Parasitic Oscillation

[gearhead]

What causes high frequency (above audible) parasitic oscilliation?

Enough to sap the life out of an amp?

[skyboltone]

What causes high frequency (above audible) parasitic oscillation?

Enough to sap the life out of an amp?

It's an amp. I amplifies AC voltages. Sometimes we deliberately make them amplify radio frequencies on purpose. It's called a transmitter then, only real difference is that we do the impedance transformation with a CL matching network instead of an output transformer.

What's an oscillator? It's an LC or LCR circuit where the output is fed back into the input. With certain combinations of L and C you end up with frequency multiplication.

When your (audio frequency) guitar amp multiplies frequency do to combinations of circuit overload, stray L and C, and positive feedback it oscillates. It can try to produce more power at the radio frequency than the PS can supply.

[gearhead]

Thanks for the technical rundown.

From a practical standpoint, should you be looking for capacitance in places where there shouldn't be? Are there target points where this typically happens?

[skyboltone]

Thanks for the technical rundown.

From a practical standpoint, should you be looking for capacitance in places where there shouldn't be? Are there target points where this typically happens?

I wish it was that simple. As a general rule I think just looking for some sort of positive feedback is the first step. Somehow, the signal is crossing over itself and running back through the same stage.

An old trick with transmitters is to tape a neon lamp to the end of a chopstick and run it around tube bases and wiring inside the amp, or around the tubes themselves. If it starts glowing you've found the place where the parasite originates. Then look there for potential positive feedback; wires running parallel, something that might radiate signal into an ajoining component. I dunno. In principal you're looking for a tank circuit that shouldn't be there. Radio uses oscillators like mad to add frequencies together and subtract them from each other. These tuned stages are tuned with adjustable coils or caps. In an amp I think we're just looking for a grid to plate or grid to cathode circuit from the same or a following stage that's in phase with the guilty stage.

To be honest, I don't know how often audio amps are beset with radio frequency parasitic oscillations. I assume you're still dealing with your underacheiving express. The slow train......

Hang in there
Dan

[gearhead]

Yeah, it's my miscreant express. Have spent enough time on it to build another one. Literally. And a few feet of new wire and a fistfull of shotgun parts.

The only thing I can think of is a parasitic oscillation or some related solder joint issue.

[skyboltone]

Somebody here mentioned one time a bad tube socket. Intermittant on a tube pin.....

Are you using Beltons?

[gearhead]

No, some no-name, middle of the road quality. How much better are the beltons? I want to get some good ones for my Liverpool build anyway.

Intermittant? Wow, that's a hard one to find. Have done lots of continuity checks (pin-to-socket, and between pins), but no go so far.

It seems to loose it's sensitivity (even with no input signal) somewhere in the V2 to V3 area.

Replaced the PI socket and all it's connecting wires (as well as all the PI caps, coupling cap to V2, and coupling caps to power tubes). And yanked/redid the entire pot assembly and any wire off of it, replaced the input jack/wire, and whatever else made sense, LOL.

[Luthierwnc]

You can make a handy checker by shrink-tubing a 250 puff cap (with an adequate voltage rating) on the end of a chopstick. Have one lead on the tip and attach the other to a ground clip. Then just touch various points in the amp. You'll kill some audible highs but it might help you find frequencies you can't hear. Basically you are hanging supression caps without having to solder them in and out.

Good luck, Skip

[gearhead]

Do you use it with or without an input signal?

[skyboltone]

No, some no-name, middle of the road quality. How much better are the beltons? I want to get some good ones for my Liverpool build anyway.

Intermittant? Wow, that's a hard one to find. Have done lots of continuity checks (pin-to-socket, and between pins), but no go so far.

It seems to loose it's sensitivity (even with no input signal) somewhere in the V2 to V3 area.

Replaced the PI socket and all it's connecting wires (as well as all the PI caps, coupling cap to V2, and coupling caps to power tubes). And yanked/redid the entire pot assembly and any wire off of it, replaced the input jack/wire, and whatever else made sense, LOL.

The "build for profit" crowd on here swear by them. I used to use ceramics but will never again unless I can find NOS amphenols.

[Luthierwnc]

Use a signal unless the power tubes are running away without one. A parasitic is basically an ultra-high frequency being amplified to the point that it robs power from the frequencies you want. They can red-plate a tube. Usually it is a piece of the input signal being oscillated by layout or component problems.

What you are doing with the probe is shunting high frequencies to ground at various points in the amp. You'll hear some loss of highs but it should also ground the ones you can't hear as well.

Another thing to consider is dropping parts of the preamp chain to isolate the problem. In an Express, you can bypass any of the three gain stages and still get the amp to make noise. The third is easiest by just pulling the tube and jumping from the .002 to the inverter. Remember you are reversing the phase. You can lift a ground or grid on the first two stages too.

In addition to the parasitic probe, I jury-rigged a shielded jumper. It is just some coax with a clip on both leads and another clip on the ground. Although it does add some capacitance, grounding the wire keeps from adding unnecessary oscillations (or channeling Swedish ham radios) when you hop-scotch the circuit. sh

[skyboltone]

Hi Skip, thanks for the help. For a little history here, Gearhead made an Express that doesn't get wild and crazy. It's waaay too tame. He's been farting around with it for a couple of weeks and only got it to do the Express thing for about 5 minutes. So far most of the guru's haven't helped out.

Gear, did you post pics? Maybe start a new thread. Let's get to the bottom of this.

Dan

[Luthierwnc]

Thanks Dan,

My problems with this type of build has been keeping them stable. Hair-triggers on these rigs. I've always ended up trying several values in the coupler and grid leak into stage 3. My 336 copy is the acid test. If I can keep it from feeding back, I'm there. sh

[Luthierwnc]

Question:

I haven't spent much time here or on the TW side. Has anyone ever posted the definitive TW study with sine wave pictures at each point in the preamp circuit? Gearhead (and everyone else) might get a lot out of seeing what the millivoltage should be at the different stages.

Food for thought, sh

[skyboltone]

Thanks Dan,

My problems with this type of build has been keeping them stable. Hair-triggers on these rigs. I've always ended up trying several values in the coupler and grid leak into stage 3. My 336 copy is the acid test. If I can keep it from feeding back, I'm there. sh

Oooooh. Did you build the 336? Solid hog with a maple top? Pictures in the Garage section please.