Simple 6dB/octave variable high pass filter design (CR)

[XAXAU]

Hi guys, first post here! Sorry if I´m asking about simple things but I´m just getting into electronics and pedals!

I´m a bit confused about variable 1st order high pass filters. If a low pass filters´fC is 2Hz/1.98943Hz (100k pot and 800nF cap) then when you lower the resistance fully the filter barely touches the highest frequencies, right? 1 Ohm = 198943Hz.

But if a high pass filters´fC is 2Hz then it will chip off some lows when engaged, right? So should it be lower than that? 0,2Hz or even 0,02Hz?

What I wanna build is a pedal for my synths, taking the headphone output and put it through a LPF, a HPF and an attenuator and feed it back into the synths´audio input overloading the filter for some distortion.

Cheers!

[R.G.]

All RC filters have some effect beyond their nominal cutoff frequency.

A single RC will have a characteristic frequency of F = 1/(2*pi*R*C) no matter what configuration it's in. There are four: two parallel and two series, one high pass and one low pass in each.

Think about the series version first. They're easier, as the parallel versions work more on current than voltage. A high pass, with a series C and a shunt R after the C will have a "cutoff" frequency of F = 1/(2*pi*R*C). However, that's really the frequency where the power is down by half, or for voltage measurements, -6db. But the rolloff above the "cutoff" frequency actually started down; this is noticeable at 10x the cutoff frequency, and gets down to -6db at the cutoff. From there, it is asymptotic to -6db/decade.

Swap the R and C, so you have a series R, shunt C. The "cutoff" frequency is the same, F = 1/(2*pi*R*C). But now the response is no loss at DC, and it starts to roll off noticeably at 1/10 of the cutoff frequency. It's down 6b at the "cutoff" and then is asymptotic to -6db/decade on the high end.

And you asked:

But if a high pass filters´fC is 2Hz then it will chip off some lows when engaged, right? So should it be lower than that? 0,2Hz or even 0,02Hz?

In reality, BOTH high and low pass versions of the single RC filter will chip off some of what it ought to be passing while frequencies are still in the pass band. It's one of those unfortunate things Mother Nature insists on being true. And the imperfect nature of the frequencies passed and rejected is what caused all the research and development into higher order and just downright fancier filters that continues to today.

What you really want is a copy of "The Active Filter Cookbook" by Don Lancaster. Used copies are pretty cheap on Amazon and other places.

[Quackzed]

i'd think .2hz is plenty low, 1hz is 1 cycle per second, too low to hear... dont worry about it, your synth dont go that low either, and if it does, the amp it uses doesnt... human hearing starts @ 15 hertz fwiw... well above .2hz!

[XAXAU]

Thanks guys!

I thought it was -3dB at the fC, -6dB per octave and -20dB per decade? For example a low pass´fC is 2Hz (-3dB), -20dB at 20Hz, -40dB at 200Hz and so on?

[R.G.]

I thought it was -3dB at the fC, -6dB per octave and -20dB per decade? For example a low pass´fC is 2Hz (-3dB), -20dB at 20Hz, -40dB at 200Hz and so on?

Fc is -3db for power in a matched load, -6db if you're doing voltage ratios. -6db voltage is -3db power in a matched load.

-6dbvolts/octave is the same as -20db/decade.

db is a ratio. It's 10*log (Pout/Pin) and 20*log(Vout/Vin).