NFB resistor and Presence pot...

[diddymix]

Hi people,

I just have a quick question regarding global NFB loops and the presence pot, as seen on many amps. Basically I was thinking that in regard to the feedback resistor and the presence pot.. If one was to reduce the feecback by increasing the resistor... would one have to change the value of the presence pot for it to still be affective and audible??
Basically I have copied the Dumble 102 NFB and presence on my amp, I have yet to implement the presence... however because my amp is 40W EL84, I ended up using less feedback than the dumble (4.7K), I found on my amp that 10 - 15K was about right and still alot of feedback.... but now that Ive done that am worried that the presence pot (2KB) wont be affective.. is this correct that if you change the feedback amount the presence pot + cap should change also?

Thanks!

[martin manning]

The feedback resistor has a small effect, but the frequency response of the presence is dominated by the values of the cap and the resistance going to ground, in other words, the other half of the FB voltage divider.

[roberto]

Well the feedback resistor has a great effect, because if you lower it (keeping everything else constant) you'll increase the gain difference between the frequencies that are full fed-back and the frequencies grounded, so not fed-back.

Then it's just a ratio, so different combination of values give the same results, but it's not true that the feedback value has a small effect per se.

[martin manning]

Perhaps I should have said that the frequency response of the presence control, and by that I mean the lower frequency knee of the shelf, is not much affected by the feedback resistor. Further, the difference in the feedback signal (in dB) between low and high frequencies when presence is full on remains very nearly the same as the feed back resistor is varied.

[roberto]

Further, the difference in the feedback signal (in dB) between low and high frequencies when presence is full on remains very nearly the same as the feed back resistor is varied.

Ciao Martin,
can you please explain that better mathematically?

By my experience and by my logic the behaviour of the classic 1959 Marshall-like presence pot i the following:

• when you increase the value of the feedback resistor, the effect of the presence pot is diminished: logical if you think that without NFB you would have almost no effect, and by the fact that when you install a MVPPI the effect of the presence collapses when the PI starts to overdrive abruptly (so its effective gain is reduced by clipping the signal). So the value of the NFB resistor (as a voltage divider with the impedance to ground through the presence pot and cap) sets the gain at low frequencies, while the gain at higher frequencies is set by the PI + output tubes + OT.
• when you increase the cap the frequency of the presence goes down

[diddymix]

Well the feedback resistor has a great effect, because if you lower it (keeping everything else constant) you'll increase the gain difference between the frequencies that are full fed-back and the frequencies grounded, so not fed-back.

Then it's just a ratio, so different combination of values give the same results, but it's not true that the feedback value has a small effect per se.

Thanks Roberto, thank you Martin. OK so it seems that perhaps if I make an adjustment (everything else the same) to the feedback resistor.. by increasing it (to lessen the feedback)... than the effect of the presence control will decrease?? What I mean is that you will not hear it as much... So in order to compensate for this, would I have to increase the value of the presence pot?? (currently 2KB) to perhaps 5, 10, 20K for example?? To be clear... I dont want to adjust the frequency response or 'eq' of the presence... I just want it to remain effective in 'volume' after lessening feedback. Thanks again guys, just abit confused about this but I think im on the right lines..

[martin manning]

For DC (presence fully off), the attenuation of the signal from the OT is Rs/(Rfb + Rs), where Rs is the shunt resistance and Rfb is the series resistance, and the low frequency attenuation can be adjusted using Rfb.
With presence fully on, the low frequency knee is ~1/(2π((Rfb//Rs)Cp) where Cp is the presence cap. Since Rfb >> Rs, the knee frequency is nearly constant as Rfb is adjusted over some range.
The slope from the knee is -6dB/octave with increasing frequency, so for a typical variation in Rfb, the attenuation at frequencies above the knee is proportionally the same.
Usually there is more than enough range on the presence pot- the presence effect doesn't usually come on until the pot is nearly all the way up.

[roberto]

Ciao Martin,

I agree on the 6 dB/octave, and the variation in Rnfb changes the gain at low frequency, so the difference in gain between bypassed and unbypassed frequencies.
This makes the presence work between higher gain differences, so it extends both the width of the gain (so being more effective) and widens its frequency range (the difference between the lower and upper knee, that is the gain difference divided by 6 dB/octave). Isn't it?

[martin manning]

Roberto, yes when presence is all the way up, it's a single pole LPF, and above some frequency the gain will top out at its open loop level, and that will occur at an increasing frequency as the FB is increased. So I agree, the difference in HF vs LF gain with presence full up is larger when FB is increased. You would have to turn the presence down a bit (making a two pole shelf LPF) to bring the HF gain back down to the same relative boost level.

Diddymix, your conclusion is correct, but you can't do anything about it. Reducing the FB always reduces the maximum gain difference from LF to HF. The frequency at which the treble boost begins to come in stays the same, and you can still control that with the presence cap value.

[roberto]

Roberto, yes when presence is all the way up, it's a single pole LPF, and above some frequency the gain will top out at its open loop level, and that will occur at an increasing frequency as the FB is increased. So I agree, the difference in HF vs LF gain with presence full up is larger when FB is increased.

Thanks Martin, that was what I was referring to when I said that the NFB has a big impact on the way the presence reacts. Its depth and spectral frequency widens when you have more feedback (so when you lower Rnfb keeping everything else the same).