• Please review our updated Terms and Rules here

Bad UNIBUS receivers?

Dare

Experienced Member
Joined
Dec 28, 2015
Messages
97
I'm trying to get my M9312 to work with my PDP-11/05 but there appears to be something wrong with the board. By itself, if I plug the M9312 into the standard bus terminator slot, nothing works. The system wont auto-boot the console emulator and I get a bus error trying to read the ROM using the front-panel (address 165XXX).

So I broke out the scope and started to check the basics; e.g. was the basic UNIBUS read cycle being recognized. Almost by chance I stumbled across a curious behavior: If I connect the scope ground--not the input, just the ground--to one of the inputs of the MSYN NOR bus receiver, the board magically comes to life. I can read the ROM and run the console emulator.

The circuit in question is here:

attachment.php


The point at which I attach the ground is E19 PIN 10 / E28 PIN 7 (either will do). No other points on those two gates seem to provoke the correct behavior, including E28 PIN 6 (BUS MSYN L) or E19 PIN 11 (BUS C1 L).

My non-EE conclusion is that one of the two gates is bad. Question is which one? Anyone have any insights?

Thanks,
--Jay
 
I would bet on E19. What you are doing is essentially forcing the signal that should come from E19-10 to low, at which point the gate in E28 would appear to work right. So E28 seems to be fine, as long as the input signals are right. So it don't seem to be any problem with E28.
However, it would seem E19 never drives the output low then... And if it doesn't, it will not matter what input you have on E28 pin 6.
Would be more interesting to see E19-10 in general. I would look for what the level is at all times, if it ever moves at all, and so on.
In essence, E19-10 should be the inverted signal of E19-11.
Of course, the problem could actually be elsewhere. If E19-11 is always low, then nothing should happen, and then you need to check why that signal never moves...
But start checking both sides of that gate to see if the both sides are the inverted of each other properly.
 
Just looking at the schematic C1_H must be L for BUS_MSYN_L to be allowed to propagate thru the E28 and gate.
If it is H, MSYN (and thus the bus cycle) are ignored.
C1_H should be L for DATI(P) (ie, read) bus cycles) which you'd expect the M9312 to recognize (ie it has no registers to write to, just PROM to read from).
So yeah I'd bet on a bad E19 device.
 
Thanks for the input, guys. E19 does seem like a reasonable place to start. What puzzled me was the fact that merely connecting the scope ground probe to E19 10 made the problem go away. In theory this should have negligible effect on the signal, whether high or low. This had me wondering whether there was some issue with the *input* stage in E28.

Now to find a trustworthy source for UNIBUS driver chips...

--Jay
 
Thanks for the input, guys. E19 does seem like a reasonable place to start. What puzzled me was the fact that merely connecting the scope ground probe to E19 10 made the problem go away. In theory this should have negligible effect on the signal, whether high or low. This had me wondering whether there was some issue with the *input* stage in E28.

Now to find a trustworthy source for UNIBUS driver chips...

--Jay

Your scope probe ground signal is, well a logic low. Ground.
Unless you have a fully isolated scope connecting the scope ground to an active node will force that node to ground / logic low.
 
Thanks to everyone for their help.

After my brain fart passed (of course my scope doesn't have differential inputs!), I bought a few DS8837s off ebay (along with some DS8640s just in case) and replaced the suspect E19 chip. Sure enough, that did the trick!

attachment.php
 
Back
Top