fuzz on clean notes.

[SoulFetish]

I just fixed a 70's pro reverb today which had a similar issue. But before I get into that, I have a couple of questions.
First, do your EL84s share a cathode resistor? Do the cathode resistor(s) use a bypass capacitor? Where did you connect the centertap, or DC referencing resistors in the heater supply?
Does your amp use global feedback (ie 100k resistor from the output stage to the PI)?
If so, does your amp still exhibit this if you disconnect the negative feedback? (Keep in mind, disconnecting this will increase the gain quite a bit in the output.)

[islandamp]

Yes, It has a shared 150-ohm cathode resistor with a cap across it. Also have 2 -50-ohm resistors going to ground ( chassis ), for the heater reference, before the 1st el84, then the other el84, then to the preamp tubes. I used a basic Plexi presence circuit, and at 1 time I cut the lead off of the transformer, and it made no difference. The amp has a fuzz on quiet notes and breaks up on the bass notes, It's a shame because if I play it very gently, at clean volumes, its the best amp I've ever built. Wide-open it definitely rocks. I noticed today that it has a buzz, that I thought was a grounding issue, that goes away at the last 1/64 of a turn on the volume knob. Gets real quiet then. I just kind of hurried the amp together, when I built it, just to see if the Komet preamp with an 18 watt, would sound good. Has a lot of potential if I ever figure it out..

[SoulFetish]

Okay, good. Then most likely your issue is being cause by HF oscillating.
What happens when you disconnect the negative feedback? is the problem still there?

[islandamp]

I once cut the presence wire off at the speaker jack, without removing any of the other presence circuitry. Nothing changed. I remember, once, using a jumper wire, with alligator ends, and connected from the 2nd or 3rd preamp 100k load resistor to the phase inverter, and it cleaned up. I'll have to get that jumper and isolate things again. This amp sounds so good when played "lightly" or full distortion, I think it will be my main amp, if I ever get it figured out. Also there is NO resistor from the input jack to the 1st gain stage. Maybe I need to add a 68k.

[SoulFetish]

okay, good. So the oscillating is most probably in the preamp section. Can you put together a schematic of your amp?
Without that, I'll assume the power supply is as follows.
PI load connected to 20µF Filter cap -->10k dropper--> 3rd stage load connected to 20µF filter cap -->10k dropper--> stage 2 & 1 load share 20µF filter stage.
(with the caveat that you have the experience to do the following safely...)
If you were able to get it to stop by connecting the second or third stage load resistor to the PI filter stage, then this may be an indicator of a couple of problems.
First, that the filter capacitor in that particular node is faulty and needs to be replaced. You can test this by taking a 20µF HV cap with alligator leads (or solder tack it in place), and add the capacitor in parallel to each of the stages and see if the problem goes away.
The second is that preamp stages 1 & 2 need to be decoupled. This can be accomplished with out changing the operating conditions by connecting each plate load resistor(s) to their own 20µF capacitor. Then, take two 18k-22k resistors and join both of those to the 20µF supply node of stage 3.
One more question. is the coil around the wires running from the switch to the stage 2 plate resistors grounded at one end?

[Davidg]

You said there was no resistor from input jack to 1st gain stage? If this is true it is most likely your problem. You need a 1M going from grid to ground for a grid reference resistor. This also sets the input impddance of your amp. You must have a reference to ground there though and preferably a 33-68k resistor soldered directly to the tube pin( bare minimum 10k) look at typical fender marshall inputs and copy them if u need scheme of what I'm talking about. Try it and report back!

Also the coiled wire kn the Komet amp is zn old trick to stop parascitic oscillation. It was first used by Mesa in some of the old mark series amps. So that preamp is prone to kscillatio. I would copy tbat if I sere u jt may be enkugh to stop it. U still need 1M at input regardless.

[pompeiisneaks]

You said there was no resistor from input jack to 1st gain stage? If this is true it is most likely your problem. You need a 1M going from grid to ground for a grid reference resistor. This also sets the input impddance of your amp. You must have a reference to ground there though and preferably a 33-68k resistor soldered directly to the tube pin( bare minimum 10k) look at typical fender marshall inputs and copy them if u need scheme of what I'm talking about. Try it and report back!

Also the coiled wire kn the Komet amp is zn old trick to stop parascitic oscillation. It was first used by Mesa in some of the old mark series amps. So that preamp is prone to kscillatio. I would copy tbat if I sere u jt may be enkugh to stop it. U still need 1M at input regardless.

Are you talking about a coil of wire soldered to the chassis that's wrapped around the signal wires to remove noise? I've seen that in silverface bassman as well, were mesa around back in the silverface era? So fender stole it from them right after? I always thought it was at least first in fenders, but they were already known for just taking known ideas from previous amps and such that were not specifically for guitars. I think it was the CBS era that they tried pushing the designs into more hifi territory and cleaning them up more etc.

~Phil

[Stevem]

The wire deal used to suppress oscillation issues in the Mesa amps ran from the plate of a preamp tube then rapped 4 times around that gain stages input grid if I recall right, but the raps where spread out pretty far .

In the fender amps that wire rap off of ground is used as a partile shield .

[pompeiisneaks]

Oh ok interesting, different use cases, similar concept though I guess. Still controlling noise. does it behave differently than the grounding method? I.e. the grounding method just intercepts stray noise into the coiled wire, acting like a simpler 'shielded wire' but how does this case work?

~Phil

[sluckey]

The wire acts just like a small cap connected between plate and grid. This is a local NFB loop that reduces gain at high frequencies, thus reducing/eliminating hi-freq oscillation.

[pompeiisneaks]

The wire acts just like a small cap connected between plate and grid. This is a local NFB loop that reduces gain at high frequencies, thus reducing/eliminating hi-freq oscillation.

ahh that totally makes sense, cool, thanks!

~Phil