Coin battery in computer drains overnight

[DaveE]

Some of those motherboards worth more than hundred dollars for a collector.
And they are like old wine. One piece dumped, one less. So when you finally found the one you are looking for for ages, you will try everything possible...

Yes, that makes sense in that world.

Many custom PCBAs cost hundreds to thousands of dollars in the other world. But then it isn't so much the cost to make or repair it, it's the cost of reliability, down-time, field service, etc. Vendors in that world are often chosen by reliability and ability to provide good support as much as performance of the product when it's new. Suppose that PCBA was in a laser that costs $100,000 dollars and if it fails, the laser quits, the semi-conductor inspection tool quits, and a semi-conductor fab line stops. Those people will spend anything to stop the bleeding ASAP. Similar issues at car factories, textile manufacturers, big clinical labs, etc. Actually, what most of those guys do is have spares on site or a central depot, if they can afford it. Sometimes the panic is "we used our spare unit and need to replace it ASAP."

Then the post-mortem analysis starts: What, why, how likely, how to avoid repeats, etc. There are cases, rarely, where you spend more money diagnosing a failure than the replacement value, just to get the data. So the decision often goes one of two directions: just throw it away and send a new replacement; or send a new replacement and spend a small fortune in engineer's salaries analyzing what happened, then throw it away. Those companies would go out of business in a month if they had to compete in the consumer product markets.

A $200 motherboard often isn't worth the cost of the technician's time if repairs aren't simple. Unless they are obsolete, then a used one could be a great deal at (almost) any cost.

 

[Averagesupernova]

Last night I spent a little time attempting to see if there were any of the other boards inside the case that would cause an increase in CMOS battery current when plugged into the motherboard. I found no changes, which is what I expected. Still was drawing just under a microamp.
-
So today I removed the display and keyboard and brought to the bench to run as a system while watching the current draw from the CMOS battery. I applied 3 volts again through the 100K resistor and found the unit still was drawing just under a microamp. I powered the unit up and the drain on the 3 volt supply dropped to .35 uA. I assumed it would drop to zero. No idea if this is normal but it isn't a concern. I ran the unit through setup, set the clock/calendar and shut it down. This time the current draw went to about 1.4 uA when powered down. I have to assume this is normal. My guess is that the real time clock that is buried in the south bridge/wherever does not run until the computer boots up. That's the reason just under a microamp before. When powered up again the CMOS battery current dropped back to .35 uA again and the clock and calendar were correct. I really don't know what else to do other than put a battery in and go with it.
-
I have to assume one of several things:
1) There is an intermittent leak that hasn't shown up on my bench.
2) I was simply lied to about the battery draining overnight. I know it was dead when it was sent in for repair but after 20 years in the field it's no surprise. By the looks of the old battery I have suspicions that it is the exact same one that was on the board when I shipped it in for repair. I'm wondering if the big glob of RTV silicone was simply put there at the last minute before they shipped it back so it would be difficult for me to tell what wasn't done.

 

[Tom.G]

At 1.4uA, that 235mAH battery should last just over 19 years. I think you got your money's worth out of it.

Datasheet here:
https://data.energizer.com/pdfs/cr2032.pdf

Cheers,
Tom

 

[Averagesupernova]

At 1.4uA, that 235mAH battery should last just over 19 years. I think you got your money's worth out of it.

Datasheet here:
https://data.energizer.com/pdfs/cr2032.pdf

Cheers,
Tom

Since I have not owned it for more than a small portion of those 19 years I didn't cash in on that battery life but it's likely someone else did.
-
The latest is I installed a new battery Friday evening and everything seemed to work. I checked it two days later after having been powered down with no change. All works as it should. I think the conclusion will be that the story I got from the repair tech is a big fat lie.