Edward Randy Repair Logs
Once upon a time, many years ago, I owned an Edward Randy PCB. It had a slight whine/tone behind the sound, you couldn't hear it during gameplay but it was noticeable at the attract screen. I didn't know then what I do now, so I tried reflowing a few parts of the audio section - no change. Learned to ignore it, eventually sold it. Since then I've owned quite a few Data East PCBs of that sort of age, and the more modern ones do seem prone to this kind of problem - noisy audio sections which cause a background sound of some kind, usually very very quiet. Recently, I bought my Edward Randy PCB back, saw it on ebay with my name written on a tie-tag on it.
After checking the board, it turned out to be noise coming from one of the OKI M6295 sound chips. These have an analogue output on pin 36 (DAO). Looking at the datasheet, the suggested use shows the chip connected to a low pass filter (a job often performed by low value caps in sound sections). I noticed that when I touched pin 36 with an oscilloscope probe, the sound vanished. The small amount of capacitance created was enough to filter the audio line.
Pin 36 connects (in the case of Edward Randy) to pin 5 on a C3403C op-amp, they're a quad part so have 4 input/output sets of pins. It doesn't go anywhere else, no caps or resistors, nothing. I had a look at a few more DE boards, and found similar arrangements Rooting around in my parts boards box, I found a long dead Captain America - very similar sound section including 2x M6295 and a Yamaha YM3012 DAC. While the OKIs as usual had nothing between them and a C3043C, the YM3012 had some caps to filter the sound.
I took one of those caps from the dead game, and fitted it between pins 5 and 11 (ground) on the C3043C on Edward Randy. The annoying sound ceased. My guess at the moment is that DE designed them this way without much caring about the fact that some OKI chips are noisier than others (I have seen this before, even on new ones there's a certain amount of tolerance). If you've got a game with a not-so-good one, the op-amp amplifies the actual game sound and the noise. Perhaps it's that the OKI has degraded over time, I don't know either way - can't say. I had the game near enough 9 years ago and the sound hasn't changed in that time until I applied this fix. Interestingly, with the cap in place, turning up the volume even more shows a second noise tone, so quiet that at normal levels you would never ever hear it. That's coming from the other OKI chip.
I suspect the correct fix is to fit another OKI M6295, but that's not a particularly fun job and after you've finished putting the new one on, it might be noisy too potentially. This seems to do the job for now and might help someone with a similar situation. I tried a couple of different caps, and I found that in my case a 0.0015uf resulted in a reduction, while 0.0033uf killed the noise entirely. Since those are commonly available in thin designs, I've fitted it underneath the PCB - the cap is shallower than some of the surrounding chip legs in fact.
NOTE: obviously this will vary massively by PCB, the fix is only suitable in cases where probing an analogue output results in the complete reduction of noise. Use a non polarised capacitor, start with the smallest value you have first and work your way up.