Gates SA-39B (M3529B) not compressing (FIXED)

[arthurdb]

--------------
FINAL EDIT: Got it fixed by replacing the 6H6 tube that rectifies the side-chain signal. This was what I initially suspected but turns out the first replacement tube was a dud which needed a much higher AC signal to start developing a DC voltage. Its behavior in circuit was similar to the original one so I though it might have been something else....
--------------
Hello folks.

I've been given a SA-39B compressor for repair and I can't find out what's wrong.
It seems to be the same as the one at the beginning of this thread and working off the schematic linked in this post (attached below)

Power supply voltages are more or less in line with the values on the schematic as well as signal AC values. The neon bulb was at a much higher voltage (70VDC instead of 55VDC) but I still could get the 275VDC output from the voltage regulator. I tried with zener diodes for 55 VDC (and adjusted the regulator to suit) but didn't really change anything.

The audio tubes are well matched (within 0.1mA) except the 6V6 output tubes which have a few mA difference.

R3 is adjusted to show 0dB compression with no input signal.

Applying an input signal I do get signal at the output however there is no compression unless I drive the signal hard enough, way beyond clipping (clipping starts at app. 45Vpeak on the jack output on the front panel). If I feed a really crazy high input signal I can get the needle to drop, so the meter seems to be working fine. I can get an indication of beyond -20dB compression with a high enough input signal and the output is really completely distorted at that point.

So anyways, this led me to believe that the 6H6 double diode valve, which rectifies the AC signal into the DC control voltage for compression, was not working properly.
Indeed, the schematic seems to indicate that with 35VAC on each side of the rectifier input, I should get a -24VDC* control voltage, however I had 0V with 35VAC.

Received a new 6H6 valve today and it's not working any better. Same behavior. With both the old tube and the new one, I need to drive the input of the rectifier with 80VAC on each leg to get -1VDC on the output (meter indicating -5dB). This identical behavior leads me to believe it is safe to assume that they are both working okay, right?

Checked the whole circuit at the output of the rectifier, no shorts or anything, everything seems to be just as indicated on the schematic. There isn't much to it so I don't really see what could go wrong. S1 (release rotary switch) and associated resistors, R31, R3, C1 and the input pot all test ok. I checked heater supply for the 6H6 valve and its fine, if a bit high (6.6VAC). I have 41VDC on both legs of the rectifier input (schematic says 39VDC) and yes, as mentioned above, the audio signal is also present and symmetrical on both legs.

I have tested for continuity between the 6H6's socket's solder lugs and the tube pins with the tube half plugged in.

What am I missing? Is it possible that there is some sort of leak on the screens of the 1612 tubes explaining that I'm not getting the expected CV voltage (connected to screen of the 1612 - PIN5)? I have measured resistance between pin 5 and all the other pins on both tubes, unplugged of course, and they all read open circuit.

*Should that be 2.4VDC?

EDIT: Forgot to mention but this unit had already been fully recapped

EDIT2: I originally attached the wrong schematic, this has been corrected

 

[Rob Flinn]

use your scope to look at the side chain after the 6h6. That way you can see if you are developing a control voltage.

 

[Rob Flinn]

Also check you have continuity through the release switch.

 

[arthurdb]

Thanks for the quick reply.

Sorry if it wasn't expressed clearly in my post. I am developing a control voltage, but only at insanely high signal levels. In other words, the threshold is much higher than it should be.

With the input pot fully open, I start developing a negative control voltage with app. 0.3 Vrms input, at which point the output is clipping (over 80 volts peak to peak). The control voltage continues to decrease in a seemingly linear fashion as I further increase the input. I can hit the -20dB spot on the meter with app. 0.9 Vrms input, at which point the CV approaches -6VDC

I do have continuity through the release switch, I measured the resistance between ground and the common lug of the switch and I get the correct resistance values in all positions.

EDIT: I have also continuity between the common lug of the release switch, pin 5 of the 6H6 valve (rectifier output), pin 5 of both 1612 valves (screen), the wiper of both decks of the input pot and the filter cap C1.

 

[Rob Flinn]

What is the maximum control voltage you can develop with a 1.28Vrms signal ? i.e 0V to i/p transformer secondary centre tap. Also do you have pads on the output ?

 

[arthurdb]

With 1.28Vrms input signal I get a CV of -7VDC (as measured between the input transformer secondary center tap and ground when the input pot is set to maximum)

I do have pads on the output but I've been mostly checking the output at the front panel jack (so before the pads). The output voltages I've given so far were measured with a high impedance meter at the jack output.

The output starts clipping at 35 Vrms (measured from the from the front panel jack with high Z meter) or 2.5 Vrms measured after the pads (loaded with 620Ω).

The input signal to reach that clipping point is app. 50mV rms and the CV is 0V. The CV barely starts developing when I raise the input to 300mV rms.

 

[Rob Flinn]

Have you checked that those old carbon composite resistors are still within 10% of their rated value ? I have fixed 2 or 3 of these units & a couple of Sa38's. All of them had some bad resistors, & some of the coupling caps (not the electrolytics) were bad too.

 

[arthurdb]

So I checked all the resistors today. Most are within 5%, at worse they're off by 7-8% (edit: a few resistors were discolored from the heat but these ones were all within 5% or better)

The coupling caps had already been replaced previously with sprague caps. The recap job was done before the current owner got the unit a couple years back and it was working fine until now.

I double checked all the DC voltages against those in schematic, most are spot on, at worse off by 10%.

I double checked the tube currents, they are really close to the values on the schematic and are really well matched except for the output tubes which are 10% off from one another

The AC voltages look pretty good too except for two things :

One, I'm missing around 10dB of gain on the output stage. I tried with another pair, which incidentally happens to be better matched (5% difference), but the difference was negligeable.

Second, I have a much lower voltage on the input of the rectifier. With 56VAC on the output of the 6V6 tubes, the schematic says I should get 35VAC on the input of the rectifier, however I only get 25VAC, whether I plug the original 6H6 rectifier in or the replacement one I got.

This is starting to drive me crazy...

 

[arthurdb]

Out of curiosity I tried to inject an external DC voltage from my lab supply for the control voltage with the 6H6 pulled out of circuit. The unit seems to react well, sounds great. Current draw at -20dB compression (-10VDC) is under 1uA once the filter cap settles.

I wonder if I should get a third 6H6 tube, maybe the first replacement was a dud...

 

[arthurdb]

Received the other 6H6 tube, seems to be working like a charm. Just got to fix that imbalance on the output tubes and it'll be good as new. Thanks a lot Rob for your help.