I have an ECS K7S5A motherboard that not only doesn't tick the clock when the power cable is disconnected from the PC, but also eats up CR2032 batteries like there's no tomorrow. I can measure full 3.2v voltage from the battery going into the first resistor in its path (R392, 1K ohm), but at the other end of the resistor the voltage drops to something like 1.5v. Then it goes through a diode (D18) and is further reduced to ~0.8v. This is probably way too low already. I'm thinking maybe some component further down the path is shorted (or way off tolerance), causing the voltage to drop more than it should... A short could also explain why the battery is quickly drained, going from a fresh 3.2v to 2.9v in just one day! I have tested both the relevant diodes off-circuit (D18 and D17), and they both work as expected. Also tested a ceramic filter-cap between battery voltage and ground (C517), no issues there. Both transistors (Q38 and Q39) were also tested with a multi-meter in diode mode, though in-circuit.

I'm wondering if anyone has the same problem, and/or if there is a known solution to the problem? I have cleaned the battery holder connector, and also force-bridged the CMOS clear jumper with solder so that it's guaranteed to be in the "normal CMOS operation" position at all times. Flakey jumpers on this connector is a known problem that causes strange issues.