maniacminer wrote on 2025-02-22, 02:01:

Now I am figuring out what to connect this to, I don't think there is a lot of space inside the chassis for a CR2032 in a holder to fit. I'll have to do some fiddling about and see what works.

Is there much room above the Dallas chip?
What I've done (on a number of occasions) is to dremel enough of a circle in the
Dallas casing (which is quite thick on the one's I've done) to hold the CR2032 in
place, then using a bit of "black tape" around the ends of the chip, hold the new battery
in place with flat - loops filled with solder and pounded flat - leads pressed against it via
pressure from the tape (which is a bit stretchy)

Doesn't take that much room over the chip, and I've had some in service this way for
a few years...

But I've mostly done battery-backuped RAM chips, not so much RTCs - so I don't know for
sure how the internal layout differs (I figured these out with very slow/careful dremeling -
and some - of the same part number/different date codes - were a bit different!
so just be very careful!

Dave ::: https://dunfield.themindfactory.com ::: "Daves Old Computers"->Personal

jmarsh wrote on 2025-02-22, 02:28:

What you should do is cut the exposed positive trace and connect the new battery to the lower half - otherwise you're trying to charge the old battery from the new one.

I cut the old one if I can ... but I've got at least one where that would have been riskier to do
than I wanted...

So I drew upon my experience with Logitech "Solar Keyboards" ... the school where my wife worked
tossed a few of these after the battery died (so naturally I brought them home).

They use an ML2032 which is a rechargeable 2032, but hard to find and pricey.
So I just put in a CR2032 and change it every couple of years...

I initially was worried about the solar cells trying to change the CR2032 --- Others who
have done this have taped over the solar cells to prevent this ... but I wanted to keep it nice looking.
(and internal construct doesn't make it easy to disconnect them without messing stuff up).

I built a "fake 2032" with fine wire leads so I could assemble the unit and have the battery working
on the outside... using a 4.5 digit DMM with it's lowest range being 2mv (1.9999) that makes the lowest
digit 1/10000th of a mv ... at those scales random noise usually causes the lowest digit to "bobble" but
you can tell if it's tending to be !0 --- and in full sunlight, I could see voltage from the solar
cells but no measurable current to the battery (ie: < 0.0001mv)
so I tossed it back together - and have used it for a few years, probably replacing the CR2032 twice
(I see it's due again) - no noticeable damage/leakage etc.

but ... you should do what you are comfortable with! (and I of course cannot guarantee anything or take any
responsibility for what you might do - I'm only telling my experience)

Dave ::: https://dunfield.themindfactory.com ::: "Daves Old Computers"->Personal

The point is to not have the new CR2032 wasting precious current fighting the internal resistance of the old dead battery, especially when the RTC is being powered by the PC.

jmarsh wrote on 2025-02-22, 03:05:

The point is to not have the new CR2032 wasting precious current fighting the internal resistance of the old dead battery, especially when the RTC is being powered by the PC.

Agreed, and my point was that with the CR2032s that I've tested (and that's what was in the Dallas chip
I hacked and had to leave the old one connected) no measurable charge current flows .. I was able to
measure down to 0.0001 ma .. and I've had one in operation like that (with the old one still
connected) for a few years and seen no noticeable decrease in the new battery life.

but .. ymmv .. no guaranteed .. you should definitely clip the old battery from the
circuit if you can.

Dave ::: https://dunfield.themindfactory.com ::: "Daves Old Computers"->Personal

There's not much room above the Dallas and I don't want to have too much height as it would be awkward to mount the PCB back into the case. As for reverse charging a lithium metal cell, it needs a higher voltage than the chemical reaction produces to charge, only an issue if one was to use 4 AA - up until around 6V no discernable current flows. I mounted the battery holder with a blob of hot glue onto the Flash BIOS IC, it is easily removed by just twisting off (unlikely I'll want to undo this mod though...)

My next task with this machine is to remove the RAM IC from the memory expansion board, I assume it is duff. The parity from the memory ICs is not connected to anything, so that is why it can accept parity and non-parity RAM.