Tube Amp Gain Stage Decay - Crackle/Fizz?

[Joncaster]

When an amp is set to edge-of-breakup or slightly overdriven, what makes the distortion fade away with a crackle/fizz type texture?

I don't mean like a problematic sound that's due to filtering or oscillation, etc. but a healthy amp that just has that ragged decay.
I hear it on a lot of amps, actually, now that I am aware of it.

In my research I came across a video of a Bartel amps prototype that doesn't have any of that, just goes from overdrive to clean smoothly.
He mentions something about an interstage transformer coupled phase inverter, but I have no idea why that would contribute to a smooth decay.
Maybe is has to do with source impedance and lack of NFB...

The amp i'm building/tweaking has a bit of this ragged decay going on, I'm trying different things to see why it happens.

Bias has an effect but more to do with whether the decay is on the fizzy or fuzzy side (cold/warm).
Keeping in mind I'm tuning the finer details through headphones (load box into a small amp with headphone out, or into an interface into DAW), so its more apparent than in the room through a cab.

Is it bias shift, the NFB knee, IMD, grid blocking, or something else?

Thanks

[teemuk]

Because interstage transformer cures it then most likely blocking distortion. Sustained overdrive will shift bias points and when overdriving stops original DC bias points are not recovered instantly. Then lowest amplitude signal portions will not be amplified.

This has a lot to do with grid conduction clipping, interstage capacitances, cathode bypass capacitances and overall RC filters involved in aforementioned.

Interstage transformer is one method to eliminate capacitive coupling between PI and grids of output tubes. This capacitive coupling is one of most common sources for overdrive induced DC bias shifting and related blocking distortion.
Alternative techniques to reduce blocking distortion in this point include setting up a different bias point closer to class-A operation (which is not always possible), adding considerable series resistance to grids (this has side effect of making grid clipping harder), or reducing coupling capacitance (which has side effect of "thinning" the overall tone).

[Joncaster]

Thanks for the reply Teemuk,

The more I listen and read, the more I think i'm actually just focused on the distortion too much...
I think what I'm hearing is literally just what a distorting tube sounds like, but I became fixated on it.
The quality of the distortion is tuneable, and I'm getting it to a place where it sounds better and more integrated into the tone (not feeling separate like a diode clipping can be).

You mentioned Class A and large grid stoppers, which is exactly what I have going on, and I'm not hearing any blocking (until I intentionally increase the bass response at max drive, but even then its musical).

Listening to a circuit on headphones, and being a mix engineer by trade, I definitely get too close to a project, i think.

One thing I will say, and I didn't think it would make such a difference, is Polypropylene coupling caps...
I changed all the coupling stages from Polyester to polyprop...crazy difference. Same values (measured pretty close)
There was an edge to the decay with the PE caps. a fuzzyness.
The PP caps have more top end, but less fizz.