Adjusting resonance on tim escobedo's "ghost dance"?

[Top Top]

Is there a way to increase (or better yet, add an adjustment for) the resonance in this circuit?

I'm adapting it for use in a synth, btw... I changed the .0015 cap to a .01 to get deeper sweeps (sounds to me to give it lower cuttoff frequency capabilities)



What type of filter is this, by the way (lowpass, bandpass, band reject, etc...). I am familiar with the types of filters and using them, but not what they are by design. It sounds to me to be lowpass, but I notice that at the lowest settings (with the filter cuttoff at the lowest), it seems that some of the higher frequencies come back a little bit? Maybe it's something to do with the resonance...

[earthtonesaudio]

It's a bandpass filter, similar to a twin-T configuration.  Without adding any active components, the only way to make it more resonant is to more closely match the components.  Notice that there are two 220k resistors with a 680pF cap to ground between them.  That's one T of the "twin T," and the other is the two caps with the variable resistor to ground between them.  If you make those two caps equal in value, resonance will increase.  Note however that as you increase resonance you also increase the possibility of oscillation, which may or may not be desirable. 

Normally a twin-T filter is used to create a notch, but by placing it in the feedback path of an inverting amplifier, it creates the opposite response.  There is a way to increase the Q of the notch response version, by bootstapping the ground point with the output signal.  I don't know if this would also apply in the inverted configuration, but it might be worth breadboarding to find out.


There are a couple other additions to the circuit that affect the overall Q of the sound, such as the parallel 500k/680p at the input.  Those two components create a notch response, which would seem to reduce the bandpass effect somewhat.  You might try removing that cap.

[earthtonesaudio]

I should add, that in addition to matching the individual components, if you tune each T section to the same crossover point, the Q will be greatest.

[Top Top]

Thanks, that gives me a place to start... I have it on a breadboard at home so it is no problem to try these things out, I just usually don't have much luck with randomly replacing parts without any starting point to go off of. So I guess the first thing I will try is making the .005 a .01 (that's what I've got in place of the .0015 currently), and then I'll try taking out that parallel input cap.

What about making the two "T's" identical in terms of their parts - including the the variable resistance to ground? Is this what you mean about tuning each T to the same frequency? Do they need to be two caps and one resistor on one and two resistors and one cap on the other? Or could they both be like the upper "T"?

I'm sorry, but what does "bootstrapping" mean in this case? I looked up the meaning of the word in electronics, but still not sure how it would be applied in this case.

Also, there are still two inverters on that side of the 4069 that are unused (the other side has three taken up by an LFO). This would be the brute force method I believe, but what about just putting two of these filters in sequence? Could I use the same pot to ground for both, or would that mess with the signal path?

Finally, unrelated in a sense, I just wanted to note that there is a schem by Christian (Hemo) floating around for making a really simple envelope follower out of three of the inverters on a 4069 - I have read that people tried it with good results. That means one could throw together a very simple, one chip envelope follower... I may give this a shot once I am sure my LFO is completely worked out and I can free up that side of my 4069.

[earthtonesaudio]

More info here:
http://www.geofex.com/Article_Folders/EQs/paramet.htm#twin_t
...covers the twin T notch filter, bootstrapping it, etc.

For the steepest notch, the "middle leg" of the T is 1/2 the impedance of the "arms" of the T.  Steepest notch inverted becomes the steepest bandpass, a.k.a. the most resonant.


I'm not entirely sure this is the reason, but it seems like the different value is because, the signal travels through two parallel C's on its way to ground through the single R (so the third C is equal in value to the parallel addition of those two caps, or twice), while through the other T it travels through two parallel R's on it's way to ground through the single C (so the third R is the parallel value of the other two, or half).  For the steepest notch the signal has equal-impedance paths to ground in both directions.

[Top Top]

Hmmm... so I tried a few things.

1) Taking out the 680pF at the front just made it weaker
2) I replaced the 680pF that goes to ground with  .01 and this made it oscillate... especially when set to lower frequencies. So there is some resonance... but too much, and the filter didn't sound any better, just had a mild sine wave sitting on top of it.
3) Strangely, replacing the .005 with a .01 to match the .01 that I have on the other side (which replaces the original value of .0015) did not make any major change in sound - maybe slightly more focused filter, but lost some of the "fatness."

What I am wondering is if by increasing the resonance in this circuit, I am actually loosing some of the character I like about it as is. It might be that some of the deepness of the filter that I am getting is happening as a result of it NOT being too resonant. It is still plenty "farty," so I think I am going to leave it alone.

[Nasse]

I wonder if lowering pwr supply for cmos down to 5 volts would do something...

[earthtonesaudio]

My next thought would be to replace the 680pF cap to ground with something a little smaller.  With the other values you've used something like 470pF might be a bit more resonant.