C-R-C filters

[EFK]

I don;t really know where to put this, but I'm looking at it theoretically so I suppose this ought to be the place. Typically you find a single cap off a gain stage, and of course there are calculations that can be done to determine the rolloff point. However: what if you utilize a cap, a resistor to ground, then another cap? C-R-C. I realize the implications of the caps in series but I'm not sure how to figure rolloff and all that. The only place I have ever personally seen this done is in some Supro amps, but I'm wondering if it could be useful to reign in some of the gain between stages that may have too much gain. Problem is, I'm not sure other than trial/error, how to figure out rolloff points. ????

[loudthud]

There will be quite a bit of interaction unless you really load down the tube. The first cap needs to be something like 10 times the second one with 1/10 the resistor value to ground. What you end up with if you don't follow that rule is a 6dB/octave rolloff that starts early with a slower rolloff. Best way is to use a modeling program, or cut-n-try-it.

[EFK]

Thanks - but I could use some clarification if you don't mind. "Interaction" between what and what? I understand the 10X comment, and that is interesting b/c all of the Supro amps in which I have seen this implemented utilized values of .01 and .001 for the caps (the smaller being the second cap) with a 270K R to ground between them. "1/10 the resistor value:" 1/10 of what? Trying to get my head around this.

[loudthud]

A single R-C high pass filter, cap in series, resistor to ground gives a 6dB per octave rolloff (when driven by a low impedance). The -3dB point follows the 1/(2*pi*R*C) formula. If you put two if these in series, you'ld think you could get a 12dB per octave rolloff. To do that, the second RC needs to be driven by a low impedance. Without adding a buffer stage, you make the first cap 10 times bigger and the resistor 10 times smaller (keeping the RC time constant the same). If you don't do that, the second R-C doesn't behave right. It's almost like the two caps are in series, you just get 6dB to 9dB per octave rolloff after a weak start. The second R-C will raise the -3dB frequency.

Checkout this article by Nelson Pass: http://www.firstwatt.com/pdf/art_coherent_xvr.pdf

It's about active crossover design using RC filtering but to make the filters behave, each section uses progressively smaller capacitors and larger resistors.

In the Supro, what is the impedance after the .001? If it's 2.7meg, the filter should act fairly well. If it's 1meg or 500K, you'ld get a slower rolloff than expected. That's not necessarily a bad thing, it's just hard to predict without some pretty complicated math, so make life easy, use a modeling program.

A High pass of .01-270K gives a F3 of 59Hz. That's below guitar frequencies but it will make a little difference in the sound, maybe 2dB at 80Hz. The .001 makes the rolloff start a little sooner, moving the -3dB point higher to maybe 100Hz.