Help needed in solving high pitch noise

[stephenl]

The subject amp is quiet and stable up to a relatively high level of gain...then it makes a very high pitched noise whose frequency can be changed with the treble and hi-cut controls. I guess it's a high frequency oscillation?

I've swapped tubes and chopsticked wires - no change.

Its basically a KF50 circuit.

I've been doing a lot of tweaking and playing with the gain at 2:00 or less, so I haven't run into the issue previously.

Any suggestions?

[Stevem]

If you swap the wires on the grids of the output tubes, or the swap the wires on the plates of the output tubes, does the issue get better or worse?
Turning the amp up to the point of the noise taking place and then placing a short across each preamp tube plate load resistor will help you to pin down what stage/section its starting in if its not the phase thing I posted about before with the wire swapping.

[martin manning]

Turning the amp up to the point of the noise taking place and then placing a short across each preamp tube plate load resistor will help you to pin down what stage/section its starting in if its not the phase thing I posted about before with the wire swapping.

Short the plate loads? Why subject the tubes to that abuse? A better way to isolate the place where the oscillation originates is to ground the output side of the coupling cap, or if you have made a listening probe using a 600V cap, use that to AC ground the plate or grid of each stage.

[Stevem]

Just for a second to turn off that stage, your not gonna burn up a tube that fast.

[martin manning]

You can easily exceed plate voltage limits by connecting the plate directly to the supply node voltage.

[tubeswell]

It might be HF oscillation that's there all the time but you can only hear the lower frequencies of it. If you disconnect the global NFB for a moment, is it still there?

Edit - I'm not asking that because of thinking you may have the wires crossed. I'm asking because if it does disappear when the NFB is disconnected, it may be unstable at HF and you may need to 'slog the dominant NFB pole' (in Merlin's parlance).

[stephenl]

Thank you for your help/suggestions. I didn't use a NFB loop on this one. The PA is KOC dual pot bootstrapped MV > 12AT7 LTP > PP KT66 w/ cathode bias and no NFB.

What does a "listening probe" consist of? Sounds like something I should make.

[bcmatt]

Turning the amp up to the point of the noise taking place and then placing a short across each preamp tube plate load resistor will help you to pin down what stage/section its starting in if its not the phase thing I posted about before with the wire swapping.

Short the plate loads? Why subject the tubes to that abuse? A better way to isolate the place where the oscillation originates is to ground the output side of the coupling cap, or if you have made a listening probe using a 600V cap, use that to AC ground the plate or grid of each stage.

Hey Martin, I hadn't heard of the listening probe idea before. Is this what you are referring to?:
http://www.el34world.com/Forum/index.ph ... 893.0;wap2
It seems like a great idea. Now I need t figure what I did with my old computer speakers... pretty sure I ditched them. Will have to keep my eyes out for some more.

[martin manning]

Yes, you can find drawings showing how to construct one on a chopstick or some such thing. I made one a while back and posted it here somewhere. Be careful about the signal level you are feeding into a solid-state input though. The cap blocks the DC but you could be sending 100V pp AC in. You can add a 1M pot to attenuate the signal.

[bcmatt]

Yes, you can find drawings showing how to construct one on a chopstick or some such thing. I made one a while back and posted it here somewhere. Be careful about the signal level you are feeding into a solid-state input though. The cap blocks the DC but you could be sending 100V pp AC in. You can add a 1M pot to attenuate the signal.

Cool. Thanks! I found this while trying to search for images:
http://billmaudio.com/wp/?page_id=1254

[stephenl]

I had a chance to put a scope on it. The noise looks to be about 10kHz.

The first place it shows up is on the grids of the power tubes

Nothing in the preamp or inverter.

I have 10k grid stoppers soldered directly to the socket grid pins.

My shared cathode resistor is unbypassed - so does that generate a little local feedback? Can that cause oscillation?

Should I maybe try putting 220pF caps from grid to ground?

[Tillydog]

I have 10k grid stoppers soldered directly to the socket grid pins.

How about the screen resistors?

[stephenl]

This really isn't making much sense to me.
With a guitar plugged in, but no input signal to the amp other than the usual white noise, I put the scope probe after every coupling cap and saw nothing other than a little "hash". Between the pi output coupling caps and the PA grids I saw fairly sinusoidal 10khz-ish signal. The frequency and amplitude could be varied a bit by changing the volume, treble, mid, master volume, and hi-cut pots.

If I pull the guitar cable out, grounding the input, the noise goes away.

[stephenl]

1k5 screens directly on the pins as well.

[stephenl]

The screens are fed from a mosfet amplified zener stabilizer...could that be inducing the noise?