Gameboy Development Forum

The circuit is not supposed to run through a capacitor. You'd need very fast drivers for that to work, because the caps would act as a high-pass filter. The circuit should look like this:



XIN is pin 19 and XOUT is pin 20 of the MBC3 and you cannot switch them. The resistor is important to hit the right frequency.

cYa,

Tauwasser

Honestly, it could be bad luck, but I would expect some oscillation in that case. I usually just go with 22pF for MHz range oscillators and never noticed any issue. Personally I find 5 pF stray capacitance a pretty high assumption, but it depends on the PCB, the chip process etc. In general I also find capacitor selection hard and usually try to adhere to device manufacturers' recommendations.

Some farfetched things I can think of is that the MBC3 has a weird pin that is only ever described as #WP (write protect?). It is usually connected to a high-active CS from the backup battery IC. There is also the PHI pin that is connected to the edge connector. I could imagine this somehow influencing the clock circuit, because it is supposedly used for synchronizing the clock registers.

EDIT: How do you actually know the crystal isn't oscillating? You aren't measuring with a regular oscilloscope probe, are you?

Last edited by Tauwasser (2018-04-17 18:09:21)

Ah, condensation can affect the clock crystal and cause it to stop swinging mechanically. Nintendo even advises that this might happen:

3.5.2 Condensation
MBC3 uses a crystal oscillator for its clock counter operation, and condensation on the oscillator may halt
its oscillation, preventing the clocks from counting up. Once the condensation disappears, the clocks will
resume counting up from where they stopped. However, please ensure that the counter stoppage does not
result in a loss of program control.