Wednesday, October 24, 2012

Korg Poly800 Joystick module

UPDATE 5/10/13!!!  This project is due for a rebuild.  Since I built this module, I've learned a bunch more cool/useful stuff about circuits.  I except a redesign in the next month. It will be even better!  Please stay tuned for my better KORG POLY800 joystick module in the near future.....

The first real synthesizer I ever owned was a Korg Poly800.  I must admit that I am partial to them--specifically the joystick, which controls Pitch bend, pitch Modulation, and VCF Modulation. I've gone through five of them over the years. I broke two and I sold three, and now I have none. But, I do still have some parts from the broken ones laying around and I thought it would be a fun/challenging project to build a module around a salvaged Poly800 joystick.

In the modular world, modules are all controlled by CV or Control Voltage. Control voltage ranges from around -5vdc to around +5vdc depending on the type of signal, gate or pitch. Joysticks work by affixing two potentiometers to the joystick, as you move the joystick, you change the values of each potentiometer.

 Above is a diagram of a potentiometer. A potentiometer can be wired as a voltage divider if you attach a negative voltage to lug A and postive voltage to lug B.  When the potentiometer is in the center most postion, Lug W will be 0 volts.  As you turn the Pot counter-clockwise, towards A, the voltage at lug W will become more negative.  As you turn the Pot clockwise toward B, the voltage at W will be come more positive.

This basic principle is what the schematic I found on Music From Outer space is based on.

And this will work without any major modifications on the pitch bend aspect of the joystick.  but this will not work for the pitch modulation and VCF modulation aspect of the poly800 joystick I am trying to duplicate.  

I think it will be easiest to describe the movement of the Joystick in terms of North, South, East, and West.

For pitch bend, with the joystick in it's center spring loaded position, the output is 0 volts.  As we move the joystick west, we start to get a negative voltage at our output, which when attached to the linear frequency input of an oscillator will cause the pitch (frequency) to decrease.  If we move the joystick east, passed the center position, we start to get a positive voltage which causes the pitch(frequency) to increase. done deal.

The first thing to realize about the how the joystick works on the Poly800, is that when you move the joystick north of its spring loaded default position, it affects Pitch Modulation, but if you move it south of its original position, it affects the VCF modulation.

This roughly translates into this idea, We want a way to separate the positive(north) and negative(south) values into two different signals which we could assign to two different modules.

I know, from building power supplies, a way to do this. when turning AC current into DC current, we can separate the positive side of the wave from the negative side of the wave with 2 diodes.

I breaded my ideas up on a breadboard and this is what I came up with.

R3 represents the Potentiometer inside the joystick.  R6 is a multiplier, it allows you to adjust the amount of voltage produced by the joysticks movements.  You may be wondering why one diode is before the op amp buffer and why the other is after.  It's was the only way I found would work.  The diode for the positive side of the joystick adversely affected the buffer when it was placed before the op amp, so I moved it after.

also I used an inverting buffer for the south movement of the joystick to convert the negative voltage into a positive, which I thought would be more useful than just a separated negative voltage.

something not included in the schematic was a tiny 1k trimmer pot I used in between R1 and R3 to "tune" the center joystick postion to 0 volts.

This schematic needs to be duplicated twice.  One for the X axis (east,west) and one for the Y axis (north,south).

Below is the actual build of the module.

Now I can control different modules using positive and negative voltage via North and South, or East and west movement (Ideal for pitch bend).  I can also divide all 4 directions into 4 different positive voltages (idea for whatever).  pretty wild.

This was one of my more complex designs attempts.  Most the stuff up to this point was designed by someone else and then I modified it.  This one was mostly all my ideas beyond the original voltage dividing potentiometer schematic from MFOS.  I don't believe that this method was how Korg set up the abilities of the joystick in the original poly800.  It's also not a perfect system.  It bends Slightly farther north than south, and slight farther east than west, making precise pitch bends rely on your ear more than the position of the joystick, but I would still call the module a success, and adds to the overall "rawness" of my Modular Synthesizer.

DC Voltage Distributor

The more you learn about electronics, the more you can figure out how to modify a circuit to suite your needs.  I'm beginning to understand what an amazing resource Music from outer space is.  Although the circuits are all designed around ±12v, some of them work with ±15v without modification and others can easily be modified to work with ±15v (what my system is).

So I built this DC voltage distributor. 
This circuit actually requires no modifications to work with ±15v.  A DC distributor circuit is ideal for using a single control voltage source to control multiple modules without loading the incoming signal.

when a signal is split into separate paths, the voltage stays the same but the current is divided.  A buffer circuit relays the voltage through an op amp, but the current is supplied from the power source, rather than the signal source.

 It's basically an input with 4 buffer op amp circuits.   I assembled the circuit on a breadboard first--tested it with my modular power supply, then assembled 2 of them on perf boards and finally mounted them for my module.

Here's the module in my system (3rd from left on bottom row).  I can now send a single gate signal to my ADSR module to shape my VCA in addition to sending the gate signal to my slew limiter circuit and then on to my VCF.  (in a pinch, a slew limiter can be used as a minimal envelope generator.)

I also found this module to be helpful when I was testing the module I made after this, my joystick controller.  having a buffer circuit that you can plug into immediately, helped me diagnose the problems with my joystick circuit. I needed buffers to keep the joystick circuit isolated from the loads I was plugging into it.