repl pots with vc

[Bissell, Harry]

Been there, done that. See National Semiconductor AN-20 PG8 for "analog
multiplier" made with two photocells and a lamp. This works real well for a
state-variable filter, where fets can't handle the dynamic range and offsets
limit the use of OTA's. I had to have honest DC feedback for stability in my
filter.
But: Don't use the incandescent lamp, at real low levels the filiment
response time is so slow, it will oscillate slightly (not good). I used two
LEDs in series, one facing each photocell. The ideal would be to have two
matched cells in the same package, at the same temperature, illuminated by
the same LED. But since I was replacing a light-bulb and had two cells
facing each other, I used two LEDs instead. The feedback around the driving
op-amp makes the control nice and linear and stable. I use this to preset
the Q to a fixed value, so speed of response isn't real critical... but it
is not noticibly slow. The worst case is as the cell gets dark... the
response gets slower. Mine works from about 10K ro 3Meg resistance. Hope
this helps.  :-) Harry Bissell

> -----Original Message-----
> From:	Martin Czech [SMTP:martin.czech at intermetall.de]
> Sent:	Monday, February 08, 1999 4:06 AM
> To:	synth-diy at mailhost.bpa.nl
> Subject:	repl pots with vc
> 
> Some modules have pots that are not so easy to replace with voltage
> control, e.g. Q in filters. No change in sound should occour. Over the
> weekend I tryed to figure out, how this could be done.
> 
> 1. Use a fixed load resistor and a LDR to ground as a voltage
> divider. Works, but range is limited to 2 decades, which seems to
> be enough for Q control. The scaling looks also god, faster in the
> beginning, then flat. Exactly what  you need for Q, but maybe not so good
> for other things.  Problem: the LDR gets very slow at the 1:1 divide
> point, it takes a noticeable fraction of a second, even too slow for
> hand adjust. But what's more negative is that a eg. sinuoidal modulation
> will change the average value is the modulation gets faster, towards
> more attenuation , ie. more Q. Undesireable, not intuitive
> for the user. "Through" resistance changes with LDR.
> 
> 2. Use two LDR in a divider scheme, with op amp circuitry to drive
> the LEDs. Gives more attenuation range, but suffers also from
> nonliearity in time behaviour. Level may stick at one end
> for faster modulations. Also "through" resistance changes with LDR.
> 
> 3. Use ota. Better use 13700. Has anybody tryed to give a state var filter
> a ota resonance control ? I must try this.
> 
> 4. JFET attenuator. Problems with large signal amplitudes, much
> distortion one is able to see on the scope, distortion changes widely
> with attenuation, ie. bias.
> 
> 
> Any comments ?
> 
> m.c.