6 Jun
2009
6 Jun
'09
10:51 p.m.
Roger Holmes wrote:
To cut a long story short, we have put a 240v/50Hz motor from an IBM
56 verifier in the keypunch and that part works great. We transplanted
the motor cradle and motor start relay too. The manual showed how the
DC power supply could be configured for 115v or 208v or 240v, though
all at 60Hz. Well we tried that and the DC produced was about 41volts
instead of 48v, which was more or less what was predicted by
collective list wisdom. The relay logic intermittently seems to work.
The existing capacitor in the ferro-resonant power supply is marked
15MF (incidentally, how come MF does not mean Mega Farad), and rated
at 330VAC. I think this means ideally I need (90/50)^2 * 15 which is
21.6MF. I have ordered two more capacitors to connect in parallel, one
5uF and one 1.5uF, to give 21.5 total, both at 440VAC.
It will be interesting to hear how the additional caps work out, both as to
how it works out for the ferro-resonant power supply and whether the
proper voltage improves the relay circuitry behaviour.
IIRC, the IBM relay practices document someone pointed to earlier in the
discussion made some mention of a large variation tolerance, but it may have
meant the system tolerance was fairly narrow so as to accept a wide variation
amongst individual relays, rather than implying a large tolerance on system parameters.
In my own hacking/design experience with semi-complex relay circuitry I did
find them more sensitive to voltage levels than initially expected, when trying
to get optimum performance. There were some conflicting characteristics
involving voltage and the stored energy in the relay solenoid vs clock speed
and system design that resulted in the power supply requirements being tighter
than expected.