help with EV 7600 power amp

[Colonel Sanders]

i recently aquired an old electrovoice 7600 power amp. unfortunately i wasnt lucky enough for it to work right away. as it turns on the protection lights come on (normal) but do not turn off (not as good). i believe this amp is similar to the 7300, which was a bit more popular.

i poked around a little, but cant find anywhere to start looking. any advice to start troubleshooting would be much appreciated. i dont have much for lab tools, just multimeters and one good scope.

channel one and two are on separate, identical boards, and both stay in protect. i immidately noticed that R21 and R30 get very very hot. voltage coming out of the rectifiers is 205Vrms measured between the + and -. i believe there is some funky floating ground in between, putting the rail voltages close to the ±99V spec.

thats about all i got for now, the manual can be found here:
www.electrovoice.com/downloadfile.php?i=969623
which is kind of readable, schematics are at the end.

many thanks in advance

[sgrossklass]

That sort of concept is not uncommon in PA amps (e.g. QSC). While requiring separate floating supplies for each channel, it allows ideal thermal contact for the power transistors since they don't have to be isolated from the heatsink (as their collector connects to ground).

These "inside out" circuits really throw you for a loop at the beginning; it'll get better eventually. Ultimately it's only a matter of which side of the load you connect ground to.

It is normal for R21 and R30 to get hot - they have to dissipate almost 2 W each. Make sure they are dropping the +/-99 V rails to +/-15 V.

The clipping indicator does not light?

Stupid question: Do you get any output, and if so, how does it sound / look on the scope (ground to 1-)? There is nothing keeping the signal from getting there, AFAICS. Do check DC offset, just in case (if there is any, nothing but a leaky electrolytic could cause this).

The number of semiconductors isn't that big - any indications of any of them being dead short?

Make sure that the bias pot makes good contact. Otherwise bias would creep up - without disastrous consequences usually, as the NTC keeps things in check, but the amp would run hotter than it should.

[Tekko]

I'd say the output transistors are toast, most common type of failure in PA amps as these are always driven to and beyond their limits.

[Colonel Sanders]

voltages with respect to speaker (-)
voltage drop across R30: -126volts one side, -15 on the other
drop across R21: 74volts to 15 volts

the speaker output is about 12V but it fluctuates a lot


and something interesting happened this time: started it up, protection light is on.... clip light flashes for a moment then goes out, the protection light goes back on, but rather dim.

and now something went pop.... let me take a look and see what happened.....

[Colonel Sanders]

so from what i have read and tested, im guessing output transistors.... R32 at the output is getting very hot, and there is DC at the output.

whats the best/easiest way of testing these transistors?

i read somewhere i can take out all but one pair, and add them back in until i find the defective pair... is this a valid method?

[sgrossklass]

Originally Posted by Colonel Sanders

voltages with respect to speaker (-)
voltage drop across R30: -126volts one side, -15 on the other
drop across R21: 74volts to 15 volts

the speaker output is about 12V but it fluctuates a lot

I'd say you've got a bad filter cap there, on the positive supply side. These are 4700µ / 100 V types, and as such operated at rated working voltage. It is not unusual for old electrolytics not to withstand full rated voltage any more.

Try unsoldering C18 or C19 (resolder and try the other if no effect).

But wait...

so from what i have read and tested, im guessing output transistors.... R32 at the output is getting very hot, and there is DC at the output.

Wait a microsecond... if R32 gets hot, the amp must be oscillating like mad. Not sure why that would be though - parts degradation, bad electrolytics?

Testing transistors properly is not easy. A diode test will tell you whether something is disconnected or dead short. You will need more fancy tools to determine hFE. And then you still haven't found out whether the transistor can withstand full rated voltage.

[Colonel Sanders]

Originally Posted by sgrossklass

Try unsoldering C18 or C19 (resolder and try the other if no effect).

i was thinking the same thing before you responded. i ended up unsoldering all four caps, and jumpering in a pair of 10,000uF caps that i know are working. no change in the operation (or lack thereof). im currently soldering the originals back in, should be done in a few minutes.

in my research i found one guy who said that because output transistors like these are just in parallel, i can unsolder all but one pair, then add pairs back until i find the one that is shorted.... is this a valid method and worth a shot? (i dont mind unsoldering and resoldering the whole board if i have to)