fuzz on clean notes.

[islandamp]

My 18 watt build uses a Komet preamp and a 2-el84 trainwreck power amp. The amp has a pretty killer distortion, but turning down the guitar volume, or amp volume produces a thin, brittle sounding amp with a weird distortion that rides on each note and kills the sustain almost instantly. Happens at all amp settings which include, Hi-cut, presence, Bass, mid, treble, and volume. Any ideas ?

[tubeswell]

Check that your preamp filter caps are actually grounded. (If not, the preamp gain stages will interact/couple with each other through the power supply rail, because the dropper resistance ends up acting like another load resistance in series with the plate resistors, causing a fizzy mushy sound)

[Stevem]

Your amp is Oscillating when you have a fuzz harmonic riding on top of a note that kills sustain.

This is caused by many things , but the most common with a build is the wire and component layout.

What you have is a higher signal level from one gain stage impressing itself back down stream into a lower gain stage.

This issue most times means that 3 gain stages are involved, this is due to the signal inversion that takes place thru a gain stage.
For example the signal enters a gain stage grid as a positive and when it leaves the plate on its way to the next gain stage its phase has been reversed.
To have a Oscillation problem the signals involved need to be in phase, this means the signal must go thru 3 gain stages to have it be in phase.

One kinda easy way to find out what stage is the issue is to solder a length of wire to the preamp tubes plate that is long enough to allow you to rap it around that gain stages grid wire 3 to 4 times.

This will apply negitive feedback and will reduce the or eliminate the problem you have.

This is not a cure , just a crutch to help you track down the issue.

As far as the wire layout goes any signal carrying grid wires that span 3 gain stages should not run parallel with a later plate wire and need to be seperated by atleast 1/2", and if they need to run across each other they should do so at right angles for minimum interaction!

Remember this , grid wires are high impeadance and act like a Antenna , especially the grid wire on the first gain stage of the amp where the Guitars signal output is small.

In general for a good build layout you want the grid wired as short as can be due to then as stated above being high impeadance.

If you need to rework your layout do so by making the plate wired longer and the grid wired shoter!

Also parts wise coupling caps can be installed in a way to help or hurt keeping a amp out of Oscillation problems.

On most modern coupling caps if the lettering is facing you then the input side signal wise will be on your left.
Installing caps this way means that they are self sheilding and not a big tubular signal radiator!

Also high resistances can act like a Antenna in regards to how good the contact is between tube pins and the socket.

Some contact inhancing spray can at times go a long way to getting this better.

If the pins on my preamp tube pins do not shine I take my Dreamel tool with plate type wire brush loaded into it and shine them up as best as I can and then wipe on a conditioning spray with a Q-tip.

There are more ways to trouble shoot this issue so if these things do not help then post back.

[Stevem]

Sorry about the double post, I have no idea how I managed this, lol!

[pompeiisneaks]

Sorry about the double post, I have no idea how I managed this, lol!

I deleted the second. I've accidentally clicked on submit but didn't see anything happening and thought maybe I missed it, so I hit it again and it posted twice. It happens

~Phil

[Leo_Gnardo]

Let's not overlook the obvious.

1. A bad speaker may sound OK with the amp blaring through it, but sound naff at low volumes.

2. Insufficient bias current. Similar symptom as ^^^.

3. 1 bad output tube, or 1 output tube not amplifying for whatever reason - could be an open screen grid resistor or a number of other things.

[islandamp]

I'll check it all again, this weekend. Built this amp years ago and got pissed every time that I rechecked everything, only to find no problems. Could it be the presence values? I am using a 50 watt preamp { Komet } and an 18 watt power amp. I kinda cobbled them together. I may have some weird values that are sending some overpowering negative feedback.

[pdf64]

...To have a Oscillation problem the signals involved need to be in phase, this means the signal must go thru 3 gain stages to have it be in phase...

Just to note that only 2 cascaded common cathode (ie polarity inverting) gain stages are required for an inadvertent positive feedback loop to form, ie the signal at the 2nd plate will be in the same polarity as the 1st grid.
To expand on the points mentioned, the perquisites for oscillation are unity gain and 0 degrees phase shift between input and output.
For any implemented schematic, there's an additional overlay of parasitic capacitance connecting every circuit node to every other; the inverse square law applies, so that the capacitance increases significantly as the distance between any 2 nodes decreases.
Given the high input / grid impedances of of guitar amps and a potential voltage gain from 2 stages of ~3600, it takes very little parasitic capacitance for sufficient signal from a '2nd' plate to 'impress' itself on a '1st' grid, such that the conditions for oscillation are then met.

Hence layout and lead dress are key aspects of successfully building any given schematic design; and even if a build doesn't actually exhibit free running oscillation, inadvertent positive feedback loops can cause unforeseen resonances that affect the frequency response and cause blips and chirps of parasitic (ie not free running) oscillation at certain signal levels / control settings / input / output conditions.

Of course all layouts involve some degree of compromise, and if positive feedback loops are an issue, then the strategic addition of a suitable grid stopper or 2 can help to damp those resonances down and achieve a workable margin of stability, without messing with the schematic design.

[tubeswell]

^What pdf64 said^

And use your R-meter to confirm you have DC continuity between each filter cap ground lead and the main ground return point.

[islandamp]

Everything looks good. The only thing that I can think of, along the lines of what everyone is saying, Is that coiled wire that Komet shows, in their gut shots. Looks like it goes around the gradual/fast switch and the associated components. That 1st and 2nd preamp stage use a common 300 volt power node. Could that coiled wire be grounded on one end, and used to kill oscillation?

komet gutshot.JPG

[Stevem]

Well from the looks of things to me your PI output leads from the caps to each output tube grids is likely too darn long.

I would get some good shielded cable and remake those 2 wire runs and ground each shield at where the output tubes cathodes go to ground.

Also your cable tying of the lead off the input jack and that wire off of that gain stages plate to the pot could be a issue also.

[islandamp]

This is a picture of an actual Komet, not my amp. Just trying to show how they have a coiled wire doing something. Maybe they are shielding that shared plate voltage node, if so, I didn't do that. will try wrapping a few turns around that are tonight and see if it helps..

[Scumback Speakers]

It could be the speaker or the cable running to it. What speaker is in the amp?

[islandamp]

It's hooked to many working cabinets, over the years. This has frustrated me for years. Eventually, I get tired of searching for the problem and put it back upstairs. Shame, because at 3/4 volume and up, It's a great sounding amp.

[M Fowler]

That coiled wire is just a shield, it is soldered to ground on one end and coiled around the fast/gradual wires. On the other end the coiled wire end is shoved inside the coil.

How about 47pf cap across your power tube grid caps. Fizz cap and increase the plate voltage.