That would be the crux of the issue, the battery inside the Dallas chip is starting to give out. You can either replace it if it's socketed, or hack a external battery onto it as explained here: http://greyghost.mooo.com/ds1397rework

Vynix wrote on 2020-02-16, 20:25:

That would be the crux of the issue, the battery inside the Dallas chip is starting to give out. You can either replace it if it's socketed, or hack a external battery onto it as explained here: http://greyghost.mooo.com/ds1397rework

Agreed... what do you recommend in the way of absolutely-certain diagnostics to confirm this is the case? A replace/hack will be scary work for me, so I don't want to go down this road lightly. For example, if the 'floppy drive not seeking on POST' issue is a function of a weakening PSU, I'd like to figure out how to test the PSU or eliminate it as a cause. Unrelated, I wish I could find someone locally (SF Bay Area) who could completely tear down, clean and rebuild the PSU since I'm sure it's another weak point that's well outside my ability.

I saw your link before posting; it wasn't very helpful as to what needs to be done and it doesn't look like the poster actually finished the job. On the other hand, this link seemed helpful: https://www.vdheijden-messerli.net/sgistuff/p … 5.19-io4-nvram/

I bought an old, cheap NOS DS1397 chip, I figured I ought to practice on it before taking the knife to my computer.

Swiego wrote on 2020-02-16, 20:54:

Agreed... what do you recommend in the way of absolutely-certain diagnostics to confirm this is the case? A replace/hack will b […]
Show full quote

Vynix wrote on 2020-02-16, 20:25:

That would be the crux of the issue, the battery inside the Dallas chip is starting to give out. You can either replace it if it's socketed, or hack a external battery onto it as explained here: http://greyghost.mooo.com/ds1397rework

Agreed... what do you recommend in the way of absolutely-certain diagnostics to confirm this is the case? A replace/hack will be scary work for me, so I don't want to go down this road lightly. For example, if the 'floppy drive not seeking on POST' issue is a function of a weakening PSU, I'd like to figure out how to test the PSU or eliminate it as a cause. Unrelated, I wish I could find someone locally (SF Bay Area) who could completely tear down, clean and rebuild the PSU since I'm sure it's another weak point that's well outside my ability.

I saw your link before posting; it wasn't very helpful as to what needs to be done and it doesn't look like the poster actually finished the job. On the other hand, this link seemed helpful: https://www.vdheijden-messerli.net/sgistuff/p … 5.19-io4-nvram/

I bought an old, cheap NOS DS1397 chip, I figured I ought to practice on it before taking the knife to my computer.

If I am not blind, your Dallas IS socketed.
So just replace it 😁

Is it?! That would be pleasant. It does look like it's on a pedestal which perhaps is the socket. Just use some tongs or other fixture to pull straight up? (Tried the finger wiggling test and it's solidly attached.)

Here are some better side views.

And update: yeah, that was rather easy to remove! Excellent thank you! Now to the next step of checking if the battery is dead.

Normally if the battery is dying the time would have started to drift away, and when the battery finally dies, the BIOS will go back to its default values, and the time will roll back to the default date.

That's not a very surefire way to see if a RTC chip is dead, but it at least gives you an idea if the RTC chip is dying or not.

Vynix wrote on 2020-02-16, 23:52:

Normally if the battery is dying the time would have started to drift away, and when the battery finally dies, the BIOS will go back to its default values, and the time will roll back to the default date.

That's not a very surefire way to see if a RTC chip is dead, but it at least gives you an idea if the RTC chip is dying or not.

That complicates things. On each reboot or power cycle, the clock appears unaffected. The machine clearly thinks the configuration is not set, and goes through a routine of asking some default questions. Could it be a corrupted CMOS or something similar?

I reset to default settings in the F10 setup screen, but it didn't help.

I've noticed that when I make some changes such as pulling memory sticks, on power cycle the system errors out that memory size has changed, adjusts it in F10 setup, then remembers the setting (and doesn't not error out) on subsequent power cycles. This also suggests the battery may not be the culprit... or at least the only element at work.

PROBLEM SOLVED.

It was the floppy drive cable... not properly connected to the pins on the motherboard (offset one set of pins.)

Vynix - your comment helped - realizing the clock is not resetting and therefore the battery must be Ok, I went back to basics. When I reset configuration in F10 Setup, the BIOS asked me not only to identify the floppy drives but also some other things e.g. power-on Num Lock state. Whereas on a power cycle, it only asked me to verify the floppy drive configuration. This led me to think it was about that 3.5" drive. All fixed!

And fortunately I now know how to pop off the RTC and have a second on the way to practice for the inevitable future where that battery dies.

That being said, it's amazing how long-lived the lithium cell in this thing is.