LTPI Grid stopper value affect on harmonics

[jjman]

On my project guitar amp I have a post LTPI master volume that allows the PI to clip while the outputs are not. ("Larmar type.") No feedback installed at this time, which wouldn't really be significant at low output settings anyway since the feedback signal would me minimal.

I've been messin with the tail and bias resistor values recently in an attempt to reduce the "fizz" I'm getting when the PI is overdriven. I employed 2 pots in line with resistors for some scoping and real-world testing and they have been resolved to 33k and 680R based on sound, clipping point, and bias voltage (1.7) results. PI outputs are balanced in strength. However, fizz reduction was not achieved.

A recent thread included a link to a guitar amp scheme showing a grid stopper of 47k on it's LTPI. I usually don't see such a grid stopper on guitar amp schemes. This got me thinking so I tossed in a pot and tested thru 250k with this variable grid stopper. I can see and hear a change after reaching around 50k then thru to the end at 250k.

The change on the scope is minimal but visible with shape/width of the (squared) wave. Max grid stopper reduces the visible clip slightly. The change in sound is with the harmonics. I think one of the harmonics in particular is being affected but the strength of the input signal seems to change how the stopper-adjuster will behave. That harmonic I hear is reduced when the stopper is increased into this upper zone. My hope is that I can arrive at a value that helps with the "fizz." I don’t know much about harmonics but have read many say that some sound good and some don’t regarding a clipping guitar tube stage.

I know there could be many variables but has anyone ever gone thru this insanity with a grid stopper's value on a guitar amp that has a LTPI that clips during low “master-volume” settings? Is it definitely particular harmonic(s) being affected or is that dependant upon other variables? Is this a case of where one should use the highest value that does not reduce treble measurably? Am I crazy to play with this?

PS node is 320V. I know the 0.47u input cap is big but reducing it to .001u did not reduce the fizz:

[SemperFi]

Hi. I played around a bit with post PI mastervolume, but hated the fizzy dist so quickly went back to the pre-PI master.
The fizz is due to a symetric clipping of the LTP. Large gridstops will create a lp-filter due to grid-plate capacitance. (u probably know that). So the result is some smoothing out of the highest frequencies, sounding nicer.
An unsymetric compression/clipping creating more even order harmonics generally gives a nicer dist, but too little of the odd harmonics and the dist gets too mushy for most rock styles.
I've wanted to build a fully balanced guitaramp (including balanced pre), but suspect it will just sound too fizzy and more like SS than tube. I might still do it tho just to verify, besides I may be surprised and learn a thing or two. Like how does such a balanced pre-
pwramp compress?
Anyways, good luck hunting harmonics and let us know how you get the post-PI master to sound.

[trobbins]

Are you saying that the change in harmonic spectrum is dominated by different harmonic levels moving relatively, but with little change in general LPF slope? Have you been able to measure the level of LPF impact using a chirp type fft spectrum sweep at different signal levels?

It would be interesting to check the change with also putting a grid-stopper in the other side, which I think would generally lower even harmonics.

Different 12A-7 valve types have different grid current-voltage curves, which may change the harmonic character significantly with your available level of overdrive.

[jjman]

I’ve definitely heard fantastic sounding preamp distortion from such a MV but the internets told me to use a post PI type for better sound. Maybe that’s my next experiment with adding a pot.

I don’t have equipment to sweep or do any dynamic type quantification. I changed to a 1meg pot after referring to the Valve Wizard's site. He mentions possible benefits of big stoppers with an LTPI although that page is mostly on the cathodine. Check out the banner going across the top of the screen on this page:

http://www.freewebs.com/valvewizard/cathodyne.html

Going close to 1 meg is more noticeable and it seems to be that the harmonic being affected is one octave up Once the stopper becomes real big, that harmonic dissipates. But I think I still hear higher harmonics and wonder if this is merely a LPF result or something else. The wife came home and she is not interested in hearing test signals being modulated for more than a few minutes. You know how they can be with that! So more play tomorrow.

I need to systematically switch between a 12ax7 and a 12at7 to see what I get there. Some time ago I thought a 12at7 might reduce fizz under the assumption that it’s higher max current rating was relevant. A real world test is needed on that.

I’ve been testing around 400hz so I can hear the harmonics better. I will also move the frequency up to see if/how much higher frequencies are affected. It is a guitar amp so I won’t be going to high for that. It’s now easy to hear what’s happening on a test signal but whether it’s an improvement or not for guitar isn’t clear to me yet. Even if it’s merely a LPF, it may help me.

[jjman]

I changed the test pot to an input volume control to the PI. Backing this down and turning the preamp's volume control up let's me hear only the clipping of the 3rd preamp stage. I used my guitar and was disappointed to still hear the fizz. So I'm now thinking the previous adjustable-stopper test, to the PI's input grid, was merely acting as a LPF that can mask the fizz being generated before it. I guess it's actually some sort of oscillation in the preamp. It sounds somewhat "swirly" as it decays. It appears when the preamp is at higher settings.

I don't know how to look for this on my scope. I'm ignorant regarding that. I’ve never seen any “ringing” on the clipped or unclipped sine waves in this amp, but I don’t know how meaningful that is.

How to look for possible oscillation with a scope? Do I set my scope a particular way? Do I use a square wave? Are there any sites on this kind of thing?