Personally, I'd design the card to be able to
switch 100mA with an
output voltage of 20V (IIRC DEC ran the relay between the +5V and -15V
rails). Maybe overkill, but I'd be suprised if it didn't work...
Actually, DEC derives a +0.7V supply with a 150 ohm resistor and some
diodes, then uses a transistor to close the switch between *that* and -15V
through the teletype mod. This is the 150 ohm resistor only rated for 57ma.
Moreover, *another* diode clamps the reader enable line to ground, so the
actual voltage on the positive side of the reader run option is at most one
diode drop above ground. I have no idea why the reader run circuit is run
from with 15.7V instead of 15V, unless the drop across
the transistor is
supposed to leave 15V for the reader run circuit. That, in turn,
would
imply that the reader run circuit won't run on, say, 14.3V, which is only 5%
under. A circuit with less than a 5% margin seems like it would be flaky
over long cables, etc.
All of which just convinces me I don't understand the issues well enough to
bet on the results.
I have a redesigned version of the circuit, but with no way to test either
one, the situation is still not satisfactory. If I have to build it without
testing, it would be reassuring to at least know what the schematic of the
reader-run option is. I have the LT33 drawings, but they don't have a
schematic (with parts values) there. More like a picture of the thing, with
some cryptic notes about it.
Vince