PDP 11/40 Serial Connection using M7856

Starting a new thread.

First of all...how do you check the memory management is OK? Sometimes address light 16 stays on (stuck?). I need to report back some specific test results so I can say I have a correctly working RAM. With my M7856 installed when I check 777560 and 777564 I don't get 000000 ans 000200.

Not sure if what you'd see in the 11/45 should match the 11/40. I have not attempted to hook up a serial terminal yet.

UPDATE: Found a working M7856. I got the expected 000000 and 000200

UPDATE - got echo characters program working. ... And then... plorch. All I get from RAM is all 777 777 777 777 at every address. I guess my RAM card blew out. It has a lot of caps I should check them. It's always something...

FINAL UPDATE - I think I have to return to the h742a power control board, fuse blew AC LO again...more work to be done here.

b

Hi All
Bill, to make things simpler, take out the Memory management Board, and put in a G7273 or Equivalent.. For now, it will make things easier.. Once You know that the first 28K or whatever you have is working, then we can go on from there..
The same thing with the EIS Board if you have one.. Keep it simple, less variables..
Next I would, put in Your "Counting Program" and watch it Count up for All digits..
Then I would Write to each Address, the first 16 addresses with a single bit moving from right to left, Depositing only a single switch in the up or (1) position, counting as you go..

at address 000 deposit Nothing (all Zero's)..
at address 001 deposit switch one (1)
at address 002 deposit switch two (1)
at address 003 deposit switch three (1)
at address 010 deposit switch four (1)
at address 011 deposit switch five (1)
at address 012 deposit switch six (1)
at address 013 deposit switch seven (1)
at address 020 deposit switch eight (1)
at address 021 deposit switch nine (1)
at address 022 deposit switch ten (1)
at address 023 deposit switch eleven (1)
at address 030 deposit switch twelve (1)
at address 031 deposit switch thirteen (1)
at address 032 deposit switch fourteen (1)
at address 033 deposit switch fifteen (1)
at address 100 deposit switch sixteen (1)

I think I got it correct ?? Anyway the Address and the Data switches and Led's should all show the same thing, when you examine the same locations..

Next, start at address 000000 deposit 000000 then do a deposit, so that it counts automatically, but do the single switch thing again, so that after you have enter all of them and then do an examine, at each incremental address, it should show again only one data light on..
Hopefully, between these two different programs, they will help you determine, if it is memory or not..
Also, If You have an M9301 or an M9312 With Don's ZZ program, it has a basic Memory testing program in it.. If You have a M9312 without a ZZ prom, Contact Don for a copy..

"" UPDATE: Found a working M7856. I got the expected 000000 and 000200 ""
Good for You ..
Also, remember IF you are using a newer computer which has no serial ports only USB, that a USB to Serial converter may NOT swing the full RS-232 Range and so You Input may not Work..
But, I you are using an Older Computer with a Real Serial port or a Real Terminal, then You should be OK..
Remember, I do not have my PDP 11/40, nor its documentation anymore, So I could be Wrong on some things, my memory might be Hazy !!

THANK YOU Marty

I was able to make my own cable using a spare Berg connector and EIN male with null modem, attach to the M7856 and hook up to a VT102 on the other end. Set the jumpers on the M7856 to 9800 baud to match the VT102 power on defaults. Entered the program to make A appear on the screen. I then toggled in a loop to print a series of A's then a series of D's, etc. Seems to be working just fine. The memory issues have also vanished. I re-seated everything maybe something was loose.

The Fonz says "AAAAAAAAAAAA..."

I'll post photos on my site when I have the chance.

Bill

Hi All;
Bill, Good Work !!!

"" I re-seated everything maybe something was loose. "" Been there, done that, I also cleaned the Fingers with an INK Erasure as well as cleaning the Pins in the Backplane.. I have some of the Richest Erasures around.. (Gold Impregnated)..

THANK YOU Marty

Hi All;
Bill, Sorry to see that it Blew up again !!
But, as stated in the other previous postings, The Caps are OLD, and Anything with Stress on it, Should be Replaced !!!!
Then I would check for Resistors that are out of 'Spec'.. And Replace..
Then I would look at ALL Transistors and Diodes, anything that has had a Resistor or a Cap Replaced or looks Tarnished or Burnt, Replace..
I know that this sounds Harsh, but How many times do You want to go thru this Blowing up the Poser Supply, and Possibly one of the Boards, just because, it seems easier not to until You have too..
Look at the amount of times I have been thru this.. So, before I go back to Core, I will get when I have enough pennies saved up new switchers for my -5, +5 and +20 Volts that are needed for my Core Boards..

THANK YOU Marty

I completely agree except in this particular case these are very good caps and resistors. I have most on order, I already expected to have to replace. I was just playing with my baby while I was waiting for the parts to arrive in the mail. I ordered all the caps and resistors for the power board, and I also plan to replace all of the caps on the RAM card. To get caps and resistors that are true replacements will be expensive, but worth it. I got as far as echo chars at least before the AC LO went low again. Fun while it lasted.
Bill

Hosted a vintage computing hacking workshop, with help from Kyle O. he helped me debug the h742a power board. The problems with the board were resolved as far as we can tell, every chip was checked and a few replaced, but nothing major stuck out as completely failing. We were stumped as to why we were getting AC LO while the board was in circuit. We set up a sort of a test harness to make it easier to probe important electric junctions, and I printed out a nice large schematic to follow. A few hours later, finding nothing else to fix or replace we started removing cards. When we removed the M7856 serial board and the DC LO AC LO problems vanished. ug.

I need to learn more why a fault in a M7856 would cause AC LO in particular.

But now I am closer to being able to use this PDP 11/40 to test cards made for it from other 11/40's.

b

Hi All;

Bill, Good work and Thanks to Kyle as well..
Since I don't have any schematics in front of me, I can't say for sure, But I would check with a meter from the DC LO and AC LO signals to the M7856 Board.. And then, check the M7856 Schematic, If they do show up on the M7856 schematic, trace the signals to the appropriate IC(s) and check it or them as possibly being Bad.. That should fix the problem with the AC LO and DC LO unless it is problems in other boards as well, that haven't shown up..
Maybe, You could send (Postage Paid) Kyle out here to Help me with my Various Computers.. As, two heads and viewpoints are better than one, when working out a problem..

THANK YOU Marty

Marty said:

Hi All;

Bill, Good work and Thanks to Kyle as well..
Since I don't have any schematics in front of me, I can't say for sure, But I would check with a meter from the DC LO and AC LO signals to the M7856 Board.. And then, check the M7856 Schematic, If they do show up on the M7856 schematic, trace the signals to the appropriate IC(s) and check it or them as possibly being Bad.. That should fix the problem with the AC LO and DC LO unless it is problems in other boards as well, that haven't shown up..
Maybe, You could send (Postage Paid) Kyle out here to Help me with my Various Computers.. As, two heads and viewpoints are better than one, when working out a problem..

THANK YOU Marty

Click to expand...

Pretty much I think the serial board is pulling down the UNIBUS voltage in a way that triggers AC LO, but I need to verify. At this point just speculation, it could also be the bootstrap terminator board. Left off there "why" but there are caps on this board, probable culprit. Just need to work through it. M7856 is not supposed to directly get or impact AC LO that I know off the top of my head. I have the manual, will find out for sure. Fortunately I have more than one of this board.

b

Thanks for the workshop this last weekend; sorry I did not have time to stay longer. Think I was there only long enough to cause confusion! The power supply card you are working on provides a regulated reference voltage to the high current supplies and also provides an AC Low and DC Low signal, but to be more exact the AC and DC Low lines are only pulled to ground by the card and the H level voltage is provided elsewhere. I may be all wrong being that I have no experience on the 40 family of products but in other Unibus systems the AC Low and DC Low are only used by the processor cards, Bootstrap and core if so equipped and have nothing to do with any power functions and recall that you mentioned that there may have been an issue with your 5 volt bus. I would first confirm that the primary 5 and 12 volt buses are clean and stable before anything else and then maybe start with the idea of pulling everything except the absolute minimum of cards required to bring up the processor. Have not tried it yet with a Unibus system but used Jörg Hoppe PDP-11 GUI on a ton of Qbus systems and that allows easy examination of memory and modifications of the same that would be way faster than the front panel.
Your collection of forties are monster systems and have to say anyone working with that size hardware is truly Hard Core, but myself have tended to notice that on a lot of that older large equipment things like week connections on high current DC power lines and corrosion can cause all sort of issues and always require close examination. You may have a bad bypass capacitor on one of your cards but when everything is working correctly your primary five volt buss it’s got more than enough capacity to blow the offending capacitor on keep on going and I have also seen where bus drivers and the like fail and instead of doing normal things like shutting the system down they will pop the chip and keep on trucking. The AC and DC Low pull downs on the regulator card that you’re working on do not sense the primary DC bus, they are feed by the secondary Low AC Bus in the system and just provide a delay to hold the processor off line before the primary DC power supplies stabilize and then allow the system to start. On any system I work on always check the power supply first and be certain that’s one hundred percent good because until that good none of the other things matter.

It was good to see you again, thanks for coming. We worked Sunday for a few hours on the 11/40's h742a. We went through every component of the h742a and made some repairs, but nothing major was needed. We put in an extension cable to make it easier to probe test points while the machine was operating. We had replacement parts for all of the questionable components, replaced a few but most were ok and did not really need to be replaced. Were pretty satisfied that the h742a was more or less OK when we put in back into the system. DC LO was resolved but we were still getting AC LO when everything was installed in the system.

On a whim we removed first the bootstrap terminator and then the M7256 serial card from the backplane. For reasons unknown at this time, that action eliminated the AC LO condition. This was not expected. I did some research and so far I have found no reason to believe that these cards impact AC LO, but that does not mean there simply isn't an old-fashioned bad cap problem on one of these two boards pulling down the power supply enough to trigger some sort of feedback that the h742a did not like. I have a 32K solid state RAM card installed at the moment, but I have core I can install as well, later.

I should add - the voltage regulators were OK (h744, h754, etc.)

Hi All;

Bill, may I suggest, go ahead and put the Bootstrap terminator Board back in, and check everything again.. That should decide whether the terminator Board would make a difference, since You have had it out while checking the M7856 Board.. So, do the opposite the Terminator Board in and the M7856 Board out.. And go through all of the tests, checking the power supply Voltages and Ripple..

THANK YOU Marty

Naturally if the problem "goes away" by removing two boards you have to at least narrow down the problem to one or the other

Being thorough I plan to test ALL boards, ESR tests of the caps, test diodes, resistors, etc. The underlying point I was trying to make is that I was surprised after we thoroughly tested and supposedly restored the h742a board that there was still a problem with AC LO. We had already been testing with and without the m7835cpu board, which actually relies on the DC LO and AC LO signals.

B

Remember that the card that controls the AC Low and DC Low cannot generate any voltage on those busses; it’s only capable of pulling the bus to ground. The high state is provided on that buss from some pull up resistor located on the bootstrap or CPU card.
Both status lines are generated by the same circuit that’s feed from the little bridge on the card that charging the capacitor that was blown. The output of the bridge charges the capacitor via the limiting resistor in series from the rectifiers and when the capacitor reaches charged state the several transistors on the card change the state of the DC Low and AC Low at that time, both are triggered from the same point, that capacitor.
If the card were removed from the system the two control busses would always be high state although that would be an issue being the same card also provides a reference to the high current supplies.

We measured voltages at numerous points for expected subtle changes in the voltage caused by the transistors, etc. I am with you there.

At this point I am moving on, the system is correctly working as a CPU / RAM / Console. Now I will be focusing on the peripherals, drives, etc.