Q about netagive feedback capacitors in condensers

Why does a small pF capsule between the drain and gate (or plate and grid) like in the KM84 simply reduce the gain rather than reducing the high frequencies first?

Is it that the impedance to ground is so high that the cap holds a positive voltage on one side, raising the gate's voltage? This is the only thing that I can think of that accounts for why a higher capacitance reduces the gain further. Is the source bypass capacitor necessary to make this do its thing?

The main reason I'm asking is that this capacitor has a different effect in e.g. guitar circuits, which still have high input impedance (>=1M) but obviously not 300M (or whatever the gate/1G is). There it cuts high frequencies first, but also I've had situations where it creates a positive feedback loop when it gets too large AND the source bypass capacitor was also included. (Both with FETs and MOSFETs.)

It has this effect because the impedance on the input of the amp is a capacitor too. The input impedance and the feedback impedance have the same frequency dependence, so HF rolloff effects are essentially cancelled out.

Check this out:
https://groups.google.com/forum/#!topic/rec.audio.pro/jQC3wLhKWo0

...C4 (8 pF) is placed between tube’s AC 701 plate and grid. So it is
parallel to the tube’s own capacitance Cag. This capacitance according to
the tube’s manufacturer’s literature is 2.2 pF. As a result we get 10.2 pF.
At stage’s gain (20 dB)  - in other words – 10 times (it is a typical
volume) we get 102 pF (10.2 pF x 10 = 102 pF) known as Miller’s capacitance
that is connected between tube’s grid and cathode...

Not the most clear explanation but it hints at totalling capacitance between anode & grid with the Miller capacitance.

Oh, okay. That's clear enough.

This also helps explains why parallel tubes or FETs don't exactly double the gain -- because they also double the capacitance.

http://www.audioimprov.com/AudioImprov/Mics/Entries/2015/4/23_Basic_FET_Microphone_Circuits.html

Scroll down to the KM84 circuit analysis

Khron said:

http://www.audioimprov.com/AudioImprov/Mics/Entries/2015/4/23_Basic_FET_Microphone_Circuits.html

Scroll down to the KM84 circuit analysis

Click to expand...

I actually e-mail him about that -- he said he was wrong about the gain reduction calculation and he has a later post with actual measurements but without an explanation of exactly how it works.

midwayfair said:

Why does a small pF capsule between the drain and gate (or plate and grid) like in the KM84 simply reduce the gain rather than reducing the high frequencies first?

Click to expand...

It's a capacitive voltage divider into the parallel combination of the capsule and the FET's gate->source capacitance.  In short, the divider's frequency terms cancel.  I derive it here if you aren't too bored by the math.

By the equation in that thread, if you increase the value of the feedback cap, then gain will drop across the pass band.  That's why adding in the 15pF cap adds as a pad, not a LPF.