Many thanks for all the useful comments and links.
Pics here :
I assume the first picture shows the battery - helpfully stamped
What I'm not sure is whether that battery just unplugs or whether the MB
has to come out, for the battery to be unsoldered. If the latter, I'm
not that brave.
It appears to be a socketed chip that contains a lithium battery.
Just a question of carefully pulling the old chip out and plugging a new
one in, but whether a replacement with a good battery would be easy to
find, I don't know..
You may have to change the whole chip (I posted about this previously).
But since yours is socketed the phrase "piece of piss" springs to mind.
If you can't boot the PC with an inoperative RTC, you may not be able to
use the PC to switch on a new chip.
Alternatively, you might be able to just add a battery in parallel. You'd
need the spec sheets for that.
OK all, very many thanks for the comments, again. This appears to be
the spec sheet, with the power pins top right and bottom left (Vcc/GND),
presumably Vcc +ve and GND being -ve.
Have I understood this correctly? Assuming I can get the whole chip
out, I solder battery power (button, or 3 x AA) to those pins, pop the
chip in, job done?
Or, even with new battery power, how much will I have to reset? I can
see I will have to input details of the drives, but beyond that?
Finally, never having removed a chip, do I just grab it and pull gently?
On Wednesday, 30 May 2018 18:02:10 UTC+1, Graeme wrote:
Vcc pin requires 4.5-5.5v and consumes 15mA, too much for a lithium coin cell to last long.
I didn't see which pin to connect a cell to. You could also use a rechargeable charged from the +12 line, regulating it to 5v to run the IC. Flip the Vcc pin out to do this as there would otherwise be other loads as well.
On 30/05/2018 18:15, email@example.com wrote:
It is going to be a bit of a fiddle but a Schottky diode in series to
isolate the pin from the rest of the board a resistor to trickle charge
3x small Nicad cells to get 3.6v (>3.2v and <4.25v) might just work.
Depends how much current the thing draws in its half disabled state.
I'd be inclined not to bother and just rely on reentering the ram
contents as needed.
Probably safer to lever it out gently with a jewellers screwdriver
inserted under each end. Pulling with the force needed to get chips out
of turned pin sockets has a bad habit of bending the end pins if not
done *very* carefully and square to the board.
Replacing the chip may or may not work. According to the data sheet,
they are supplied with the oscillator turned off (to save the battery)
and need a specific code loaded into a register to start it.
Powering it as above may, but ity is possible that the chip defaults to
not running the clock after loss of power and a completely drained battery.
Many PCs will not be able to boot into DOS until it is running and will
not start it themselves, without a manufacturer's setup disk.
I think this chip is a replacement for the popular Motorola MC146818
realtime clock chip. The key phrase on the datasheet is "Drop-In
Replacement for PC AT Computer Clock/Calendar". There are other chips that
do that job which are still available - eg
I think I have one lying around never used, though it's probably 15-20 years
old. I can check next week - if I can find it and it's the right one, you
can have it.
If the RTC clock is not running, doing the hotswap trick may be enough - I
don't see anything that would cause the chip not to respond, though it might
not keep time out of the system since it needs the crystal for that.
Someone got in touch about an RM Nimbus 386 the other day - apparently to
boot from HDD requires the BIOS setup program which was provided on a floppy
disc - it appears lost. So the caution is to backup your CMOS RAM if you
If the system has a ROM BIOS setup program (can you go into setup without a
floppy/HDD?) then chances are good you can reset the RTC chip from there.
Many of them used 80186 CPUs, which is the embedded CPU version of the
8086. In today's terminology it would be close to being a "system on
a chip", since it included the functionality of the CPU, Interrupt
controller, programmable interval timers, some address decode logic and
other stuff. The main incompatibility was the programming of the
embedded ancillary devices was not the same as the standalone devices
used on the "standard" PC.
(there was also a big bank of registers called the Peripheral Control
Block that had to be accommodated in address space somewhere)
On the plus side it dramatically lowered the component count in the days
before high integration south and north bridge chip-sets, and by rights
should not have mattered if MS-DOS compatibility had followed the
roadmap that CP/M compatibility did. Alas applications programmers soon
started hitting the hardware directly in order to get extra performance
and with an anticipation that the hardware eco system would be less
diverse than was the case in the CP/M days.
 I have used them in a number of embedded applications, and to be
fair they are quite good... (if you want an Intel x86 platform anyway!)
That is extremely kind, and well worth trying. Thank you. Do let me
know if you find it. I have managed to get the old chip out, but if the
final solution is that I have to dump the 386 and try the hard drive in
something else, that will have to do. I do have an XP box lying around
somewhere. Be good to get the 386 running though, but certainly not
It turns out the chip in my junk box is a MAX1691CHE515, which is not IBM PC
However the DS12C887A+ is, and is currently available at the usual suspects:
(there are also some on ebay, but with old date codes - probably pulls from
equipment and likely with flat batteries)
On Sat, 09 Jun 2018 21:12:15 +0000, Bob Eager wrote:
I think it must have been twenty years ago now when the Dallas clock
chip issue came up. I remember perusing a circuit diagram that showed a
pair of Lithium coin cells connected independently to the chip which had
'battery management' built in, presumably the later version with the
additional 4K cmos ram built in.
ISTR digging away the potting compound from one of these Dallas chips,
probably just out of pure curiosity rather than to effect a repair. It
didn't seem too difficult a job to replace the cell(s) in one of these
things, you just needed to proceed cautiously if you wanted to actually
reuse it again after replacing the dead cell(s).
Theo, you are a star. You are all stars. Chip arrived yesterday,
plugged in and away! Well, almost - I had to re-enter hard and floppy
disk details, but luckily, in the box , I found a printout from
Seagate I downloaded in 1999, with full details. Updated time and date,
and away we went. Booted to the DOS prompt, and I was quite pleased I
could remember a few basic commands, without recourse to the manual.
Typed WIN and 3.1 loaded, with much disk crunching. Rebooted this
morning, and time/date still correct.
What surprised me was how little was on the disk - I must have deleted
all docts etc.  when that PC was replaced. Turnpike (Demonites will
understand) was still there, so I ran that, only to see a warning that
the date was far in the future! Well, it would be. There were some
newsgroups postings, but unfortunately, TP automatically deleted them as
they were more than three days old - actually 20 years!
 The PC arrived in two huge boxes, one for the monitor, the other for
base unit, mouse and KB, not to mention manuals. System manual, DOS
manual, W3.1 manual. Even a mouse manual. Wonderful stuff.
 No network (or card) then - I vaguely remember transferring
everything via 3.5 inch floppy, very slowly.
HomeOwnersHub.com is a website for homeowners and building and maintenance pros. It is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.