Tweed BluesMeister needs a tweak

[acheld]

Well now, more than a tweak.    I built this amp more than a year ago, but had to put it on the shelf.  I've recently started playing it, and I'm not so happy with it.

Problems:  1) This is a noisy amp.  When using only Stage 1, it's got a sort of ticking and a high frequency hiss when it is turned up past "4" on the volume dial.  When turned on, the Overdrive section (Tube 2) and/or the pre-amp boost just make this all the worse.  2) The overdrive section is not functioning as planned.

Today, I got off of my behind and put it on the bench. 

My question for today is in regard to the two scope images.  The test signal was input through the input jack at 30mV, with the OD section completely bypassed.  The images are taken from the output across an 8 ohm dummy load.

There is a clear notch in the waveform both on the way up and on the way down.  When I changed frequencies, I got a similar picture -- eg, this is not "noise" but unintended distortion.    Any ideas where or why this might be happening?   Initially I was thinking crossover distortion, but I don't think so.

Apologies in advance for the schematic, which I could not for the life of me fit on one page.   For the purposes of this particular question, the signal goes through page 1, skips page 2, and goes through the output section on page 3.

This amp was inspired by da Geezer, and Tubenit . . . 

[AlNewman]

Do you have a schematic of a working amp that you started with?  Can you trace the signal back to see where you lose a clean sine wave?

[Helmholtz]

To me it looks like the amp is at the edge of HF oscillation.
What happens if you disconnect the NFB from the output?

[acheld]

This amp came about when I was perusing work done by Da Geezer and by Tubenit on this forum. And I can't recall the exact links, but there were quite a few.  You can start by reading https://el34world.com/Forum/index.php?topic=32419.msg359347#msg359347   chatting with Tubenit and pdf64.   Going back to that post,  Tubenit is saying that the OD in this style of amp sounds harsh, and there is no question that holds true for this amp, probably the main reason I put it back on the bench.

I had never seen notches like this on any amp I've built, so I panicked.  I should have traced back, and then posted. 

I'm thinking an unbalanced PI or bias issues on the 5881's, or both.  We will see!  (I'm usually wrong with my predictions . . .).   Definitely oscillation is in the differential diagnosis.   Good idea about removing the NFB.

[Helmholtz]

Those notches are actually single cycles of HF oscillation. Might be triggered by the class A to class B transition.
This could mean the amp is conditionally unstable.
A square wave signal would show better.

Is the cathode cap good?

[tubenit]

I found a soundclick with a recording of the Tweed BluezMeister I built yrs ago.


https://soundclick.com/song/12803806


With respect, Tubenit

[acheld]

And the answer is none of the above!  I probed the first triode, and found the notches, and then probed the input signal and saw very subtle notching.   I had been using the signal generator from a Picoscope module -- at low amplitude and low frequency, the DAC within the system is not able to function correctly, apparently a known issue.    I tried to capture the input signal waveform, but the notching is so subtle as to be tough to see in all the noise, so I didn't post an image.   Apologies for the excitement, but a good reminder that test equipment has its own limitations.

Tubenit: thanks so much for the sound clip (and your great ideas over the years). 

Now back to the amp.  I tried the sig gen on my Hantek o'scope, but it is crap.    I do have an old Heathkit signal generator that I rebuilt years ago that should work ok.

As an aside regarding oscilloscope selection, I'd recommend either cheaper FNIRSI or more expensive Rigol units.

[Helmholtz]

Oh well,
so just a bad input signal and garbage in -> garbage out.