17 Aug
2010
17 Aug
'10
9:32 p.m.
I cleaned up the backplane and all edge connectors
with progold, tested =
out the power supplies, and the machine sprang to life. It would appear =
that all the bulbs except the LSbit in the address row work, which is =
good as the bulbs seem pretty unobtanium. I'll keep digging to find a =
source of spares for the 2176D. The bulbs have no markings on them, but =
they are fed via an 11v supply. It looks like they are tied to ground =
via a 1K resistor which I'm guessing is a warming circuit. I do find it =
rather odd that they are soldered in place instead of socketed, and =
since the bulbs have no metal at all - just two very thin leads coming =
right out of the glass - they seem pretty fragile.
Wire-ended bulbs are not uncommon, and are still available new.
My guess is that if this is an 11V supply then the bulbs are at least
12V, and posisbly 14V. You would want to underrun them to get a long
life, and they will be plenty bright enough.
Now for the first trick that I would use. Connect a milliammeter across
the 1k resistor for a working bulb (machine on, of course, but with that
bulb turnd off my the machine, so if it's an address indicator, have a 0
in that bit of the address register). The bulb should like to the normal
brightness, powered via the milliammeter. Note the current reading.
This will give a good idea of the sort of rated bulb current you are
looking for (filamament lamps aporoximate a constant current load fairly
well). So if you find it's, say, 50mA, you would look down the lists for
12V 50mA nr 14V 50mA bulbs. Get a few and try them...
I can't, alas, help further, I have a Nova 1210, but alas it's mising the
front panel PCB. I found it in bits in a skip (dumpster) and rescued all
I could find (chassis, front panel bezel, CPU, core memeory and a couple
of non-DG itnerface cards).
-tony