• Please review our updated Terms and Rules here

PDP-11/44 Restoration

If you catch it in a good mood... be sure to check the line voltage and compare it with the last time when it wasn't so happy.
 
I checked the line voltage, this time it was not where the AC enters the chassis, but at the un-switched output of the power distribution module with the PSU on. The voltage was 122VAC.

A few minutes later when I went to check the voltage where it enters the PSU, it was 120-121VAC both with the PSU on and off.

The power light was solid for the PSU on tests.
 
Now the voltage seems to always be within tolerance according to the Console Interface Module, as well as the rest of the voltages that it does not check for. The light does not even flash briefly even just after turning on the system.

Could the problem have resolved itself? I am thinking that since I left it connected to the power for a long time it reconditioned some capacitors that could have been going bad. I checked it with the scope and I still get the same trace that I got before, but since the filter capacitors are not constantly energized like the input they may not have received the same benefit.
 
Last edited:
I'm tempted to tell you not to look a gift horse in the mouth.

Is the system fully re-assembled and functional?
 
When I received the CPU HUNG error in conjunction with the bad power light (occurred after a power cycle too if I remember correctly) I removed the cards to protect them from any more bad power until I figured out the problem. I could easily reassemble the system though, and I don't think that any of the cards would be damaged.

I have had the PSU running for at least 10 minutes without the power light flashing, and during that time I checked the voltages and they remained constant.
 
I'm sorry, but not having seen your setup I keep going back to power source quality. What kind of facility do you have for A/C power?


  • Any line filters or power conditioners in play?
  • What type of wiring are we talking about here?
  • 200A panel
  • Cu or Al wiring in the mix?
  • What other power was being consumed at your place when tests were failing?

Maybe you are right, and exercise has helped distribute the electrolyte in the capacitors. It's not something I've seen, but hey - it's plausible.

Still, if there's nothing to troubleshoot - either use the system or provoke it somehow. Since I don't think we're quite to the latter, go ahead with the former. (at least until a problem becomes evident)

Is there a site prep guide in the docs for the 44? Maybe you could read the power requirements and comment on if yours meets them or not.
 
Last edited:
Put the cards back in the system and get back to work, if there is going to be a failure in the power supply it will more than likely at this point only occur under full load. The primary AC thing is a good point, epically once you get drives and other parts in the picture. With these old systems it’s a good idea to have them on a dedicated circuit or at least try to be certain that they are the only large current drain on the circuit they are plugged into. Out in my shop I installed a sub panel that provides several separate 120/20A circuits with one that powers the 11/23 and the 11/34 with all there assorted drives with no issues but if you have the system on the same circuit with a window air conditioner or other high current load you may run into issues. Not an easy thing to do in a regular house or apartment but that’s what you get when you get into the world of large systems, the next thing is to start thinking about floor loading.shack12.jpg
 
Old Computers said:
Now the voltage seems to always be within tolerance according to the Console Interface Module, as well as the rest of the voltages that it does not check for. The light does not even flash briefly even just after turning on the system.

Could the problem have resolved itself? ...
Click to expand...

If you still have a problem and are able to identify the cause, be sure to post the answer here. The check engine light on my car seems to be acting the same way and I figure you might be able to short-cut the troubleshooting process for me. :)
 
Thanks Ray.

To apply some numbers to the picture, the 11/44 TM specifies:

H7140-AA (We do have an AA here - right?)


  • Line Voltage 90V - 128Vrms (120v Nominal)
  • Frequency 47-63Hz
  • Current 15A rms Max at 120Vac
  • 55A peak max at 120Vac
  • Power factor 0.6 at full output load@90v in
  • Inrush: 65A peak at 128Vac for 1/2 cycle followed by repetitive peaks of decreasing amplitude for 8 cycles
  • Power 1350W with max load at nominal voltage out

The 34 by comparison is about 800W for the system box.

Aside from the actual 1350W the 44 requires, the power factor is less than ideal, so thinking about the wiring and other loads is a reasonable concern. Add to that his line voltages have been on the high side of nominal when measured, plus his system has been "happier" since the heat wave broke.

These make me want to assure myself that his power source isn't soft in some way that while acceptable for other duty, is manifesting problems here.
 
The PSU label says AA/AB. I can give the specific board revisions in the PSU if needed.

The system is on a dedicated 30A circuit with copper wire and the outlet is the 30A outlet that matches the cable from the power distribution box. That is probably the largest thing coming out of that fuse box.

A fridge is on another circuit that comes out of that fuse box. I do not know what would have been running when I was testing it, the worst case was probably a fridge and the AC (on a different fuse box), but given the randomness of when the power light would flash and when it wouldn't I am guessing that both solid and flashing DC lights occurred with the above things running. Also I know that sometimes I had a modern computer on to check up on some info for running the PDP.

There are no filters on the circuit except for the one in PSU, and there are no conditioners. The PDP is connected to the power distribution controller.

I hope this is enough information to help troubleshoot. I am going to go ahead and re-fill the system to test it out.

@KM11 I know what you mean. It is always quite annoying when a problem seems to randomly fix itself because there is always some doubt as to whether it is actually better or not.
 
Last edited:
Well I re-filled the first backplane including the 2 other memory cards that I did not put back in the system and it did not work. The power light started flashing right away when I turned it on. I measured the +5V line and I read around 5.6-5.7V. I did not check the other voltages because I do not know the diagram for the second backplane. The other bad news is that I get a ?22 CPU HUNG error. I am really hoping that everything in the system is okay, and it is just the power causing the error. It almost seems like the cards are not drawing enough current to load down the PSU enough. I read higher voltages when I first tested the PSU with a small load before. Should I try filling the second backplane to see if it loads the system enough?
 
Sorry I haven't been watching more actively this evening. Um.. no, let's not do that. I think the supply may be going overvoltage on that 5v rail. It's important to know where you're measuring this however and what's your minus reference.

I need to look at the system schematics to be of any further help on it. I'd like to follow how the lamp is driven and what the monitored voltages and criteria are. Seems the easiest way to continue. Perhaps someone else will already know without having to research it.

Studying Schematic Pg 41, CONSOLE INTERFACE I see a series of Window comparators monitoring -12v, +12v, 5v, -15v and +5VB ... My thinking is to trace back through these to determine which is unhappy.

A little complaint here - whichever DEC engineer designed this damn thing should have included trace points to add blocking diodes and LEDs to this panel. Would make it real easy to see at a glance who's unhappy. As it is, all the outputs are ganged so it's impossible to tell who's firing except by their inputs. Grrrr!

Me be changing this puppy if it were my system.
​

The circuit seems to be depending on VB+ being at 13.8 Vdc (float charge for a lead acid or Gel Cell battery) to develop a reference for other comparisons. If VB+ is not above 13.3v it appears the circuit's conclusions will be invalid.

This has only to do with the generation of the DC ON signal (DC OK?) not the actual regulation of voltages or BUS signals.

I would also like to know the condition of BUS AC LO L (BF1) and BUS DC LO L (BF2) (description Pg 6-33 thru 6-35) It appears this may be describing your CPU halt.


More study...

I've just read the entire thread from the beginning... I'd like to return all the way back to your failed C4 on the BIAS Board...

Now I saw you got lots of excellent troubleshooting advice about tooling, clip leads and absolutely dangerous voltages. I'm a bit hesitant to have you go poking around there, but there are some fundamental things that bother me about those photos of the bias board. (scorched quadrant in the area of R2 the 15K carbon composition 2W power resistor - used in LTC gen)

BTW - Do you know what mod is on this vicinity?
​

Back to measurements - I'd particularly like to verify voltages across D12 the 16V Zener (although I think someone said 14v) and across C4. (two readings)

C4 doesn't look to me to have been the source of that heat, just a victim. (Q6 maybe?) I'd like to figure out who was the hot guy and if maybe C4 was fed overvoltage somehow, perhaps through the BIAS XFMR FDBK (Bias Transformer Feedback) in which case D34 is suspect.

J1-1 and J1-A on the LOGIC INVERTER BOARD appear to be the 5V remote sense lead inputs. Can we follow these to be sure they're hooked up?

Still need to find the original source of the VB+ mentioned above. Like to know it's value though.


I know it will be difficult for you to access all these readings. You should probably let me finish my read before diving in, unless one of those is easy.

Think I'll print out this schematic tomorrow. Helps me to scribble notes on them.
 
Last edited:
RSX11M+ said:
Studying Schematic Pg 41, CONSOLE INTERFACE I see a series of Window comparators monitoring -12v, +12v, 5v, -15v and +5VB ... My thinking is to trace back through these to determine which is unhappy.
A little complaint here - whichever DEC engineer designed this damn thing should have included trace points to add blocking diodes and LEDs to this panel. Would make it real easy to see at a glance who's unhappy. As it is, all the outputs are ganged so it's impossible to tell who's firing except by their inputs. Grrrr!

Me be changing this puppy if it were my system.
​
Click to expand...

Exactly how would you mod that card (back to the schematic & considering the board layout :->)?
 
Ooo... A challenge!

Do we have clearance on the component side of the CIM for a short stack?

I see it's up in the corner of the box - so probably yes.

I'd suck out the E9 E10 E11 and probably install low profile sockets. I'd make up a small board with machine pins to plug into those positions, on one side and put the ICs on that board with necessary reroutes and other components. Going to all that trouble, I'd probably want to remote the LEDs so I'd include a 20 pin ribbon header. I'd really like dual color LEDs but it might be difficult to figure out a reasonable color scheme. Red for what?.. too High and Green for too Low? Not intuitive. Maybe I'd go for a simple LED DIP bar display. Might even plug directly into the cable.

I'd try to locate em inconspicuously behind the front bezel.

No board cuts unless that DEC engineer did put those positions on the board already - in which case I got some crow eat'n t do. (but I'd still come up with a way to remote the LEDs)


Did I pass?
 
RSX11M+ said:
Ooo... A challenge!

Did I pass?
Click to expand...

Conceptual design: A+ (I was assuming the board-cut approach; I like the pull-n-machine-pin approach -- especially if clearance-issues require a separate mount via a ribbon cable)
Mechanical design: A (you considered clearance and alternatives, including user acceptance)
Electrical design: D (you evidently "thought about it" but didn't document your work)

Overall grade: B- (I weight D's fairly heavily -- no "social promotions" here ...)

So, yes you passed but if you'd like to revisit your Electrical Design, I suspect that you'll be able to advance to the top of the class :->!

Thank you,

paul
 
pbirkel@gmail.com said:
Electrical design: D (you evidently "thought about it" but didn't document your work)

Overall grade: B- (I weight D's fairly heavily -- no "social promotions" here ...)

So, yes you passed but if you'd like to revisit your Electrical Design, I suspect that you'll be able to advance to the top of the class :->!

Thank you,

paul
Click to expand...

Documentation? Oh, yeah... my dog ate it. (gave him the runs)
 
I will definitely try to find the BUS AC LO L and BUS DC LO L signals. If I understand the description correctly it sounds like the memory is not in operation when it finds that the DC LO L signal is asserted. Will I risk damaging anything in the computer by turning it on again with the bad power?

Is the VB+ inside the PSU or outside of it? I am assuming that it is a bias voltage.

As for the Bias and Interface board I only replaced 2 parts on it. The first part was the burnt looking capacitor, and the second part was a capacitor that looked like it was bulging (not sure if it was bad, but I figured I should be safe). The modification was a resistor that jumped two traces. I can't remember exactly where it was at. I will have to take a look at the board. I don't recall any scorched part on the board. The only part of the computer that matches that description is on the power distribution controller where the cord was against the heater.

EDIT: I looked at the picture, and I see what you are talking about with the scorched spot.
 
Last edited:
VB+ is to be found on the CIM (PDF pg 41) in grid C6 of that page. It appears to be derived from +15V -- I do not know what other nets it has at this time but it's so near the ideal voltage for a lead acid battery - I figured it had something to do with one. (has blocking diodes and everything. I'll look some more to see if I can find other references to it.

I still feel the best way to proceed is to trace back to the root cause for the DC ON light to go off. That will tell us which section is the problem. [+5VB too high I'm guessing] I THINK (didn't follow completely through yet) negation of DC ON will result in BUS DC LO L which would trigger the (M7095 CONTROL) CPU stop (schematic pg 63) described in the TM (reference above 6-33).

So the thing to understand is that the CIM will ultimately initiate BUS DC LO L if any of the supply voltages are outside the levels monitored by the CIM Window comparator we've been discussing. If I am correct, even a +5VB that is too HIGH will cause this.

Which is what I think is happening to you.

5.7v is too high for my comfort, and it may be out of spec on the CIM. The check of VB+ was to assure myself that the window comparator on the CIM has the correct reference voltage it requires, so we could then attempt to determine that it was indeed the +5VB scramming us.

Then we look at why it's too high (tells me the feedback is ineffective)

I cannot say at what level of +5VB damage will occur, but I don't like it already. It appears that more load is causing the voltage to rise (again if your point of measurement is correct)

I'd ask you to remove the boards, but we need at least the CIM to continue troubleshooting and we can''t fix something that's not yet failing. The only alternative I can think of would be to sub your modules for more dummy load in hopes of driving the +5VB out of spec like the real ones seem to.


To the group helping on this thread: Please - I'm running with about 4 hours sleep here, so anyone who has a comment on this reasoning is invited chime in. I'll return when I have rested.
 
Several years ago I went and did a tolerance analysis spreadsheet of the voltage monitor logic on the 11/44 CIM module. Here are the limits I computed:

Code: 
          +15V      -15V      +12V      -12V      +5V     +5BV
  MIN   +13.730   -13.731   +10.984   -10.984   +4.781   +4.668
  MAX   +16.235   -16.235   +12.988   -12.988   +5.412   +5.520

Here's a screenshot of the whole spreadsheet. If anybody wants a copy PM me and I'll send it.

cim.jpg
 
You must log in or register to reply here.
Back
Top