I Finally Built an ASM-1 Synthesizer.

[gstopp at fibermux.com]

     Hi DIY,
     
     Finally I have put together an ASM-1 behind a real front panel with 
     all the controls hooked up, played by a real keyboard...
     
     Up to this point I have only analyzed the circuit behaviour with an 
     O-scope or listened to the audio signals one at a time. A while back I 
     did track the VCO's with a 3-octave keyboard, with no problems. But 
     *all this time* since I have been selling the PCB's I have not had a 
     working complete system to play with - until now. I've always had all 
     the parts, I just needed a front panel and some time.
     
     The front panel I made is not one of the Christopher List panels 
     (sorry Chris, I was working on this one before I got the panels from 
     you). This one is made from a piece of aluminum with some overlays 
     made out of white card stock, with laser-printed graphics and 
     laminated in plastic. It looks real nice yet bears no resemblence to 
     any commercial machine, maybe the closest thing is like a Serge with 
     large (3/4") calibrated dials. Banana jacks were used - I'll try to 
     scan a picture sometime.
     
     I built this ASM with the same recommended parts as I use in my 
     documentation. To this I added the following:
     
     1. 1K output resistors on the banana jacks, for the four VCF outputs 
     plus the ADSR outputs. These signals normally come straight from 
     op-amp output pins and I'd rather keep 1k outputs as standard.
     
     2. 100K resistors on the VCA "INITIAL GAIN" pot wipers, as the 
     equivalent to using 200K input resistors on the PCB. This limits the 
     initial gain function to a more comfortable range.
     
     3. A few op-amps epoxied to the back side of the panel, to provide a 
     4-input mixer, two reversible attenuators, and a simple A/R envelope 
     generator.
     
     4. Two gate input jacks plus a manual gate pushbutton on each ADSR, 
     OR'ed together with diodes to the gate input.
     
     Here's some of the things I noticed, that may be of use to the rest of 
     you ASM builders:
     
     A). One of the VCA's had a *real* quiet output - there was a bad 
     solder joint on the 1K output resistor on the PCB.
     
     B). The ADSR's need tweaking in order to work properly. This might be 
     the biggest obstacle to anybody but the designer of these EG's (me), 
     so I'd better explain. When the gate arrives, the attack flip-flop is 
     set to start the attack. When the attack phase reaches its peak, the 
     attack flip-flop needs to be reset. If you look at the schematic you 
     will see that I cheated and did not use a comparator to detect the 
     peak-reached state, I just shove the ADSR output back into the bottom 
     CMOS NOR gate with a trimpot. This means that the logic threshold of 
     the chip is used as the comparator (sneaky, huh?). If this trimpot is 
     not set properly then the flip-flop may stick in the attack state even 
     after the ADSR has reached peak, so it will not decay and just release 
     when the gate goes away. The first adjustment is to turn down DECAY to 
     the minimum and trim the "SUS TRIM" trimpot to maximum resistance, so 
     that the max sustain voltage equals 1/2 the positive supply. Then 
     adjust the "ATK PEAK" trimpot so that the ADSR signal at the NOR gate 
     is at maximum level, them bring it back down until the ADSR peaks at 
     around 7.5-8 volts. Now go back to the "SUS TRIM" trimpot so that when 
     the attack phase ends, the ADSR output jumps to the same level as the 
     last voltage of the attack. I hope I remembered this right...
     
     C). The glide processor input is unterminated, so that if you just 
     attach it to a jack and don't plug anything into it, it will float and 
     drift around and probably pick up the LFO square wave or something... 
     maybe neat for special effects but annoying in the long run. My advice 
     is to ground it with a 100K resistor on the back of the panel. It 
     won't hurt anything and it should prevent input drift when no signal 
     is present.
     
     D). The "OFFSET" trimpots on the VCF *are* useful - don't leave them 
     off! This whole system is DC coupled all the way to the output of the 
     VCA's, and the multimode filter *does* tend to have 
     component-dependent voltage offsets. I was using a guitar amp for 
     monitoring and the DC distorted the signal pretty badly. You can put 
     AC coupling caps in if you want, but I'd rather keep it all DC. (Those 
     of you with Crown DC300's, watch out for speaker smoke.)
     
     E). While we're on the VCF subject, you may notice that multimode 
     filters have a pretty hefty gain at the center frequency when the Q 
     pot is turned up, causing some pretty raspy distortion. This sounds 
     just like our beloved MS-20 screaches, but sometimes clean is 
     beautiful in which case you will need to reduce the volume of the 
     incoming VCO waveforms by a factor of about ten to one. For those of 
     you who did not include input level controls, you may want to change 
     your input resistors on the VCF from 100K to 470K or even 1M. The 
     output of the filter will be quieter if you do this, so your VCA's 
     will not be running at their best signal-to-noise.
     
     That's about it as far as my first system tweaks went. I noticed that 
     everything is working great, and the overall system performance is 
     quite good. I was pleasantly surprised at the fact that the 
     stacked-banana patching scheme lends itself well to many whacky 
     Forbidden Planet soundtrack-type spacey noises. The ADSR's seem 
     extremely fast - I could drive the gate input with an audio-rate VCO 
     output and drive the other VCO with the ADSR output, and mess with the 
     EG controls to make some great changing-FM noises.
     
     Anybody who has an ASM-1 system up and running and has some questions, 
     please post to the list! I can now go straight to the workbench and 
     verify problems or whatever.
     
     - Gene
     gstopp at fibermux.com