Building a replacement 386/486 CMOS battery

Necroware has a replacement design for Dallas RTCs: https://github.com/necroware/nwX287

I'm not sure if it's worse, since AFAIK these don't leak. I have a IBM PS/2 system that I inherited that I plan to restore some day. Sure it's a pain but glad there was no damage to the board.

The only thing worse than that is when those things are soldered and not in a DIP socket. They still make them (or a compatible version) by the way, sealed battery and all.

Ah yes. I've had to repair one of the Dallas ICs before. You have to dig out a dremel and bite down till you get to the metal then add some hook up wire. Unlike OP I chose to use rechargable batteries that used to be used for home wireless phones - they are the right voltage and often they supply a battery charger.

If you're lucky you can desolder the the old battery and use a couple of pin headers to get you back up and running. I do find it irritating some retro machines wont even boot without a working battery.

There were lots of test equipment with variants of those, and unlike CMOS settings, they held critical calibration constants.

Yeah, those are a thing on old Sun workstations. They have a clock and they also store a small amount of data.

https://www.sun3arc.org/FAQ/sun-nvram-hostid.faq.phtml

I remember replacing one in the mid 1990s. The machine's ethernet address is stored on this chip, and I believe it's battery-backed RAM (not flash, EEPROM, etc.), so when the battery dies, the data on the chip is lost.

When you install a new chip, you need to program in the ethernet address so the machine can get on the network. The proper thing to do is use the factory-assigned ethernet address to ensure that it is unique. But you can't get the address from the old chip. So, like many people, I just picked some random numbers. It worked.

Dedication and a cutter knife do it. I did it myself, twice.

Once while it was soldered into a mainboard :P

>no way

you drill/dremel in two spots, solder tow wires and you are done.

https://www.classic-computers.org.nz/blog/2009-10-10-renovat...

Is this some attempt to get around "contains batteries" shipping restrictions, since if anyone looks inside it'll just be like any other chip and there's no way to tell unless you start googling part serials lmao.

"No batteries here, officer! Move along."

Just a note with those holders. They shouldn't be used in anything that will be shipped as the battery can get knocked out. The original shipped batteries were soldered in so they could be shipped safely. In fact a lot of the time the cells were taped or cable tied as well as soldered in on some bits of hardware!

The newer CR2032's have much lower mass and can withstand many more G's before it's an issue so they are usually in holders.

Not OP, but with supercapacitors you're talking about a wildly different amount of energy. A 5F/5V cap stores about 62 J, while a 2500mAh 1.2V AA battery stores 10 kJ. With three of them you can store almost 500x as much energy.

Super capacitors have very high self discharge rates relative to a lithium battery. For something like an RTC which only uses a tiny amount of power, you’d quickly discharge the cap and have no way to charge it again.

What advantage does a supercap have over batteries?

AA or AAA batteries will leak.

That's my recollection as well. RTC batteries from this era often came in four possible flavors:

* Varta barrel battery, subject to leaking and corroding traces

* AA/AAA style brick + leads to motherboard, easy to replace

* Dallas DS1287 epoxied battery + chip - can be replaced, cut into to tap directly into the chip, or replaced with Necroware's nwX287 RTC module [0]

* Lastly, the standard CR2023 style. I started seeing these appear in later 486 Socket 3 boards, very rarely on anything earlier

[0] https://github.com/necroware/nwX287

They’re using rechargeable NiMh, not alkalines. I would have used the older Eneloops or current Fujitsus, but maybe rechargeable Duracells have better quality than their alkalines.