Data Patch Bay – Yamaha PSR-6

circuit bending

While browsing some circuit bending sites some time ago I found an excellent tutorial over at Circuit-Bent.net for adding a patch bay to a retro Yamaha Keyboard. The tutorial dealt with the PSS-170, 270, 140 and SHS-10 but not having any of these models on hand I wanted to see if I could apply the same technique to the a PSR-6 I recently scooped up on the cheap at a local pawn shop. Finally this past week after the arrival of a big bag of banana plugs from E-bay I decided to give it a try.

– I also came across a very helpful forum post on Electro Music from another circuit bender (user name Dnny) who converted this mod for a PSR-6, It is definitely worth a read if you are attempting this bend.

circuit bending

To understand this bend you have to understand a bit about the inner workings of these keyboards. The PSR-6 (and many others of the era) operates using two primary IC chips. The first of these is the Yamaha XE323B0 CPU chip. This chip is easily identifiable as it will usually be the largest chip on the board and has a huge number of pins (I believe around 60). This CPU chip is the brain of the keyboard. Beside the CPU there is a smaller 18 pin FM synthesizer chip (the YM2413) which is responsible for actually creating the audio signals you hear out of the keyboard.

If you study the back (trace) side of the circuit board you will see there are 8 traces which travel directly from the CPU to the FM synthesizer chip. These 8 traces carry instructions from the CPU to the synthesizer chip and tell it when to make noises, and what noises to create. Further each separate pin carries information on a different aspect of the sound. These 8 traces are what I will be hijacking with this bend.

circuit bending

The first step I took was to cut each of the traces running between the CPU and the FM Synthesizer chip of the PSR-6. This will interrupt the normal flow of data. These circuit boards are very ruggedly made (as opposed to some more modern boards) so you may need to cut across them several times with an Exacto knife. Use a multimeter to confirm there is no connection between the two sides once the traces are cut.

Circuit Bending

Next up comes the really fiddly part. I soldered lengths of wires to each pin on the CPU and each pin on the FM synthesizer. I used yellow wire on the CPU and blue for the synthesizer so that they were easily identifiable. It is very important at this point to keep the pairs of corresponding wires together. I used pieces of painters tape to temporarily attach the pairs of yellow and blue wires together and numbered them in order from 1 to 8.

circuit bending

Once I had drilled and mounted the components for the patch bay I was ready for wiring. I soldered the blue and yellow wires to the middle and lower lugs of a row of switches. I then ran wire from these switches to two rows of banana jacks. All of the CPU connections I ran to the top row of jacks and all the synthesizer connections to the bottom row of jacks.

From here I closed up the keyboard and powered it up. To my delight it worked great. When the switches are all in the up position the data lines are connected and the keyboard works the same as it did without any modification. However when you flip the switches or connect any of the red jacks to any of the black jacks… things get weird.

Circuit Bending

When you are getting started with this patch bay it can be a little intimidating. It will crash and it will spew loud garbled noise, don’t be discouraged. You will also find strange, new and truly bizarre¬† noises you never knew the PSR-6 had in it. A good way to get your feet wet is to choose any instrument on the keyboard. flip down the second switch and then change to a different instrument. The turned switch will stop some of the information regarding the new instrument from reaching the FM synthesizer and you will be left with a strange hybrid of the two voices.

The more I play with this keyboard the more fun I have. Slowly I’m beginning to understand what each data line controls and I’ve been able to create more and more interesting patches. One surprising thing I’ve found is not all but many of the patches are repeatable allowing you a level of control over the noise I didn’t expect. This makes the device quite viable for live performance.

Now that I’ve gotten my feet wet with this bend I’m incredibly interested in taking it a step further. I’ve seen other circuit benders set up LED’s associated with each data line allowing you to see the data traveling through each connection. This seems like it would be extremely helpful in developing your understanding of what each connection controls. I’m also going to do some tests injecting a square wave oscillation into the data points to see what kind of results this elicits.¬† Finally I’d also be interested to try running the data through an inverter or other logic gates. If I have success with these tests I will create a second post detailing what I’ve found.

Share and Enjoy

  • Facebook
  • Twitter
  • LinkedIn
  • Reddit
  • RSS
  • Pinterest