The mobo supports and came with 2 batteries. The system had both in place in 1992. And new replacements for each are now installed. I did detailed measurements and have complete confidence that the circuits that control battery voltage for charging and switching/usage are working fine.
It's only in the BIOS/CMOS screen where it displays the time that it becomes inaccurate. You can see it stop for seconds at a time. And over a 1 hour period it can lose 20 minutes. MS-DOS keeps the time. And no time is lost/gained when system is powered off, even overnight.
---
1st is an onboard 3.6 NiCd Varta bomb which I removed. And in its place I put a connector to use a rechargeable Li-Ion @ 3.7v. I tested the charge current and it supplies a tiny current of a couple of mA @ 3.9v. When the system is on. So the charge circuit is working. Under a charge load it pushes about 1.1uA current into the battery so to speak.
2nd is an off-board 4.5v alkaline non-rechargeable pack from Rayovac. It was long dead. And I made an equivalent 3x AA pack. I tested the diode circuit and it doesn't supply charge current, so that is good. The load on this battery is about 15uA. And it calculates out to about 6-7 years. It started out at 4.72v in April, and now it's down to 4.67v in July. A drop of about 50mV over 3 months. I think that is alright. When powered on, the current draw is 0uA, also good.
The "onboard" battery has a drain of 7uA if I remove the external AA pack. And it has 0uA consumption with the AA pack connected.
So that's where I'm at right now. Both batteries are connected and the mobo is receiving proper "CMOS/CLOCK" battery power.
