Rather than the conventional 'short something out and see what happens' this is a little more subtle. Most chips have a manufacturers test pin, if you get the part number off a chip and google it, you can get the manufactureres data sheet as a .pdf. The test pin is normally connected to 0v in the toy / keyboard, but if you put it to +VDD or inject audio into it, you get some very interesting results. The Casio SK CPU's have a test pin ( pin 42 on the big SMD chip ) likewise the SK-5 percussion chip has a test pin, and I came up with the 'spacial distorter' mod based on this. Most chips do have a test pin, A/D's D/A's Mask MPU's, Codecs etc.
Attached is the info for implenting this on the SK-5 percussion chip which I hope people will find interesting! Cheers Paul
Hmm... I can't seem to get mine to work. I double checked both the service manual and the instructions here, but nothing really did the trick. All I ended up with was a slight fade in volume.
Is the trimmer set to something specific and left there?
Hi, that is correct.. if you do the mod without disconneting the test pin from ground, you will just get a reduced volume as you are grounding the audio output.
If you've still got your track cutting blade handy, you could try this ;-)
Pattern Mutator...
There is an 8 bit digital bus between the SK-5 CPU and the percussion chip ( basically 8 wires that send numbers in binary format, i.e. the number 10 would be represented as 00001010 and 255 would be 11111111 etc). If the SK cpu needs a bass drum sound, it places a number on the data bus, say 11001001 and sends another signal to the percussion chip to read in the 11001001 and make the noise. The percussion chip registers the number and the 'go' signal and produces a bass drum noise. The pattern mutator is simply a patch panel on the data bus between the CPU and the percussion chip, which allows us to 'modify' the number sent from the CPU, before the percussion chip reads it. For example the CPU sends 10010010, but if we swap over the first and last two bits ( by moving the wires ) the percussion chip sees 01010001 and makes a different sound. It also affects the drum pads, as pressing a drum pad makes the cpu send a signal to the percussion chip. This is also a very easy bend to implement, directly under the chip on the PCB are some nice test points ( round silvber blobs on the PCB track ) so you simply cut the track after the blob and before the pin of the percussion chip. All the wires from the blobs go to 8 patch connectors, and another 8 patch connectors are wired back to the percussion chip on the other side of the track that you cut. Then make up 8 jumper wires to connect the wires from the test blobs back to the wires to the chip pins.
That's perfect, as I was trying to figure out a way to modify the bass drum sound in the spacial distorter mod. Very exciting work you're doing here, Paul.
Paul, regarding your comment about swapping the data messages for the drum firing - does this mean then, that it would be possible to exchange the data lines, so that, where there would normally be a hand clap in a rhythm, you can get it to be the cymbal?
I ask this in relation to my SK-8 (almost identical to the SK-5 circuit).
The SK-8 doesn't have the drum sound pads - but you can still access these sounds individually from the matrix grid, and put external switches to trigger them, just like the SK-5 pads.
The drum sounds , in this case, on the SK8 are kick, snare, hihat, handclaps. I would like it to be: kick, snare, hihat, cymbal - because I can use the cymbal as an open hihat.
Since I have these 4 drum lines and the keyboard keys MIDI'd with a Highly Liquid MIDI SK board, it will then mean that I can finally access the cymbal over MIDI, and sequence my own rhythms with the cymbal (open hihat).
Below is the matrix table for the SK8, if that helps. Where the matrix table refers to "base" that means kick drum, not bass sound.
Yes very possible. I have sent you a .wav of various fiddlings of the patterns. You just need to identify the Byte that relates to a cymbal sounds. If you have an 8 channel logic anlayser you can hook it on the data lines and get it to strobe from the OKI chip enable pin, you will see each byte in realtime..
It is also possible to 'play' the OKI chip direct, if you connect the data lines to some SPST switches you can load up a byte i.e. 00100101 ( off off on off off on off on ) then you can fool the chip into playing whatver sound it has loaded at a specific address.
So I did some preliminary poking around with the pattern mutator mod and came up with some pretty ok results. First, I put 8 switches between the test points and the spots on the chip and tried that out for awhile. There was some decent stuff to be found just by trying out random combinations, but more often than not it produced a variant of the really busy spastic barking bend that you can find with almost any two points and a drum beat. I got the impression that if I had a better grasp of what information was going where (or even what the most significant bit was) that I'd be able to have better control of the output.
Maybe knowing the binary numbers for different drum sounds would have helped, but I didn't know which side to read from. I assumed that since I was looking at the bottom of the chip, the first digit would be all the way to the right, but I wasn't entirely sure.
I had much more luck wiring RCA jacks to all the points and leaving the data lines intact. The sounds stayed more rhythmic and did some nice filter type changes without the mildly off beat stuff.
In any case, I have a happy amount of work ahead of me.
Yes I've done this to an SK-1. Depending on where you connect the test pin, it will have a nice boomy effect on the drums, or other wierd randomness. I do have a .wav sample but it is to big to attach on here. I can mail it to you if you want to hear.
Author
wiring details.jpg
Logged
