Polytone Power Amp Oscillation

Hello,
I have a Polytome Mini-Brute 3 on my workbech.
Preamp is fine but the Power transistors were blown.
I changed the Power and driver transistors but the amp
shows oscillation (under load only) now.
Any Idea what to do ???

I'm a tube guy and overstrained with it .

Here's the Schematic:-sorry it's upside down
http://www.murchmusic.com/Polytone Info/schem2.JPG

thank you for help
 
It's junk.

Throw it out buy a new Behringer.
It's junk too... ;)
Keep the speaker and cabinet.

That's the glib and fast response.

you didn't state what sort of oscillation it shows under load??
LF, HF, riding on the signal, audible, ultrasonic, what??

What parts did you use to replace the old MJ802/4502 with?
Perhaps since they are not followers here, the gain of the new devices plays a role... or not.

How about the input pair??

Etc...

It's the same as tubes, but think triodes and positive bias, not negative bias... higher current, lower voltages... the short version...

_-_-bear
 
thanks for your answer bear.

It is HF oscillation which is riding on the signal.not audible(maybe a litle distortion)
It dosn't shows up on small output voltage swing.(up to 4Vpp into 4 Ohm)

I changed with original (recomended) parts.
MJ 4502/802 for power and MJE 172/182 for driver .
I didn't change the input transistor.should I ?

May there is a easy way to mod the circuit to a more stable one.
Loosing some HF wouldn't be a problem.
thanks
 
Thanks Elwood !

I just put a 100 pF bc-cap across Q4 ( driver on neg. rail)
and this seems to solve the osc. problem.
I saw this bc-cap on the in the elliot P3A wich is similar to the polytone power stage.
Could someone explain why only one driver is HF-bypassed with
a bc-cap.
Would it be safer to spend a second bc-cap for the pos.rail driver?

Thanks for your help .
 
Hello again...

I'm relatively new to this as well but as i understand what you have done is create local feed back on that one transistor so that one transistor only processes 1/2 the signal so doing the same to q3 would at least make it symetric. That being said i think either c7 or c12 may be leaky. play with it and see what happens.


Regards, Elwood