It only took about two years to complete. And another 15 months or so for this silly web page I put together to share my mods and panel control structures. It's from the original run of ASM-1 pcbs. Woo hoo.
The front panel is my own design. I'm not a graphic artist but so what. I wanted to keep the cords out of the way of the knobs. It doesn't seem useful to have the jack right by the knob since it's not immediately obvious what the other end is plugged into.
The back has tons of wires and features hinged standoffs so the pcb can be flipped down for hacking. (Hmm, that picture must have been taken before I hooked up the triangle outputs on the VCO -- and the power supply!)
I can't imagine you being here without knowing what it is already, but for the record, it's a home-built analog modular synthesizer designed by and built on a pcb by Gene Stopp. Find out more at Magnus Danielson's asm-1 page with schems and lotsa links to other builder's pages (and one rather large image).
Well, the filter is a little disappointing but other than that it sounds like a synthesizer. Tweaking the resonance limiting circuit might help but what would really help is complementing it with a ladder filter.
*What good is that? Since the sustain level is always zero volts, I can run the envelope output into a knob and change the amount of envelope CV without changing the sustain CV setting.
I like hard sync. I thought the original sync circuit was pretty lame. It was dependent on the fall time of the signal driving the sync input, and it would "unsync" on the rising edge which caused noticeable glitchiness as the frequency of the sync signal was swept. I wanted a totally hard sync, so I redesigned the sync circuit and stuck it on a daughterboard that handles the sync duties for both oscillators. I forget how it works, but I remember it was fun coming up with the resistors in the transistor circuit. I do little circuit boards like this with a dremel tool now and they turn out much nicer than the etched stuff.
I spent a lot of time getting the control ranges of the various knobs to work nicely. Here's what I ended up with.
I find octave switches way more friendly than coarse tune knobs. Here's the octave switch circuit I designed after trying unsuccessfully to find a predesigned circuit. On the lowest setting, the VCO becomes an LFO and the sensitivity of the tune knob is increased. The on-board pitch trimpot and whatnot are set up to allow the trimpot to vary the pitch by a bit over one octave, allowing a 0V input to produce anywhere from C0 to C1 with the lowest octave setting. The range of the tune knob on the panel is about an octave and a fifth.
I have two exponential controls. EXP1 is low sensitivity and is
suitable for vibrato, EXP2 is high sensitivity for wacky effects.
(gif or pdf)
The linear inputs are AC coupled to avoid pitch shift. The summing
resistor is chosen to allow a 10Vp-p input to start making things sound
terrible with no attenuation.
(gif or pdf)
The Pulse Width pot allows a sweep of somewhat greater than the complete
0% to 100% duty cycle range.
(Uh, it should be a linear pot on my schematic.)
(gif or pdf)
The frequency pot gives a range of about ?Hz to 15kHz. The upper limit is actually set by the current limiting resistors in series with the 3080 control current inputs. The 68K resistor matches the pot rotation with this limit.
CV1 and CV2 are center-zero and invert the control signal in the left
half of their rotation. The 500K pots respond slowly near zero
with increased sensitivity at the extremes. 250K pots were a bit too
sensitive near zero. Note the 100K resistors have been changed to 47K
to reduce the chance of U1-B's output saturating.
(gif or pdf)
At the moment the resonance control (not shown) needs to be cranked quite a bit to get the cool sounds, then it gets rather touchy. I've been working on tweaking the resonance-limiting circuit in the filter to try to make the response smoother.
The attack, decay, and release pots, which all figure into RC time constants, are 1M audio taper. Sustain is 100K linear. Attack has a 10ohm series resistor to avoid envelope overshoot due to delays in the logic and switches on fast attacks.
The init control varies from just below off to somewhat greater than
unity gain.
(gif or pdf)
Hmm can't remember what kind of pot I used here. If I would have had my head on straight I would have added a tune control into the glide somewhere as a master tune control. Oh well, I can always make another module. The Exp/Lin switch you can see on the front panel is non-functional. I don't like linear glide too much so it may never be functional.
I twiddled with both the pot and the current setting resistor, and
here's what I came up with. Control is from ?? to ?? Hz but the
control isn't very good on the low end. I'm not particularly happy
with the LFO because the duty cycle is noticeably off from 50%. This
is probably because the maximum output swing of the TL082 is not
symmetric. Some better LFOs are right up there on my todo list.
(gif or pdf)
I looked at some of the front panel layouts on the net, but didn't really like any of them. I decided to copy the knobs on one side, jacks on the other approach that a few people have mentioned. It made sense to keep the knobs free of patch cord clutter, and having the jack near the knob didn't seem to have any mnemonic value or whatever.
The front panel was laid out with xfig. I'm not super crazy about xfig but it did do the job despite the fact that I was constrained to work with the built-in postscript fonts. Here's the knob panel in xfig and postscript formats, and the jack panel in xfig and postscript. Notice that I screwed up and put the logo right where I'll be drilling the rackmount holes. Ooops!
A couple interesting things to note about the jacks are that I used 1/4" jacks in addition to banana jacks at the points that will be the usual interface to the external world: the glide input, gate inputs, and VCA outputs. Also the banana jacks are spaced 3/4" apart and strategically laid out so I can make some commonly occurring connections with dual banana plugs as can be seen in the picture.
Don't use Johnson banana jacks like I did. The metal sleeve rotates in the plastic part which will eventually cause the wires soldered to the jacks to twist and break unless you're really careful when inserting and removing patch cords.
I did this front panel with multiple CV inputs, but I've since decided that single inputs and CV mixers make a lot of sense because you can send the same mixed CV to multiple modules. But I sure have a lot of knobs.
I'm driving the ASM-1 from my JKJ CV-5. MIDI is produced either from my SY77 or from my Axon AX-100 guitar to MIDI converter. I've also got a WX5 wind controller but I've been too lazy to drive the ASM-1 with it. The CV-5 is pretty cool because in addition to the normal high/low/last priority modes, it has a mode that can output a CV for the highest note and the lowest note being played, which makes things work like an Arp Odyssey.
The AX-100 is pretty cool, but the CV-5 really needs some simple software tweaks to work well with a guitar controller. JKJ was thinking about doing them but apparently got bogged down. Too bad.
I've built up circuit boards for a dual ring modulator and a dual, cascadable S+H. They are waiting for front panels.
Magnus Danielson maintains the ASM-1 page.
You have been duly noted:
Questions or comments? Email me.
First published May 2000, last updated June 2004
