[sdiy] Re: Guitar Synth update

[Ingo Debus]

Am Dienstag, 08.06.04 um 00:29 Uhr schrieb harrybissell:

>  It has envelope follower
>
> output faster than the EDN circuit (since I just happen to know the 
> exact
> peak of each cycle, and the zero cross).
>

This sounds like a similar thing I've done with my MIDI flute interface 
recently. The envelope follower is a full wave rectifier followed by a 
peak-hold circuit, straight out of Horowitz & Hill. The peak-hold is 
just an opamp with a diode in series with the output, cathode of diode 
goes to hold cap and to inverting input of opamp, this is also the 
output of the peak-hold (has to be buffered). The other end of the hold 
cap is at ground. The non-inverting input of the opamp is the circuit's 
input.

I used to have a "bleed" resistor in parallel with the hold cap, the 
usual compromise between lag and ripple (values 0.47uF for hold cap and 
220k for bleed resistor, lowest flute note is at about 260 Hz). There 
was still noticable ripple, and since I detect the cycle length of the 
signal anyway for pitch detection, I decided to replace the resistor 
with a MOSFET (BS170) that is reset every second cycle. In the other 
cycles (where no reset is done) the AD conversion is done. I also 
reduced the cap to 47nF. Well, inspired by the recent discussion on VCO 
reset times, I also put a small resistor (6.8 ohms) in series with the 
MOSFET to protect it. The BS170 has an on resistance of about only 3 
ohms typically, and I thought a peak current of more than 1 ampere is 
not so good for it.

This works, but behaves somewhat strange. Perhaps the opamp (TL084) 
doesn't like the capacative load. I'll have to check further.

Ingo