Odd behavior regarding variable NFB, presence, and volume in a Plexi clone

[jarrodthebobo]

Hello everyone!

My main "futs around" toy based off a plexi/friedman idea is giving me a rather strange issue/unexpected interaction.

I recently started messing around with a NFB control using a 250k pot acting as a variable resistor and a 27k resistor in series with it. Turning the knob down to its lowest setting (27k alone essentially) leads to LOUDER signal than when the pot is turned up all the way (near 300k of NFB resistance) which honestly doesn't make sense at all to me considering higher levels of NFB are typically associated with gain REDUCTION, not increase.

Additionally, I've pretty much confirmed to myself that Ive at least wired the output transformer correctly and am not accidently introducing positive feedback into the loop; when the NFB resistance is at the lower end of its rotation, the presence knob is functional and is audibly changing the higher frequencies, but when the NFB resistance is turned UP the presence knob ceases to really make a noticeable difference which again, checks out with what I know about the function of the presence control and its interactions with NFB.

What exactly could be going on here to produce this oddity? Less NFB on this amp appears to DECREASE gain here and very audibly so, which again doesn't really appear to make too much sense.

For further context, the presence control is the typical 5k .1uf presence control attached to the 10k tail resistor of the PI, and reading the NFB resistor/pot with the presence knob is circuit drops the resistance across the pot/resistor to around 4.7k, which Im assuming is expected.

There is also a LarMar style ppimv installed which, expectedly so, displays lower useful presence/NFB settings with the volume turned down and more obvious effects with the volume turned up.

Could I possibly still have something wired wrong?

Thanks for anyone who can help scratch this brain itch!

[martin manning]

Any chance that the reduction in NFB allows an ultrasonic oscillation to take over and reduce the output?

[GAStan]

I'm not familiar with this amp. Do you by chance have a schematic available to post? Pictures would also allow us to help determine if your wiring is correct.

Have you verified the potentiometer high/low resistance is actually doing what you describe with a multimeter? I ask this because I recently wired a volume pot backwards

[pdf64]

People sometimes flip the signal polarity when fitting a LarMar, they don’t realise that the polarity has to be maintained.

If the power amp gain reduces when the feedback loop is opened, then the feedback is positive.

[Stevem]

You don’t even need a scope to check for oscillations of a magnitude high enough to be bringing down signal level, just hook up a voltmeter set for ac across the test speaker.

And just for the record you should be testing with a speaker and not a load resistor as sometimes a amp will only oscillate with a inductive load on it.

[jarrodthebobo]

To answer some questions:

I'm pretty sure the MV is wired correctly as it functions as expected and there are no squeals or obvious audible oscillations. Same goes for the presence knob, but who knows I could just somehow have the world's most stable amp with a PFB loop haha!

I'll be at work for the next 15 or so ish hours but will try and scope out the amp for hidden oscillations when I can.

The build was based on a jtm45 turret board by KLDguitars but is heavily modified; the preamp through PI are essentially a 1968 superlead style circuit with a switchable gainstage at the front like a jcm800.

The power amp is a cathode biased PA because I've liked what I've heard in the Trainwreck builds that utilize that topology. The cathode bias cap is a pretty chunky 4700uf cap to try and stiffen the response to be more similar to fixed bias at times could this be affecting the NFB?

[bepone]

What exactly could be going on here to produce this oddity? Less NFB on this amp appears to DECREASE gain here and very audibly so, which again doesn't really appear to make too much sense.

Thanks for anyone who can help scratch this brain itch!

NFB- when lowering the resistor, amplitude is dropping, no other option

To test the amp, remove potentiometer (bypass), lift NFB, measure the output

Connect the NFB 27k, measure the output, if you have increase in amplitude, yu have positive feedback and you need to swap the wires on the output socket (a-a or g1-g1)

[jarrodthebobo]

So the mystery deepens!

I tried swapping the ot leads and immediately the amp began to howl! So it isn't that...

Then after futsing around a bit, I started to notice that if the tone knobs were rolled off to zero, if the presence knob isn't at full, a ground loop buzz became very audible. At other varying TMB settings, the tone becomes rather scratchy and farty sounding if the presence is turned down; only sounding normal again with presence maxed!

Playing around even FURTHER; if I switch OFF the cathode bypass cap of the 3rd stage (820ohm cathode resistor, 100k plate resistor, ie normal plexi 3rd stage) the amp begins to behave even MORE strangly; sounding strangled off and very splatty like a fuzz pedal with a low battery...

Playing even FURTHER I started moving around grounds trying to see If I could fix the odd behavior of this amp... when a loud buzz started to take over and chopsticking around wouldn't alter the sound... and THEN the cathode bypass cap for the power amp exploded!

...cleaning out the amp, the hum is now completely gone (and I haven't replaced the cap yet, just clipped it out), but the odd presence, tone control, and v3 cathode behavior remains.

I'm at my wits end here! I've tried wiring the presence knob in a multitude of different ways; even taking it out of the circuit... but the tone controls still behave very oddly and induce what appears to be some sort of parasitic oscillation if the knobs are turned anywhere under 2. Strangely enough, cranking all 3 tone knobs essentially alleviates the issue and the amp sounds great... so it HAS to be a grounding issue around the TMB circuit, right...? But that doesn't explain the behavior that occurs when removing the v3 bypass cap either....

I've been working on this thing since I got home, and I need a nap

[pdf64]

I suspect that the amp is oscillating; the hum is indicating that, you’ll probably find that the cathode voltage is increasing and HT voltages are reducing alongside it.

It may be that the OT is very low spec, has too much phase shift to accommodate a feedback loop.

But it also may be that the amp is unstable even when open loop, eg due to non ideal layout / lead dress. The high idle current of cathode bias means the output valves will have higher gain than normal.

[bepone]

something is wrongly connected, or component faulty, t is is simple amp, need to reckeck all from zero

[LOUDthud]

You didn't mention which impedance tap you used to obtain the feedback. 27K seems too low unless you have a 2 Ohm tap.

If you have a two trace scope, compare the phase of the input signal at the PI to the phase at the secondary of the output transformer. They should be in-phase.

[jarrodthebobo]

UPDATE: The LARGEST issue has no been mostly solved: my shieled input wires for v1b and v1a had initially been grounded at the filter cap grounds... moving the shield ground to the input jacks has completely eliminated most of the strange oscillations and odd interactions between the tone controls and presence.

However, I do still have one very odd issue: if the presence knob is not cranked to maximum (or slightly below maxium), I have some pretty obnoxious 120hz buzz. Additionally, switching in the additional gainstage in front of the plexi topology leads to some strange oscillations unless the presence knob is at MAX. Ive determined that my OT leads are not backwards, and also confirmed that my presence pot is not backwards (and have even rewired the presence pot to the modern 25k pot with 4.7k parallel resistor). Disconnecting the presence pot ground from the circuit and attempting to ground it at different spots has lead to changes in the buzzing tone wise, but never gets rid of it. Disconnecting the ground COMPLETLY however still allows the buzzing to stop if the presence knob is at maximum again, which is truly strange. Ive tried different capacitors and each capacitor has the same problem.

I've even taken to shielding essentially every wire in the 1st preamp stage, but the issue still remains. My best guess is some sort of ground loop; and I've tried using both a star grounding scheme AND larry grounding scheme with no noticeable change.

Really quite lost here

[jarrodthebobo]

You didn't mention which impedance tap you used to obtain the feedback. 27K seems too low unless you have a 2 Ohm tap.

If you have a two trace scope, compare the phase of the input signal at the PI to the phase at the secondary of the output transformer. They should be in-phase.

It's off of the 8ohm tap; the 27k resistor is low due to the series potentiometer used to change the nfb. That 27k value DOES show up on some plexi variant circuits however!

[jarrodthebobo]

something is wrongly connected, or component faulty, t is is simple amp, need to reckeck all from zero

If I was a sane person, I'd do just that . Unfortunately for me, half of the fun in tinkering is the chase!

[bepone]

ok lets see how did you wire that.. all wires and components are acting like antenna/transmitters in tube amp circuits,

so probably wrong speaker grounding point and some layout issues (rectification wires close to something)

for fast check outptut stage - pull all the preamp tubes, leave only PI and output + rectifier, and comform that amp is dead silent (almost dead , something can be hear but little)