Oscillation problem with Express Clone

[wizardof_oz]

Just recently I finished a build of TW Express clone, which was a fun project and still challenges me with an oscillation issue
I did quite some research on the Express in the trainwreck sections of this awesome forum before and after the build but now I'm kinda stuck and I would need some help from you guys.

The project is based on the Kelly schematic with UR12 corrections. I started with most original values but decided to change the layout of the power supply caps (didn't like the idea of caps bundled like in the original Express), for a slightly different BIAS voltage circuit and elevated heating for the tubes to avoid hum.
For more stability, I added grid stoppers plus shielded wires between input jack and V1b (10k, pin7) & between volume pot and V1a (820R, pin2), and changed the El34 grid stoppers from 1,5k to 10k.

As iron I used an OT (6.6k) and a PT (325-0-325, custom made with 75VAC sec. for BIAS and elev. heating) for 230VAC, both from Modulus Amplification in the UK.
Voltages are a little above the values from the Kelly schematic with UR12 corrections but seem pretty normal to me: wall 235VAC, B+1 413VDC, B+2 399VDC, B+3 319VDC, B+4 301VDC, B+5 286VDC, BIAS -34,6VDC, filament 6,43VAC with elev. CT (22,8VDC vs. GND).

Like many before me with this highly sensitive circuit, I have significant squeal (oscillation) with the VOL set to noon. I get the harmonics but the the annoying squeal kills them...
Even though there are many hints about extreme highs and moderate lows, mine has a moderate high end but the low end is (too) massive. Not sure what the right term is, but I'd describe it as booming bass. I lowered R7 to 120k to reduce the bass a bit.
When attennuated, the amp sounds good but without the harmonics.

Here's the schematic and layout for reference

Schematic_TW_Express_Crazy_Train_V04.pdf

Trainwreck_Crazy_Train_V6_EH.pdf

Plus a list of measures I found in Ron Worley's awesome build guide of and threads here at TAG as an overview what I tried already...

Measures.pdf

A bunch of gut shots are attached as well.

Hopefully some of you have additional ideas what I could try to get rid of the oscillation.

[Matthews Guitars]

It's a sensitive design. Use high gain preamp tubes and turn it up and it's going to squeal at some point. Hand pick preamp tubes that balance the available gain. And don't expect that this amp can ever be dimed on all controls and remain stable. This is a high performance thoroughbred racehorse of an amp. It's very high strung.

Add shielded cable for the input path and as much of the audio path as you are motivated to install. It WILL help.

The changes you've made are unnecessary. Elevating heaters serves no practical purpose but complicates what should be a simple build. Built properly you don't GET heater hum.

And I'd have recommended you stick with original component values which are proven to make a kick-ass amp like no other. Why not TRY success before messing with it?

I'm curious to know how capacitor type affects tone. You used orange drops, I used MKT1813s. Wonder if that really alters the tone to any noticeable degree?

[martin manning]

Tone stack wiring is very sensitive, so try moving leads around with a chopstick and looking for any that affect the tone of the squeal. Does the bright switch affect it? I agree on the heater elevation, none needed unless there is a cathode follower that runs Vh-k near spec limit.

[Roe]

try to twist the wires from the phase invert to the el34 grids. bigger grid resistors downstream also help with oscillation and blocking distortion. try 47k at the el34s and 68k at v2 and v3

[wizardof_oz]

It's a sensitive design. Use high gain preamp tubes and turn it up and it's going to squeal at some point. Hand pick preamp tubes that balance the available gain. And don't expect that this amp can ever be dimed on all controls and remain stable. This is a high performance thoroughbred racehorse of an amp. It's very high strung.

Add shielded cable for the input path and as much of the audio path as you are motivated to install. It WILL help.

The changes you've made are unnecessary. Elevating heaters serves no practical purpose but complicates what should be a simple build. Built properly you don't GET heater hum.

And I'd have recommended you stick with original component values which are proven to make a kick-ass amp like no other. Why not TRY success before messing with it?

I'm curious to know how capacitor type affects tone. You used orange drops, I used MKT1813s. Wonder if that really alters the tone to any noticeable degree?

You may be right for the EH voltage and the Bias voltage, but this should not have a significant effect on the audio path.
Actually I started with all "original" component values (if there is such thing...) based on the Kelley´90 schematic ´, except for a 10k grid stopper before V1b, a 22nF cap for C9 (instead of 0,1uF) and 100pF for C12 (instead of 50pF), all of which should not have an audible impact. All changes after this were done due to instability and a very bass heavy behavior, which seems untypical for this circuit design.

As for the shielded cables: I already have shielded cables between input jack and V1b, between vol pot and V1a and NFB circuit between output selector and R22. Do you have any additional specific recommendation?

[wizardof_oz]

Tone stack wiring is very sensitive, so try moving leads around with a chopstick and looking for any that affect the tone of the squeal. Does the bright switch affect it? I agree on the heater elevation, none needed unless there is a cathode follower that runs Vh-k near spec limit.

I tried as much as I could. The bright switch has an effect, more squeal intensity in the following order: neutral position - 100pF position - 500pF position.
Which leads would have the highest chance of reducing the squeal? The cables to the pots? Could you identify any other suspects from the gut shots?

[wizardof_oz]

try to twist the wires from the phase invert to the el34 grids. bigger grid resistors downstream also help with oscillation and blocking distortion. try 47k at the el34s and 68k at v2 and v3

Thanks, I'll try over the weekend and report back.

[wizardof_oz]

in the meantime I went closer to the original circuit (C9 back to 0.1uF, took out the cut control) and tried out some measures:

• twisted the wires from the phase inverter to the el34 grids

• bigger grid resistors 39k at the power tube sockets

• grid resistors 68k at v2 and v3 (pins 2 and 7)

• adjusting the lead dress

• even tried a thin aluminum shield so the components in the chassis were completely caged

No real change, noise floor is constant (high) and I get loud squeal with high volume (12:00 position of the VOL pot, no attenuator) as soon as I play a chord or single note, which gets more intense with the bright switch activated. Basically the heights are firing the oscillation.

Interesting is the bass heavy behavior of my amp, which seems untypical for this circuit. I found only one hint here at TAG about a similar behavior, but no details why this can happen and proposed measures. Would anyone of you have a theory?

Any additional measures you can think of? Some more hints about the shielded wires in the audio path would be great, too.

[sluckey]

Disconnect the NFB wire. Any better?

[Roe]

Disconnect the NFB wire. Any better?

yes could be positive feedback

[wizardof_oz]

I already disconnected the NFB wire (R22/100k detached) as one of the first measures. Sound and playing feel changed, but squeal was still there.
Let me re-check...

[eddiep1966]

I know some of you are going to think I am nuts. I have built three Express Clones and they all squealed at higher levels until I shielded the bottom of the chassis. These circuits are very sensitive to outside noise as we all know. I use aluminum foil and lay it on a speaker cab making sure it's flat then place the chassis on it. The squeal was gone in the ones I built. Something in my shop is emitting noise and these amps will go nuts on it. Like i said you my think I'm nuts but it has always work for me.

Sorry, didn't see the part about you using a thin shield.

[wizardof_oz]

yeah, tried this already Thanks eddiep1966 for the hint, though.

Any ideas at all where the bass heavy behavior of my amp could come from? The low end is booming and really annoying...

[xtian]

Sure, three ways to trim bass.

Reduce cathode bypass caps on V1b and/or V1a from 22u (full bandwitdh) to 1uf (high-pass).

Reduce value of coupling cap C8 from 0.002 (2000p) to 470p.

Reduce value of pre-PI coupling caps from 0.1u to 0.047u.

This is general to all tube amps. I'm only just familiar with the eccentricities of the Express, so perhaps others will say best approach.

[Roe]

Any ideas at all where the bass heavy behavior of my amp could come from? The low end is booming and really annoying

then something seems wrong. recheck the wiring of the tonestack