Counterpoint SA-100 Head Scratcher

[genther]

Hello everyone,
I'm looking for insight to a problem that has me baffled from anyone who is familiar with the SA-100. I'm working towards a solution for getting mine working again by replacing a blown output stage. I picked up 4 lateral MOSFETS from Exicon and tinkered with the bias circuit and gate resistors until it was working perfectly (well as perfectly as these amps can be given their design faults.), I just like the way it sounds. I wanted to eliminate the poor connections at the TO-3 sockets, so I used the TO-247 package to get solder connections. The gate resistors, zeners, and wiring back to the board hanging off the MOSFET pins offended my sense of sanitary electronics. I also wanted to replace the tired caps, use better resistors from PRP, and replace the troublesome bias voltage regulator circuit, so I invested in new PCBs to mount on the heatsink that replaced the entire output stage.

I cannot get the damn thing to work. There are 0 volts at the gates. I have checked over and over the boards against the schematic, the solder joints, and continuity between every component. The component values on the new board are identical to the components on the channel I have working in the amp. Attached is the schematic I drew for the new circuit with the voltages noted in red and green. Please let me know if you can offer any ideas, I'm hoping I'm blinded by something obvious. Thank you.

[llwhtt]

K1 grounds the gates, is K1 energized or de-energized?

Craig

[genther]

Hi Craig,
A great guess and just the type of dumb thing I would overlook; however, I have not mounted K1 yet for testing, so it isn't that.

[vilfort]

Hmm... Here is what I would do:
Remove the output MOSFETs. Then test the BIAS circuit by itself. You are right in expecting that we should see something +/- 0.7V at the gates. If the VBE multiplier transistor is fried and has a dead short, then it would act as a solid wire and short the gates to a floating zero volts. If you still see zero volts after the MOSTFETs are removed, then remove the transistor and test w/ your multimeter. If you still have a short on the transistor itself, then you need a new.
Once the VBE circuit is up and running, you should be able to see the voltage go up and down a little as you turn the BIAS trim pot. Again rough adjust w/ your meter and shoot for +/- 0.7 volt. Once you are this far, then re-connect the MOSFETs and try to fire the beast up. Then re-adjust BIAS for circa 300mA standing current in line with the positive rail and the MOSFETs. Then adjust for 0-10mV offset on the speaker terminals. Now the output stage is doing its thing and it is time to focus on the tube circuit. Hope it all goes well.
All the best,
D.

[llwhtt]

I thought of a shorted VBE transistor also but R13,R60,&R61 would keep it from looking like a dead short no matter how it was shorted. There are just so many ways to get zero volts there "assuming" the parts are good and wired correctly.
Could it be ground and NOT zero volts?

Craig

[genther]

Thanks for the input guys. It was very helpful that you didn't see any problems with the schematic (you didn't mention it so I assumed you didn't), so I was able to eliminate it as a problem source. Anyway, I had two problems, one was a late night and a bad test wire. The other was a 25R VR1 instead of a 25K VR1. Funny how you can look at the same thing many times and still see what you think you see. The persistence of memory.

[llwhtt]

I guess "assuming" bit us in the butt again, and like always it was something stupid. I "assume" it's working now?

Craig

[genther]

Yep, and the first tune out of it was "Pump Up the Volume" which I did. Thanks again.

[vilfort]

I just replaced the relays in an old SA100. Band new releays. Worked great on one channel - not on the other. Took a while to figure out what was going on... Then it hit me: the darn relays have an internal diode. So the polarity matters. The relay that had the right polarity worked - the other did not. After I cut the traces and swapped the polarity on the relay that did not click on after the NE555 timer turns on the 6V (0V at the gates like genther observed), all worked as desigend. Just thought I would share in cased anyone else have been scratching their heads over this one...

D.