Time again for a stupid question. Why replace the original?
I hate the idea of batteries inside equipment where the electrolyte
goo can leak all over sensitive components. If I had your MicroVAX,
I'd get a bunch of primary cells in a battery holder and place them
outside of the case, so the goo will leak out onto the floor.
A diode in series with the battery pack should disable the charging
circuit and drop the pack's voltage to near-optimal. (3x1.2 NiCD =
3.6v; 3x1.5 primary = 4.5v - 0.7v for the series diode = 3.8v; pretty
close). If the battery's function is to keep the clock and some RAM
alive, 3 "D" cells will probably last a year or two. Make them
Lithium-iron cells and you'll probably get 10 years of service.
FWIW. the CMOS RAM/clock battery of this PC is not the (expnsive, hard to
find) lithium thing that IBM used, but 4 AA alkaline primary cells in a
holder with a couple of diodes in series. They last about 5 years, and
they're trivial to obtain when I do need to replace them.
However, a word of warning to anyone doing this. The series diode is, of
course, essential to prevent any attempt to charge the primary cells. But
a few machines -- the Torch XXX springs to mind, but thera are others --
use the NiCd as a crude shunt regulator to limit the voltage applied to
the RTC/RAM chip. The NiCd is charged from, say, a 12V supply line, and
if it's open-circuit, or missing (or of course replaced by primary cells
with a series diode), the votlage applied to that chip will be
essentially 12V, which will almost certainly ruin it. A Zener diode
connected across the oriignal battery terminals (of a hight voltage than
the primary cell pack, obviously, but low enough to limit the voltage
when the machine is turned on) is a suitable hack for this.
IIRC, My XXX has a 5.1V zener soldered across the original NiCd battery
connections, and 3 AA primary cells in series with a Shottky diode as the
backup battery.
-tony