42S Troubleshooting

[tgwell2]

(03-08-2024 05:09 AM)brouhaha Wrote:  The address lines to the RAM chip are the same signals used to scan the keyboard (US patent 5,065,356). There is series resistance (nominally 7K) between the Lewis and RAM signal and the keyboard scan signal, so that pressing multiple keys doesn't cause direct contention of the CMOS outputs of the Lewis chip, and the software doesn't try to access RAM while keys are pressed. However, if you do press multiple keys simultaneously, it will draw more than the normal current through two (or more) of those 7K resistors. Two keys in the same row would draw an extra (roughly) 0.32 mA with fresh batteries, and for four keys about 0.43 mA. That's not a lot of current, but then, the button cells can't supply a lot of current either. (357 silver oxide and A76 alkaline button cells are rated for 0.66 mA continuous, though can provide more.)

Assuming that the problem isn't contact corrosion, the cause is probably a failing RAM chip or a failing Lewis chip. If a Lewis chip is failing, the calculator is toast. The RAM chip is an industry standard chip and can be easily replaced. You can use a 32KiB chip in place of the factory-installed 8KiB chip, and have a 42S with more RAM.

This was exactly the kind of information I was looking for! I certainly appreciate the insight.

Today I noticed that I put in LR44s instead of 357 batteries... I bought a new pack of 357s and put them in just to try... and the problem seems to have gone away(for now?). I know there's a slight voltage difference between them, and I'm assuming an impedance difference but I'm a little surprised.

Thanks again!