Image Map Image Map
Page 1 of 4 1234 LastLast
Results 1 to 10 of 37

Thread: Troubleshooting PDP8E Major Register Board M8300

  1. #1
    Join Date
    Dec 2013
    Location
    Near Milwaukee Wisconsin
    Posts
    865

    Default Troubleshooting PDP8E Major Register Board M8300

    Recently, I picked up a couple of spare DEC boards. A M8300 Major Registers and a M8310 Major Register Control. The M8310 board works just fine in my PDP8E machine. But the M8300 Major Register does not. The symptoms are, or at least as I see them; I can Deposit data into memory, but when I use examine to view this entered data, it is not the same. Actually it is not that simple. The problem is Bit #4. If Bit #4 is or was stored as a zero, an Examine will report the zero. But, if the data is or was a one, the Examine will report the one but then all memory locations examined after that, Bit #4 will remain a one. So Bit #4 is stuck as a one, but only after it was read as a one. Now remember, the only difference is that I have a different M8300 board in a working computer. So the problem must be on the M8300 board.
    So, I started to investigate the difference between the Deposit and the Examine front panel functions. They are closely related.

    The Deposit will read the Switch register onto the Data Bus, then the SR data will transfer through the Data Control Gate, the Adder and output Mux, onto the Major Register Bus and then end up in the MB Register. From here the MB Register is written to memory.

    The Examine will read the CPMA memory location to the MD Bus, From here the data will go thru the Register Input MUX, the Adder and output Mux, onto the Major Register Bus and then end up in the MB Register. From here the MB Register is written to memory.

    The difference is Examine reads memory or Deposit reads switch register. Then different data paths up to the Adder. After that, the data path is the same. So I believe the problem is in the 4th bit path of the Register input MUX or the Register Input of the Adder. So I'm looking at MD04 to the Register Input MUX E30 (74153) to the Data Selector E36 (8266) to the Adder E37 (7483).

    The Register Input MUX selects the MD Bus Bit by sensing the EN0, EN1 and EN2 line. EN0 = ZERO, EN1 = ZERO and EN2 = ONE. These selections occur during the Timing State TS2.

    My question is how can I set up my Scope to try and determine which part is bad? If I just look at the input and output of the Input MUX on individual traces, they should follow each other, when I do repeated Examines? Then I could do the same for the Adder.

    Forgive me if I'm a little wordy, but sometimes, I need to talk this out, so that I better understand how this works. Mike

  2. #2
    Join Date
    Dec 2013
    Location
    Near Milwaukee Wisconsin
    Posts
    865

    Default

    Well.... so much for today. I figured that I'd first look at the good board to get an idea of what the waveforms would look like. And that proved successful. Then I installed the bad board and did the same, but I could not get any signals on the scope. Closer inspection showed that the leads on the E30 74153 MUX were all black. This was probably insulating my chip clip. A little De-Ox and a toothbrush cleaned away this stuff. Then a little alcohol to wash the board off. The board was then re installed into the machine, but now I can not even set an address. The entire problem has changed. So, I'm giving up for today, to think about this. Maybetomorrow will be better. Mike

  3. #3

    Default

    Clean the gold fingers and the backplane connector?
    Member of the Rhode Island Computer Museum
    http://www.ricomputermuseum.org

  4. #4

    Default

    Did you try to resolder those connections?

  5. #5
    Join Date
    Dec 2013
    Location
    Near Milwaukee Wisconsin
    Posts
    865

    Default

    I did not have a chance to work on it today. Since it was warm today 45, I put up Christmas stuff in the yard. It's supposed to drop into the 20's later. And tomorrow I'm working on the 1912 Model T. So, I'll probably not get to power the board up until Monday. But, I'm still thinking about how to go about this. I did clean the edge connectors and replace the electrolytic capacitors. I'll work it out, keep you up to date as I go. Thanks Mike

  6. #6

    Default

    I usually do a deposit and trace each bit from the Switches through each part of the path to the memory to make sure that it actually gets there, and then do an examine and trace each bit back to the bulbs. If you have the switches set to all ones and then all zeros you can compare the bits to make sure that they are the same through the whole path.
    Member of the Rhode Island Computer Museum
    http://www.ricomputermuseum.org

  7. #7
    Join Date
    Dec 2013
    Location
    Near Milwaukee Wisconsin
    Posts
    865

    Default

    Mike, I had a few minutes to play with the edge connectors. The over the top connections was why all of sudden the board failed completely. I have a couple of jumpers that connect the top of the 8300 to the 8310 boards and seems that when one of these connectors is pushed on all the way, it loses connection to the board. I have to pull the connector off the board about 1/8" and then all the connections are re established. Anyway, I'm going to try to find the problem maybe this afternoon. Since this morning is warm outside again, I have some outside work that will be easier done than to wait until the cold returns. I have to think about what you said about the switches. They will work for the Deposit, but the examine reads the memory, not sure how they would work for that. The data path is also slightly different. Thanks Mike.

  8. #8

    Default

    Mike,

    If the examine doesn't work for all bits, you really don't know if the deposit is working for all bits. You could look at the address and data bits at the memory for a deposit with the switches all on and all off. That would at least tell you that the data is getting to the memory for all bits. You could then do an examine and look at the data coming from the memory. It all of the bits look OK, then you could work your way through the CPU paths to see where the bits get different.

    Don't forget that an examine rewrites memory, so if the examine doesn't work correctly it can change the contents of the memory.
    Member of the Rhode Island Computer Museum
    http://www.ricomputermuseum.org

  9. #9
    Join Date
    Dec 2013
    Location
    Near Milwaukee Wisconsin
    Posts
    865

    Default

    Mike, sorry if I was unclear. But only the Bit #4 is bad. All the others one work just fine. I'll remember that tip sould I run into that problem. Thanks Mike

  10. #10

    Default

    Then you could compare bit #4 and #5 through the path to memory and back and see where they differ.
    Member of the Rhode Island Computer Museum
    http://www.ricomputermuseum.org

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •