feedback resistor and cap bypass to eliminate ice pick

[daz]

Not really all that concerned because the amp isn't harsh, but i thought it would be a good idea for when i want a brighter sound to cut any ice pick that might occur with higher top end. Anyways, i saw this at that robrobinette site and he shows a 27k FB R and a 100pf across It. i have a 47k R and it's in series with a pot for variable NFB. When i added a 100pf i added o from the pot input to the far side of the 47k. The cap is across the 47k and if the pot is turned to add more resistance it's also across that, or in other words across the entire resistance wherever the pot is set. I mention that just so u know it's right and not a wiring fault fault. The question is, i heard no difference even after trying it on a switch for instant A/B. Anyone tried this and have any idea why it does nothing, at least nothing i can definitely hear w/o the possibility of placebo effect. (sometimes i think i may hear the slightest difference but not sure) Could the output tube or transformer or anything else have a effect on how much or little it's effect is?

[catalin gramada]

100p cap is there to stabilize the nfb meant subprime oscillations at hf due to the phase shift as time nfb is used over multiples poles, and have nothing with the tone. Download a reactance calculator and see what is doing in respect with you nfb resistor. You may want to use more 1nF to be effective.

https://www.angelfire.com/electronic...Smith-NFB.html

[daz]

I see. His page sounded like it was a HF cut. No matter, just thought i'd try it and when i heard no difference (even a 250pf doesn't seem to matter) i had to ask. Thanks

[catalin gramada]

Also note, using nfb you extend the HF reponse. Is more easy to tailor the upper high using a presence circuit like Marshall did. Selecting the right cap over the shunt you may do a simple effective high cut in you range of interest. A series nfb cap will not work in direction you want. But you need to use a pot instead a fixed resistor (if want a variable) as shunt and to determine cap value for the high cut pole using a RC calculator.
Also you may try to use a fizz cap between PI outputs to tame some hights. Instal directly between plates if you have a high voltage one or after coupling caps, it will act as same 100-500 pf. Do a try.

[Helmholtz]

A Selecting the right cap over the shunt you may do a simple effective high cut in you range of interest. A series nfb cap will not work in direction you want.

A cap across the NFB shunt lowers HF feedback and thus increases HF response.
A presence circuit cannot cut treble, only allows to boost high mids and treble.

A small cap across the NFB series resistor is often used to improve stabiliity by lead compensation.
NFB phase compensation can be tricky. The OT has a strong influence on stability and different OTs may require different compensation measures.
Also an amp can be stable with a resistive load but oscillate with a speaker. And vice versa.

My original Marshall JTM50 needs a 22p cap across the 27k NFB resistor. With lower or larger values it likes to oscillate between 60kHz and 110kHz at certain conditions.

[catalin gramada]

A cap across the NFB shunt lowers HF feedback and thus increases HF response.
A presence circuit cannot cut treble, only allows to boost high mids and treble.

A small cap across the NFB series resistor is often used to improve stabiliity by lead compensation.
NFB phase compensation can be tricky. The OT has a strong influence on stability and different OTs may require different compensation measures.
Also an amp can be stable with a resistive load but oscillate with a speaker. And vice versa.

My original Marshall JTM50 needs a 22p cap across the 27k NFB resistor. With lowers or larger values it likes to oscillate between 60kHz and 110kHz at certain conditions.

Oh yes, that-s true. I was thinking just how be effective in managing the amount of highs for a certain range. For a effective high cut the "resonance control" with a fizzy cap between phase inverter outputs may fit better

[J M Fahey]

Those who care about fizz aren´t playing loud enough

[Helmholtz]

No idea what fizz cap means.

[catalin gramada]

Is a high freq suppressor. That;

[Chuck H]

No idea what fizz cap means.

...

[Helmholtz]

Is a high freq suppressor. That;

That 47p cap can't have an audible effect on frequency response.
It produces a low pass effect with a corner frequency around 25kHz in the open loop response.
Multiply that corner frequency with the NFB factor for closed loop response (before clipping).

The purpose of the cap is to generate a phase lag to improve the stability margin.
So without the cap the amp might oscillate.

[catalin gramada]

Nobody said 47pF is. But fizz cap may do an effective high cut control like vox did. 4.7nF is ok ?

[Helmholtz]

Nobody said 47pF is.

I thought the 47p cap should be an example for a fizz cap??
Doesn't "fizz" imply a certain effect on sound?

But fizz cap may do an effective high cut control like vox did. 4.7nF is ok ?

Sure, but don't forget that those Vox amps didn't use global NFB.

[pdf64]

Ultrasonic oscillation can cause a fizzy effect on the audio signal.

[Helmholtz]

That makes sense, thanks.
(Still don't like the term )