HP-50G Excessive Battery Drain / Fix?

[3298]

Just chiming in to report that I had this problem as well. In my case, the power-off consumption was 55mA, idle was around 65mA. For reference, the serial number starts with CNA 038, and the calculator came with firmware 2.15, which I have replaced with 2.15+HPGCC3. It worked fine for years, until I found it with batteries drained to <50mV each a few weeks ago. (I think that wasn't exactly healthy for the rechargeables I use.)
After having "fixed" a Casio Algebra FX 2.0 Plus in 2008 or 2009 (by re-soldering a flash pin, but a year later it broke again), I'm not afraid of opening up devices and did so with the 50G. I expected it to be harder than with the Casio (8 screws at the back, 4 of them inside the battery compartment, that's it), but my 50G turned out to be a lot more stubborn than the Youtube video linked in this thread made it look like. Still, I managed to pull the front and back halves apart without breaking them.
My knowledge about electronics was insufficient for finding the failed component myself, so I called in help from my father. Long story short, after working with magnifying glass and multimeter he pointed at a 6x3mm rectangle of black plastic next to the wires from the batteries. It had the writing "100" and "10P" on it; there are two of these, and the one he pointed at was a little closer to the display than the other one. (In these pictures they are yellowish, but still the largest boxes apart from the chips. The labels next to them are CT1 and CT9.) "That's a capacitor, and it looks suspicious." He took it out of the circuit, measured it again and said "Yes, that's it." He replaced it with another capacitor from his spare parts collection (the material and capacity didn't match, but he said that doesn't matter as much for what this capacitor is used for). Then we tested whether the 50G was working - yes, it was - and put the multimeter into current measurement mode - it showed 0.02mA! Very close to the lower bound of what it could show in that mode, but it was close enough to the 27µA reported here for an intact 50G, so we didn't switch it into µA mode and just put the calculator back together and called it a day.
In the end I have a fixed 50G, and no more dust behind the display cover - I took care of that while I had access, because why not.
Moral of the story: check those capacitors as well, not just the huge one next to the batteries and the diode D6. They seem to be able to cause this issue.