Bandpass filters (again)

[ExpAnonColin]

Is it typical, even at very high Q, that signal is passed uneffected (well, at lower volume)?  I thought that if the Q was reasonably high, it would provide enough DB loss for the outer frequencies that it would drown them out, particularly with guitar.  Does adding more bandpasses in series fix this?

-Colin

[Tim Escobedo]

Sounds like something's not right. At very high Q, the effect should be pretty noticable. If it's high enough, it'll oscillate all by itself.

[ExpAnonColin]

Sounds like something's not right. At very high Q, the effect should be pretty noticable. If it's high enough, it'll oscillate all by itself.

Right-o, it certainly oscillates when I get the Q high enough, and the effect is totally noticeable.  The question is, should there always be leakthrough even with a high Q?

-Colin

[Tim Escobedo]

At high Q, audio in the passband will be at the output. But you should be able to notice a tonal effect. What schem are you using?

[ExpAnonColin]

I don't think you're quite understanding my question. The effect is very noticeable, and it's clearly working as a bandpass filter, with reasonably high Q.  My question is, is it normal that even at high Q a small amount of unneffected signal will pass through because the out-of-pass-band-lower-gain doesn't drop it off enough?

I'm using a design I've cooked up based on some envelope filter schematics.  I'll draw it up tonight.

-Colin

[gez]

Maybe the bandwidth is too large?

[Nasse]

:shock: I have no knowledge bout htese but in a building list. Just have wondered if you make classic state-variable like epfm super tone control does the Q effect just bandpass output or is it noticed at lo and hi pass output as well? :shock:

[Bernt]

Colin!
The simple answer is "YES", you can't get a analog filter with infinitely steep "shoulders". So there will always be some bleed-thro.
Unless you go digital...

Kindest regards, Bernt.

[R.G.]

Is it typical, even at very high Q, that signal is passed uneffected (well, at lower volume)? ....   The question is, should there always be leakthrough even with a high Q?

Yes. Active filters, unlike LC filters, have an initial slope of whatever they're designed to do, but then at some lower level - like maybe -20 to -40db - the signal declines at only a single time constant slope of -6db per octave. You have to stack successive active filters to clear away the down-slope signal even more.

if you make classic state-variable like epfm super tone control does the Q effect just bandpass output or is it noticed at lo and hi pass output as well?

The "Q" has an effect on low and high pass as well, although it's a bit funny talking about the "Q" of a non-bandpass filter when Q is commonly used to talk about bandwidth.

A simple lowpass filter with a high Q will have a resonant peak at the frequency where the response would be half power if the Q were 1. So the frequency response is flat to nearly the falloff frequency, then rises to a resonant peak, then falls dramatically into the stop band. The reverse happens with a high Q high pass filter. A state variable does exactly this - if the bandpass is high Q, so are the lo and high pass sections.

Interestingly enough, a good wah pedal is usually a high Q lowpass filter. The passband below the resonant peak lets the lower part of the signal through for a fuller sound and the peak adds the resonant interest.

[toneman]

depends what U R feeding it.

Sine?  Square?

If U R feeding it a sine(wave),  all signal will B booted.
Amplitue Enhancement.
If it's a MultiMode, & use notch, the sine will (escentially)dissappear.

If U R using Square(wave) there are still harmonics left outside
the passband in bandpass OR notch.
So the signal will never completely dissappear.

Hopethishelps
tone

[Nasse]

Thanks R.G.

[ExpAnonColin]

Thanks RG, Bernt!  I was wondering if my circuit was messed up... I'll draw it up later tonight for all of you kind folkses.

-Colin