Systech Overdrive Bandpass filter Q

[John Lyons]

How can I go about setting the Q or altering the Peak of this filter?
I have it set up to bypass the filter with a pot set as variable mixing resistors
before and after the filter, which works but I would like to alter the
bandwidth as well. 

[Rob Strand]

It's the Deliyannis/Friend bandpass filter.  Plenty of info on line.

The Q and gain are linked but by adjusting the resistor R7 the gain remains fixed and the Q and frequency become linked. When R7 is changed higher frequency = higher Q.  The varying Q is the consequence of only adjusting one resistor to change the frequency - that's why this circuit was chosen.

The down side is you cannot *independently* adjust Q and frequency.

Here's a design sequence to take an existing design and change the parameters:

To change the frequency is most easily done by scaling the caps without affecting anything else.

You can change the Q by changing the ratio of R6 to R7.     If you change the R6 to R7 ratio by increasing R6 by a factor of x and decreasing R7 by a factor of x the Q increase but frequency will more or less stay the same.

Once R6 is chosen you can choose R5 to set the gain at the center frequency without changing Q or frequency much.  If you keep the R6 to R5 ratio the same then the gain will be the same (more or less).

The procedure might not be simple as you expected but that's life with this circuit.

** Where I've said "more or less" means R7 is small compared to R6.  If that isn't the case the parameters become interdependent and you really need to use the precise filter design equations, which you can find on-line (eg. TI app notes).

[John Lyons]

Thanks Rob.
That gives me some info to try out.
I had not considered R5.

[swamphorn]

I like to call a single-amplifier biquad (SAB) with a single variable resistance "half-tuned"; a half-tuned band-pass filter like this varies frequency and quality factor (Q) directly proportional to one another, so a 1:10 frequency variation means a 1:10 Q variation as well. Interestingly, this gives a constant bandwidth as the center frequency is varied. A Sallen-Key high-pass filter where the feedback resistance is made variable has a similar constant-bandwidth tuning characteristic; if the other resistance is variable, then frequency and Q become inversely proportional. A S-K low-pass filter exhibits quite different behavior; for such a filter, maximum Q is attained when the resistances are equal, so by setting the equal-resistances point at the center of the tuning range, the Q remains slightly more even as the center frequency is varied.

Unfortunately, there is no simple way to achieve constant-Q tuning with a single-taper potentiometer. Perhaps the simplest way to get a broader selection of center-frequency/Q combinations is to switch in different sets of capacitors, as Rob has said; this would vary the center frequency without affecting resonance, and then once can interactively fine-tune frequency and Q about this point. This may be not so intuitive but if the ranges overlap you get two or three bandwidth settings for any particular frequency.

If the situation permits a different circuit, a proper state variable filter (SVF) can produce a band-pass response with full, independent control over center frequency and Q. Such a filter is easily constructed with three op amps, which leaves the fourth in a quad package for buffering the input, and needs a dual-gang potentiometer (or, I suppose, a pair of well-matched photoresistors ...) to vary the frequency properly. Alternatively - if one can put up with the prolific amounts of noise they put out - such a filter can be constructed with a dual OTA chip such as the LM13700 (the datasheet even provides such a circuit), though I believe such a bandpass filter is constant-bandwidth and not constant-Q (but the frequency and bandwidth can at least be adjusted independently).