More on Envelope Generators....

[gstopp at fibermux.com]

     Hi List,
     
     A couple more details about my relay-based envelope generators:
     
     1. Drive the relay coil with an NPN in grounded-emitter configuration,
        or with an open-collector TTL gate, for guaranteed pull-in.
     
     2. Buffer the gate with an input comparator to a) prevent loading of
        the gate signal, and b) allow clean EG triggering with any input
        that exceeds the input threshold, including LFO waveforms.
     
     3. The whole thing can be powered off of +5 if you use:
     
         * Low power single-supply op-amps (like LM358 etc.)
         * 5 volt coil relay
         * Flip-flop made from 4000, 74, 74LS, 74C, or 74HCT logic
     
     4. The whole thing can be powered off of +12 or +15 if you use:
     
         * Single supply op-amps
         * 12v or 15v coil relay
         * Flip-flop made from 4000 or 74C series logic
     
     Do not use 74HCT at 15 volts. I did this, a severly exothermic 
     reaction results. Duh. Also watch your coil voltages. You may have to 
     design your particular EG around whatever relay you can get. As for 
     availability these things are actually pretty darn common. I will 
     check into catalog availability and post a follow-up email. I 
     personally like 15 volt envelopes because you can always attenuate 
     them down, further masking little imperfections like overshoots and 
     glitches.
     
     One possible glitch (ha ha) that comes to mind is contact bounce. I 
     expected to see more of that but then I realized that repeated bounces 
     at contact closure are going to be severely low-pass filtered by the 
     time constant pots and integrating cap. Besides, the contact closure 
     is for cap charge or discharge and there won't be any voltage jumps. 
     Also since the relays are DPDT (at least the ones I'm using), I can 
     just put all the contacts for both poles in parallel to average out 
     the bounces and halve the currents through each side.
     
     Anyway the ADSR will be proto'd soon and I will then create a 
     schematic with known working component values and tweaked to idealized 
     operation. I don't have any 15 volt logic at the moment so I'll do it 
     first with 5 volt parts.
     
     There is a design in an early Electronotes (low 30's) that uses reed 
     relays, but they're SPST so there's four of them plus all the drive 
     logic so it's actually a little more complicated than the Analog 
     Switch ADSR. However I think the advantage of my design is the 
     double-pole relays which save on parts count and makes more sense 
     anyway when applied to the ADSR algorithm.
     
     One thing I may try is open-collector inverters for relay drive as 
     well as the flip-flop. Maybe I can cut the parts count down even more. 
     I'll post more later.
     
     - Gene
     gstopp at fibermux.com