[Visceroid] could only find overclocking resources online for TI calcs. He spent some time under the hood of his casio fx-9750G plus and sent in a write-up for us. I gave it a quick edit, but here it goes:

I opened up the calculator, which reveals the board with the processor on it,
with the RAM, ROM, CPU and a few other surface mounts visible.
The graphics are underneath and the buttons are also underneath on an underlying board.
The main batteries, backup battery and serial port are connected on the underlying board, and I also have access to them.

The board with the CPU is in the upper half, allowing the lower half of the underlying board to be revealed. The CPU is a blob of epoxy, and the RAM is a product of BSI. Don’t know what the ROM is yet.

There is a small surface mount crystal beside the CPU, 3 pins, with a resistor in parallel to the outer two pins. It’s just a small rectangular prism thing. The center is ground, and the outer two are the crystal. I unsoldered the crystal and used some thin flexable wires to extend it. The original crystal is reasonably hard to unsolder, but it’s worth it.

I made a hole in the side of the battery compartment, just above the backup battery.
The perfect place to put hack switches. I found a DPDT switch (actually has 4 poles though, but only need 2. Might use one of the other ones as an overclock indicator later) and soldered the wires from the CPU to the middle so you can toggle between the two crystals.

I connected the original crystal to one side of the switch, and a new 8MHz crystal on the other two. I insulated the bigger crystal to stop it from shorting. The bigger crystal is one of those plain big 2 lead crystals used in many microcontroller projects. I guess a resonator would also be acceptable too.

So basically, it’s a DPDT switch which toggles between the original crystal and the new faster crystal, located somewhere useful. It’s a good idea to have it switchable, because it means EVERYTHING is faster, so you’ll need to switch back to normal for serial communication.

I don’t think shielding is neccessary for the long wires. Some work is needed though to get everything fitting nicely. I am still having a bit of trouble that it is very tight fitting. one edge doesn’t close quite as much as it’s supposed to. But at least it works.

Some words of advice:
BACKUP YOUR DATA and also make sure there’s no power on board just in case.

2 times as fast
4.3 MHz to 8 MHz
probally uses more battery power though
you might get away with going higher, although I don’t recommend it. It’s probally going to make the CPU quite hot or just simply damage it. Maybe you could put a heatsink on it. The epoxy blob also means it’d be quite hard to sink the heat.

sorry if it’s a bit too descriptive!

In the future, I hope to do a ROM hack and allowing me to select between the original ROM and my own custom system in an EEPROM. That will come much later though. It is really hard to find this sort of information on the internet!