simple "wah" circuit.

[Brian Marshall]

twin T in feed back loop.

This didnt actually go as planned though.

for some reason i had to lower the 'cap to ground' value considerably from what i initially calculated, and originally.  it was originally a .047, and the resistors were originally 33k.  at that point the frequency didnt really move, and it was at the very top of the hearing range.  i think this may be because twin t's tighten their Q immensly in active circuits, but i dont know how to calculate that.

i dont have any good way of analyzing frequency response, but it doesnt sound bad.... it does have a major shortcoming as it is just a peak filter rather than a low pass with a peek.

this is a simple part of something else i am working on.  if anyone wants to play arround with it cool,  i'm sure there are fairly similar circuits out there as well.

Bri

[Tim Escobedo]

You probably don't need the 10k resistor or the 100k resistor in the feedback loop. Nor do you need the 100k resistor on the input unless you want to attenuate the input.

The resistor to ground should be OK. Keep in mind that with all the gain a op amp is capable of, Twin-T circuits oscillate very easily. A Bridged T may be easier to control. Also, configured as a non inverting amp, the gain varies greatly as the resistance to ground (andf frequency) is changed.

[Brian Marshall]

oops the resistor ot Vb is supposed to be 1M not 100k.  i dont think it would work with out a bias voltage.

i do need the 100k reisistor in the feedback loop to keep it from becoming a twin t oscilator.  the 10k resistor also helps the twin t notch from being infinite, and keeps it from distorting, and oscilating.

[Brian Marshall]

Also, configured as a non inverting amp, the gain varies greatly as the resistance to ground (andf frequency) is changed.

Unfortunately for the project i am using this for this is a side effect i have to deal with, but it works out in the end, as the next stage after wards is a lowpass filter with a prety low center frequency.

i also didnt put it in the drawing, but there is actually a 150 ohm resistor in series with the pot.

i went back and looked at the drawing as well, and now that i think about it it is almost practically a bridged t anyways with how small that .0047 cap is.

[Tim Escobedo]

oops the resistor ot Vb is supposed to be 1M not 100k. i dont think it would work with out a bias voltage.

The 100k resistor on the input is the one I'm talking about, not the one to Vb. It attenuates the input signal because it acts like a voltage divider in conjunction  with the other 100k (or 1M) resistor (to Vb).

i do need the 100k reisistor in the feedback loop to keep it from becoming a twin t oscilator. the 10k resistor also helps the twin t notch from being infinite, and keeps it from distorting, and oscilating.

The 100k and 10k resistors in the feedback loop may not only keep it from oscillating, it may keep it from being as effective a filter. The 10k resistor does this even worse, significantly unbalancing the "T". The two 68k resistors are enough. If it oscillates, try just removing the .0047 cap, turning it into a Bridged-T. You don't have to worry about a infinite notch (or boost, in this case), because a op amp has so much gain, it may be prone to oscillating no matter what, unless the T networks are tailored for less Q.

You've bumped into some of the big limitations of Twin T circuits. It's impossible to control cutoff frequency, gain, and Q independently on these things. Sweeping the frequency pot changes gain significantly, probably even more dramatically with this non-inverting configuration. And Q varies over the tunable range, even though it may or may not be high enough to oscillate through the whole range.

I'd suggest finding some Twin-T and Bridged-T circuits and copying them. The Ranger Para-Exciter is very similar to yours, a non inverting Bridged-T. It still does have that gain problem, though. I took advantage of the gain change in non-inverting mode to make a quasi synth VCF, since as the frequency is swept low, it falls down to unity gain, kinda simulating the effect you'd get with a high order VCF swept low, eventually cutting all audio spectrum.

[Brian Marshall]

i hear ya...

the input really wasnt that well though out, simply because this is supposed to be part of something else...  

this is going to end up being part of a much larger circuit though, and the gain varying gain in the circuit though is sort of something i am counting on though.  the previous stage is going to have some really nasty harmonics, and other noises to filter out, so the increased gain at higher frequencies is actaully a plus.  i am sort of afraid of what might happen with all the ripples in all the filtering.  complex filters are not my strong suit at all.  one of the reasons i used the twin t was because i knew how to calculate it.  i dont know how to calculate 'f' in a bridge t circuit.

it is wierd though, on bread board the 10k resistor didnt really "look" out of ballance, but it dose seem odd looking in the drawing.  i would put a small resistor on either side of the top half of the T.

in the end this part of the circuit will actually be envelope controlled... hopefully... havent worked on that part yet, but i'm doing it one step at a time, and working on other projects inbetween.  

in the end this may not work, or might acutally be more complex than it needs to be.  thanks for the help.