26 Sep
2015
26 Sep
'15
10:07 p.m.
As I've periodically posted, my 8/A has an intermittent power supply problem
(Power OK light and signal line flickers, so the entire system is confused.
Throwing the DC breaker and resetting it will often make the problem go away
for the rest of the evening). It's slowly getting worse, to the point where
tonight the Power light wouldn't come on at all. It's not the breaker
itself. So I decided to put the G8018 regulator assembly on the extender
card.
However, when using the extender, the DC breaker would immediately trip! I
tracked it down to the +5 volt crowbar being triggered. Careful examination
of about four different schematic pages shows a connection on the backplane
between edge connector pins BP2 & BR2 and the multiple +5 volt busses on
edge conn C. If you look at the schematic, BR2 is a test point but BP2 goes
to the anode end of the 5.7 volt zener that sets the crowbar voltage (its
cathode is connected to +5) and through a 10 ohm resistor to the crowbar SCR
gate with 100 ohms to ground.
This makes *no* sense because that connection shorts out the zener, so as
soon as the +5 comes up, 100/110 of the +5 is applied to the SCR gate, so
naturally it fires!
A close look at the G8018 provided the answer. There is half a finger etched
on BP2 which is also fully connected to BR2. (See picture) Looks like a
drafting mistake to me.
The Douglas extender board has sufficient depth in its female connectors
that the partial finger of BP2 made contact and caused the crowbar to
trigger as described. Cutting the "web" between the fingers fixed the
problem.
Naturally the intermittent flicker is gone for now, and I doubt it was an
"almost" connection to the SCR gate because once it fires, the +5 should
drop to near zero and trip the breaker, which has not been the case (except
with the extender card). But at least I have the regulator on the extender
card so I can start measuring when it comes back!
http://s1181.photobucket.com/user/DrCharlesMorris/media/PDP-8/P09-26-15_20.…
