Oscillation at start up

[Zero Cool]

I just finished my amp rebuild and i noticed during final testing that the channel i just repaired, oscillates at a very very low level at start up before the protection relay kicks in. Its a clean 3+mhz signal. i did not get a chance to measure the voltage of it, but it is less then 1 volt p-p.

The oscillation goes away the second the protection relay kicks in and the output R-C gets connected. But there is a 10 second delay before that relay kicks in.

Should i be worried about this? Once the relay kicks in the amp runs great. no signs of oscillation. it only happens when the amp comes online, stabalizes and then there is the 10 second delay before the output relay kicks on that you can observe the waveform at the amp output.


Suggestions of what to look for???



Zc

[lineup]

Originally posted by Zero Cool
I just finished my amp rebuild and i noticed during final testing that the channel i just repaired, oscillates at a very very low level at start up before the protection relay kicks in. Its a clean 3+mhz signal. i did not get a chance to measure the voltage of it, but it is less then 1 volt p-p.

The oscillation goes away the second the protection relay kicks in and the output R-C gets connected. But there is a 10 second delay before that relay kicks in.

Suggestions of what to look for???

Can't you put the R-C Zobel stabilization filter before the output relay?

[Zero Cool]

Not easily.

the factory has it attached right at the speaker terminals. to move it would require major changes. but it could be done.

[darkfenriz]

I really wouldn't worry

[georgehifi]

Did you see if it oscillates with the input shorted or open circuit?
Also check the -3db high frequency point of both channels to see if they are the same.

Cheers George

[Zero Cool]

Originally posted by georgehifi
Did you see if it oscillates with the input shorted or open circuit?
Also check the -3db high frequency point of both channels to see if they are the same.

Cheers George

I will double check that. I know the input was connected to my sig gen at that time but the level was turned to zero. I will retest and put a shorting plug in the input and see what happens.

I will also test the -3db point.

Square wave response of both channels was not terribly fantastic. and the channel that has the problem seems to drain is caps quicker on power off. Running at 1/3 power for about 20 minutes. the tops of the TO-3 devices were running about 101-103 deg F on the problem channel and about 99-102F on the good channel. Bias is spot on and varies less then .03mv from device to device N or P.


Zc

[AndrewT]

Hi,
I had not thought about this when I have been re-quoting the general advice to put the RC Zobel at the output terminals.

Would a smaller current version of the Zobel be appropriate at the relay terminals to give the amp a high frequency load at start up prior to switching over to speaker output?

Usually 10r + 100nF is used but how about 15r or 18r + 47nF at the relay and similar at the speaker terminals.

[Zero Cool]

Originally posted by AndrewT
Hi,
I had not thought about this when I have been re-quoting the general advice to put the RC Zobel at the output terminals.

Would a smaller current version of the Zobel be appropriate at the relay terminals to give the amp a high frequency load at start up prior to switching over to speaker output?

Usually 10r + 100nF is used but how about 15r or 18r + 47nF at the relay and similar at the speaker terminals.

If the relays were DPDT i could have added a second RC to load the amp down until it switches over.

The current R-C is 0.1uf and 4.7 ohms


Zc

[pinkmouse]

The oscillation will be there even after the signal is dumped to ground by the Zobel when the relay clicks over, you just can't measure it any more. I'd try a bit harder to track it down and solve the problem. What is the amp? Schematics?

[AndrewT]

Hi,
using a DPDT will ensure the amp is loaded before switch over and again after switch over, but during switch over it will be unloaded.
What if that brief period was enough to cause a component failure somewhere just as the speakers are about to come on line?

9r1 or 10r + 47nF at both the relay and output should mimic the speaker connected load condition and give a bit of loading prior to and during switch over, maybe just enough to suppress the oscillation.