oscillation...

 

[jensenmann]

Originally the bandwidth limiting with 47k and 5pF sets in at 677kHz. If you lower the resistance the frequency will go up. So it´s quite likely that the feedback network runs into feedforward, hence oscillation.
If you redesign your feedback network to a more realistic frequency (I´ve seen mostly 150-200kHz which sets the beginning of phaseshift  in the audioband to 15-20kHz, that should not really be audible) then everything should be a lot more stable. Both That application note and SSL schematic use app. 200kHz in that spot.
f=1/(2 x Pi x R x C)

 

[JohnRoberts]

#1 the 5532 is unity gain stable while 5534 requires 22pf cap on compensation ports.

#2 the instability "may" be caused by VCA output capacitance causing lag at the opamp - input, so 100pf feedback cap is not just for bandwidth limiting, but to swamp out the input capacitance.

Caveat this is mostly a guess... I have never had a VCA oscillate on me.

JR

 

[kakumei47]

Thanks all! I bet that's the problem. I didn't even realize I forgot to change the 5pf to 100 until I posted here. Will change that and test.

 

[kakumei47]

Confirmed. Changed C1003 (the cap in the feedback) and now they work fine w 5534s. Thanks for the help all!