the lowdown.
the keyboard happened to fit exactly in the case of a broken 60's philips mono record player.
that means any bends and mods were no longer restricted to the size of the SA-2 case.
for mods and bends, i went digging through too much info on the Circuitbenders forum and over at CasperElectronics, where i found a neat SA-7 schematic. turns out to be almost 100% identical to the SA-2.
different clock crystal, and a cap more or less, minor stuff.
different clock crystal, and a cap more or less, minor stuff.
sound selection is normally reduced to Tone 0, 1, 2 and 3 buttons.
now, some extra poking and wiring can complete this matrix:
that's all 100 SA-1 sounds, 5 drumsounds and possibly 5 demotunes.
more on that here.
with an old school pushbutton switching array, i wired up essentially a 5PDT switch.
in one position, you get all 10 sound select buttons,
in the other, the bottom 5 are the drumtriggers - the top 5 create random glitches by injecting the amplified audio back into datalines 11, 12 and 13. that would be the first true bend, strictly speaking.
the whole thing does seem dramatically more stable.
maybe a bit less drastic on the glitching side, but it will still do the odd aleatoric tune every now and then, which is my main goal anyway. and it crashes a LOT less.
so, two crash buttons (and a power reset button).
as seen on the schem, i found a pitch bend as well. it actually works best when it is spitting out glitchy loops already - taken too far in normal mode, it will start to behave erratically and not recover.
since i was using that 4066 and 2 crash points are also in the pitch bend, i just wired that one up there as well.
that left me with one inverter on the 40106.
oscillator! LED driver! LDR across pitch bend pot! optovibrato! HAHAAA!
i put a nice blue LED on a gooseneck so i can position it anywhere i want it against the LDR -
looks very ridiculous.
extra stuff: added a distortion bend, running AN8053 pin#5 to OKI pin #5 through a switch + pot, as found on the circuitbenders forum.
the AN8053 pin is a frequency compensation output, basically a double of the amped signal.
the OKI pin is the unamplified audio output.
so we're getting a very crude feedback loop, essentially: amp audio out to amp audio in.
not sure if i like this one so much. might need to try a different pot value (have a 1K in there now.) it does sound great on beats though, adding a kind of filter resonant sound.
it goes especially well together with a frequency shift mod: parallelling the AN8053 pin#6 cap with another cap, or even better - touchpoints. sounds nice, although i think i made the wiring on this a bit too long, adding unstable capacitance and introducing nasty static. have to try a new switch there anyway, so that part's on the ToDoList.
next up: the insert jack.
the OKI output has a jumper somewhere before it gets to the 1.5K resistor. i cut that, and put in a switching jack.
most fx pedals seem to like the unamplified signal better, so there you go: an easy fix to have your cheap Behringer pedals behave more predictably with Casio keys.
minor drawback: the feedback loop does have very different results when faced with fx in there, if it does anything at all.
but hey - i tried the led vibrato with a tape delay in the insert jack..
shit is gorgeous, so pooh pooh.
more on that here.
with an old school pushbutton switching array, i wired up essentially a 5PDT switch.
in one position, you get all 10 sound select buttons,
in the other, the bottom 5 are the drumtriggers - the top 5 create random glitches by injecting the amplified audio back into datalines 11, 12 and 13. that would be the first true bend, strictly speaking.
next, i tried the Reed G. aleatron crash mods, but felt it blacked out on me way too much.
fun for recording samples or whatever, but for playing it live it was just too fickle. so: decided to crank that bend up a notch.
since the crashes depend on throwing the clock out of whack - so the little darling doesn't know where it is in its program anymore - i was looking for a way to interrupt the clock stream with more precision than the standard Aleatron mod offers.
enter CMOS.
took most of the puzzle pieces from the excellent MFOS website.
(scroll down to Synth DIY 101 - MickeyMouseLogic.)
look at the top two inverters: first one debounces a momentary switch so we get a solid ON signal.
second one turns that push into a fixed length pulse, no matter how long you hold down the button.
that pulse then triggers a 4066 analog switch, very briefly connecting (in this case) C11 on the casio board to ground, through a 47R resistor.
i tried putting the pulse length on a pot, but it doesn't actually seem to make the crashes all that more intense.
you just get to a certain point where it will just crap out on you (until you reset the power, that is.)
the whole thing does seem dramatically more stable.
maybe a bit less drastic on the glitching side, but it will still do the odd aleatoric tune every now and then, which is my main goal anyway. and it crashes a LOT less.
so, two crash buttons (and a power reset button).
as seen on the schem, i found a pitch bend as well. it actually works best when it is spitting out glitchy loops already - taken too far in normal mode, it will start to behave erratically and not recover.
since i was using that 4066 and 2 crash points are also in the pitch bend, i just wired that one up there as well.
that left me with one inverter on the 40106.
oscillator! LED driver! LDR across pitch bend pot! optovibrato! HAHAAA!
i put a nice blue LED on a gooseneck so i can position it anywhere i want it against the LDR -
looks very ridiculous.
extra stuff: added a distortion bend, running AN8053 pin#5 to OKI pin #5 through a switch + pot, as found on the circuitbenders forum.
the AN8053 pin is a frequency compensation output, basically a double of the amped signal.
the OKI pin is the unamplified audio output.
so we're getting a very crude feedback loop, essentially: amp audio out to amp audio in.
not sure if i like this one so much. might need to try a different pot value (have a 1K in there now.) it does sound great on beats though, adding a kind of filter resonant sound.
it goes especially well together with a frequency shift mod: parallelling the AN8053 pin#6 cap with another cap, or even better - touchpoints. sounds nice, although i think i made the wiring on this a bit too long, adding unstable capacitance and introducing nasty static. have to try a new switch there anyway, so that part's on the ToDoList.
i put a second LDR on a switch, to choose between bridging the distortion pot or the volume pot (i did away with the digital volume control and just added a pot at the speaker wires. not using the speaker, either.)
that makes for either pulsing fuzz or tremolo, when lit by that gooseneck led.
next up: the insert jack.
the OKI output has a jumper somewhere before it gets to the 1.5K resistor. i cut that, and put in a switching jack.
most fx pedals seem to like the unamplified signal better, so there you go: an easy fix to have your cheap Behringer pedals behave more predictably with Casio keys.
minor drawback: the feedback loop does have very different results when faced with fx in there, if it does anything at all.
but hey - i tried the led vibrato with a tape delay in the insert jack..
shit is gorgeous, so pooh pooh.
