Fuzz/Phase shifting effect (muskito flying by) in the mids

[5150loveeddie]

M. Aiken thx for showing up around here, this is fantastic info...

I didn't try it yet but would increasing bias till the cross over notch dissapear on you sine wave (scope) would cure this also? I run that amp with a good strong pair of xf2 so I guess they can take it? I read that some tubes still had a notch at 100% plate dissipation (modern tubes) I guess this is not good at all, but if you do have high quality tubes, could you get over the notch near 60-70%??

thx again and you are welcome to stick around us anytimes!!

[raiken]

M. Aiken thx for showing up around here, this is fantastic info...

I didn't try it yet but would increasing bias till the cross over notch dissapear on you sine wave (scope) would cure this also? I run that amp with a good strong pair of xf2 so I guess they can take it? I read that some tubes still had a notch at 100% plate dissipation (modern tubes) I guess this is not good at all, but if you do have high quality tubes, could you get over the notch near 60-70%??

thx again and you are welcome to stick around us anytimes!!

No, it usually won't help because you can't lower the bias voltage enough to get rid of the crossover distortion when you are driving the hell out of the output tubes, because the bias shift is too great. If you could get rid of all of it, your tubes would likely be redplating severly. The problem is that it takes only as much peak voltage as your bias is set to in order to hit clipping, but the phase inverter can continue to put out far more.

For example, if you have your bias voltage set to -35V, you only need a 70V peak-to-peak signal to drive the output stage into clipping. If the PI output keeps increasing as you drive the amp harder, the top of the waveform will clamp at a bit above zero volts, and the signal will be "pushed down" from there as it increases, so if you maxed out at say, 100V p-p, you'd be shifted 15V down from your idle bias point, because the nominal zero-signal bias point would be at -35V, but with the 100V p-p signal clamped at 0V at the top, the "center" of the waveform would be at -50V, which is now 15V lower than you need it to be for minimum crossover distortion.

If you raised your idle bias voltage up by 15V to -20V, your idle current would go through the roof and your tubes would burn up.

The only way to fix it is to either limit the drive signal or clip it, or use DC coupling to convert the output stage into a class AB2 stage.

Randall Aiken

[Billy Batz]

Will any of those fixes not change the amps' response? If not, and depending on how it comes out, I may leave it or learn to live with the noise the way I did with ghosting. Just another vintage amp noise. Its like that blocking distortion you talked about- the splatty sound. That drives me nuts when I get too focused on it but it doesnt bother me nearly as much as this does. Will decreasing the idle bias voltage increase blocking distortion? It sounds like it the more I go over 70% the high end starts to get real splatty like a fuzz with a dying battery but much harsher.

[5150loveeddie]

Proper location of grid stoppers helps, and if the amp has global negative feedback, it can usually help to put a 33K or 47K stopper right on the input side grid of the phase inverter, as it is usually the PI/output tube/OT/speaker loop that oscillates.
Randall Aiken

At both grids of the PI or only on side A (pin2)???

Thx!

[novosibir]

Proper location of grid stoppers helps, and if the amp has global negative feedback, it can usually help to put a 33K or 47K stopper right on the input side grid of the phase inverter, as it is usually the PI/output tube/OT/speaker loop that oscillates.

At both grids of the PI or only on side A (pin2)???

As you can read in Randall Aiken's post - ONLY on the input side.

In my DINO amps I'm using even a 100K grid stopper at this spot, to obtain a slightly tone shape/highs roll off additional to the (possible) RF suppression purpose (seen on the photo below).

With lower values of grid stopper like 33K or a 47K, as Randall has suggested, you barely will hear an additional highs roll off - maybe with dog ears

Larry

[raiken]

In my DINO amps I'm using even a 100K grid stopper at this spot, to obtain a slightly tone shape/highs roll off additional to the (possible) RF suppression purpose (seen on the photo below).

Larry

Nice work, Larry!

RA

[novosibir]

Nice work, Larry!

RA

Thanks for your kind words, Randall!

Larry

[5150loveeddie]

Proper location of grid stoppers helps, and if the amp has global negative feedback, it can usually help to put a 33K or 47K stopper right on the input side grid of the phase inverter, as it is usually the PI/output tube/OT/speaker loop that oscillates.

At both grids of the PI or only on side A (pin2)???

As you can read in Randall Aiken's post - ONLY on the input side.

In my DINO amps I'm using even a 100K grid stopper at this spot, to obtain a slightly tone shape/highs roll off additional to the (possible) RF suppression purpose (seen on the photo below).

With lower values of grid stopper like 33K or a 47K, as Randall has suggested, you barely will hear an additional highs roll off - maybe with dog ears

Larry

I tried a 33k last night (in the PI input), will have to play the amp more and see if improuvement occured, it did something I like, gives a hair more warmth..

[Froumy]

I'd try the attenuating the PI /tail resistor first, only because I don't have the diodes laying around. Seems like a few guys are enjoying an improvement in tone with the extra resistance, as well. The Zener clipping option is still the most intriguing to me. It's the surest way of eliminating the crossover distortion(minus adding a tube). I haven't noticed it(yet), but for you guys that are hearing it in your sleep? I really hope you try it, and report back. Especially if you don't have a scope. I'm thinking this is a case of "diode" not being a dirty word, but the exact opposite.

Billy Batz, sounds like you feel you may be getting "splatty" blocking distortion, as well? Either that, or you're grouping them together like I did in my earlier post. Might be important for you to experiment with the diodes, for proof. The solutions for each issue can be different. A Marshall style build should have a faster RC constant, and less prone to blocking distortion. From what I've read, lowering the PI caps to .047 or lower can help with blocking. More negative feedback can help reduce crossover distortion, but less feed back can help reduce blocking distortion. If blocking distortion is part of it, I'm thinking that lower PI caps, in conjunction with a resonance circuit, could give you a similiar tone with less negative feedback to help eliminate blocking distortion. Those diodes should take care of the crossover. Current understanding anyway.....

Dan, you have a build with a Marstran 103? What spec? If I have the time, I'd like to do the same with mine and see what it looks like on the scope, to help improve my understanding, and join this hellish party.

Stephane, what's the final verdict on the 33k PI resistors? Have you tried Larry's 100k solution? Perfect opportunity to put your scope to the test(still learning on mine, by the way). Try it!

In any case, thank goodness this is getting somewhere. Ultra hat tip to Mr. Aiken for checking in to shed more light, with a proper scientific explanation! The end may very well be in sight!

-Bob

[Billy Batz]

Billy Batz, sounds like you feel you may be getting "splatty" blocking distortion, as well? Either that, or you're grouping them together like I did in my earlier post. Might be important for you to experiment with the diodes, for proof. The solutions for each issue can be different. A Marshall style build should have a faster RC constant, and less prone to blocking distortion. From what I've read, lowering the PI caps to .047 or lower can help with blocking. More negative feedback can help reduce crossover distortion, but less feed back can help reduce blocking distortion. If blocking distortion is part of it, I'm thinking that lower PI caps, in conjunction with a resonance circuit, could give you a similiar tone with less negative feedback to help eliminate blocking distortion. Those diodes should take care of the crossover. Current understanding anyway.....

I dont know I thought that blocking distortion sound is pretty common in marshalls. I just am not that concerned with that. Its not as perceptible I dont think.

Im definitely trying all fixes put forth.

[Froumy]

Forgive my exuberance. Guess I'm a little excited about all the possible cures. I may be able to revisit my previous high gain/ mid pushed spec, if any of this works. I'm thinking that one of these distortions may have been what was bugging me. Really looking forward to the report card...

Has anyone implemented a current mirror style circuit to eliminate crossover, and would this be an acceptable design in a guitar amp?

[5150loveeddie]

Ok I tried the PI input resistors 33k, 47k, up to 100k and no joy, still does it, doesn't change really the swril a bit, but it warms up the amp a little by cutting extra highs in there I guess.

That resistor act like a high end filter I guess..

[Ricky Lee]

I tried the 100k on v3 awhile back and it seemed to stiffen the amp more than remove highs, didn't care for that so I removed it.

[5150loveeddie]

I tried the 100k on v3 awhile back and it seemed to stiffen the amp more than remove highs, didn't care for that so I removed it.

I did also, I prefer the remaining highs that no resistors gives finaly......

[Froumy]

Were those grid stoppers supposed to reduce possible PO, or lower the p-p voltage? If I understand correctly, the goal is to reduce the p-p voltage to less than double the negative bias voltage. Something the diode trick should accomplish. Have you measured your p-p voltage? I'm guessing attenuating more with the Tail resistor might accomplish this. If you're almost there, I'm guessing bumping up the bias voltage just a little, might get you closer. Hope the tail resistor or diodes work.