Modifying a Casio CTK-80

Thread Starter


Joined Dec 22, 2015
Hello! My name is Seth. I am new to the forums. I've soldered kits from instructions before, but this is the first time attempting to do something on my own. I figured it's not too late to teach myself about circuits, and some help wouldn't hurt.
I am attempting to modify a Casio CTK-80 keyboard to produce new sounds, commonly known as circuit bending. Typical modifications to keyboards involve crossing wires and adding potentiometers to make noise machines. However I'm interested in adding more practical features. I found the service manual for a Casio CTK-50, which I discovered is basically the same keyboard but repackaged. You can find the manual here.
After looking over the manual I have ideas of what I can and cant do and what I can modify to make it happen, but when it comes to designing basic circuits, I have a very vague understanding.
I would hope to implement high/low pass filters, a pitch knob of some sort, and something that modulates the sound like a low-frequency oscillator.
The farthest I've gotten is turning the keyboard on without the case on reading the manual and tracing the schematic to confirm that the CTK-80 is in fact a repackaged CTK-50. I'm not really sure where I should go next. I have basic equipment but I'm unsure about having to purchase something like an oscilloscope to complete the project. Help would be appreciated, thank you!


Joined Dec 29, 2010

Below is the block diagram of a Yamaha keyboard PSR-640 which has pitch-bend control.


From the diagram, the pitch bend (yel) goes directly to the CPU's (green) ADC input and outputs control the sound chip (blue). The output of the sound chip goes to a DAC and came out as analog signal that we can hear.

But in the Casio CTK keyboard. The whole thing from CPU to the DAC are all inside one big LSI chip. The output of the LSI goes to Q201 which smooths the sound in PCM form to real analog sound that you hear. Then it goes to the power amp and speakers.

May be you can do it on the analog signal to vary the frequencies but that would be tough to design and the effect wouldn't be pleasing. You can easily bend the volume but not the frequencies of the tone at the analog side.

I guess it'd be much easier/cheaper to just buy a used keyboard with that facility built-in, rather than modifying the Casio Kbd.

Last edited:

Thread Starter


Joined Dec 22, 2015
The thoughts I had going into this was sort of like what can I do and what can I learn from doing it. The finished product would be similar to passing the stock keyboard through effect pedals but integrated. Using effects would create an interesting end product but there is something about doing it yourself that is enticing. I was thinking about modifying the processor clock to produce a pitch bend but that would probably cause more problems than its worth. Never the less I'm still interested on doing something. I want this to be a learning experience.


Joined Jun 4, 2013
Seth, welcome to Forum. I took a look at the information you provided, and did a search for the datasheet for your chip. The datasheet for the chip seems to have disappeared, or maybe it never existed because the design was proprietary between OKI and Casio.

I did find another version of the repair manual. It has a different master clock design that uses both the COSI and the COSO lines. Be sure you have the right one. CTK100 service manual.pdf

From the repair manual, I think you really have few options, but here they are. The idea is to see if any of the unused pins provide an extra function you'd like to have.
1. Remove the ground from Test1, pin1. Hook up a three terminal variable resister between Vdd and ground, and run the middle terminal to the test pin. Vary the resistor and see if it makes a difference.
2. Do the same thing on Test2, pin 2.
3. Hook up the middle terminal of the resistor to KO10, pin 20. Vary the resistor and see if it makes a difference.
4. Do the same thing on KO11, pin 19.
5. Disconnect the COSI and COSO lines, and put an independent 27MHz clock source to COSI line. Make that variable, and see what happens. I can tell you that the frequency doesn't have to be perfect, but it needs to be very stable.
Good luck!
Last edited: