Caluser2000 wrote on 2021-08-04, 06:18:
I must tell the 30 plus year old rechargeable battery pack in my 1990 Zenith 286LP Plus need replacing because they should be leaking by now......
I guess you talk about a laptop battery pack. Such packs are used in a different way compared to CMOS batteries on board. I observed no leakage problems with CMOS baterries in computers that were in regular use. Usually, I got the problems only when reviving old retro hardware. The CMOS battery pack contains of three cells in series, which are kept full by a trickle charging circuit, which puts charge into the cells all the time the system is turned on. If the cells are full, the charging current causes electrolysis of the water contained in the electrolyte, producing oxygen and hydrogen. The electrical energy is taken up by this reaction. These are gases that build some pressure in the cell. A properly working cell can contain this pressure. It also has some catalyst built in, at which oxygen and hydrogen gather together and re-form water. The energy taken up by electrolysis is released as heat.
The charging current of CMOS batteries is low enough that the heat emitted by the recombination of oxygen and hydrogen to water is low enough to not cause significant heating. Classic NiCd cells were designed to take a small amount of continous overcharge without damage, the "trickle charge" method is explicitly allowed by most data sheets. You get leakage only when the seal of the cell breaks down, so it can not contain the pressure from the gases anymore.
A critical problem with all battery stacks (not only NiCd ones, you have the same problem with non-rechargeables too) is that possibly not all cells are depeleted at the same time. This problem is mitigated by the instructions to "never mix fresh and used cells" and "never mix cells of different types" in the case of non-rechargeables. If one cell is empty, the other cells (in case of CMOS packs there are two of them) still push current through the empty cell, charging it in reverse (you could also call it "charge it negatively"). The process of charging nickel-based cells in reverse is well-known to damage cells by chemical destruction, especiallly if the reverse charging pertains for a long duration. There are studies that momentary reverse charges, if the cell gets re-charged within some minutes or hours are benign, but continuous reverse polarity slowly kills cells. If a 486 board is in storage for years, the pack goes flat and one of the cells typically gets reversed and deteriorates, which decreases the capacity of that single cell in the pack and makes it even more prone to reverse charging.
My theory is that reverse charged cells do not just lose capacity but also get severely prone for leaking. Possibly the button cell form factor contributes to constructing the cells in a way that they easily get leaky. It might also be just old age that contributes to seal failure.
Laptop cells are usually treated differently: They are not permanently trickle-charged, but only charged when you request charging. In the 286 days, schemes where charging or operating from AC where exclusive options existed. The manual usually warned you to take care of charging time yourself (or they had some timer). As there is no permanent trickle-charge, you also don't get a permanent pressure in these cells. Furthermore, the main laptop battery pack might not be used for CMOS backup, so there is no continous discharge while the system is in storage. If you completely stop discharging a laptop pack when it can't supply the required operation power anymore, you do not reverse cells for a long duration. And even if CMOS backup current is still drawn from the main pack, compared to the pack capacity, the CMOS backup current is way lower than in the barrel-battery scenario (comparing a 60mAh barrel battery to a 1500mAh 20V or 5000mAh 4.8V laptop pack).
