Image Map Image Map
Results 1 to 7 of 7

Thread: 5160 Random Parity 1 Error after Power Up

  1. #1
    Join Date
    Jan 2015
    Location
    Sydney, Australia
    Posts
    27

    Default 5160 Random Parity 1 Error after Power Up

    Hi,

    I have restored an old 5160 and it was getting a few parity 1 messages on the POST. I replaced all the memory chips and it passes the POST successfully, however if you leave it alone it will, after some time, spit out one of the following messages...

    Parity Check 1
    ?????

    Parity Check 1
    20000 (S)

    I removed the memory expansion card, which leaves only 256 on the motherboard. It powered up with successful POST. And, after some time it produced the following....

    Parity Check 1
    00000 (S)

    It never fails the POST. It's only after it has been powered up for a while before it will do this. Sometimes it can be 5 minutes and other times it can be a couple of hours. It doesn't have to be running any app. Just boot it and leave it running with the C prompt and, after a while, it will produce the error.

    I have replaced the first row of memory chips 3 times now, assuming the 00000 signifies the first row, with the lower addresses. I checked the forums and the minuszerodegrees pages but doesn't really help me, seeing I get ????? and 00000 and sometimes (S). It varies, and, as mentioned, only after running for a bit. POST is always successful.
    Last edited by river; June 8th, 2018 at 07:00 PM.

  2. #2
    Join Date
    May 2006
    Location
    Melbourne, Australia
    Posts
    5,775

    Default

    So, a read of a parity-checked RAM address is happening, and the parity calculated from the 8 data bits, does not match with the contents of the parity stored in the corresponding parity bit.

    Quote Originally Posted by river View Post
    I checked the forums and the minuszerodegrees pages but doesn't really help me, ...
    At minuszerodegrees.net, I had added 'PARITY CHECK' information for the IBM 5150 (at bottom of [here]), but have yet to do so for the IBM 5160.
    The main difference is the the 5160's NMI handler also tries to identify at what address the parity error occurred.



    Quote Originally Posted by river View Post
    Parity Check 1
    ?????

    Parity Check 1
    20000 (S)

    Parity Check 1
    00000 (S)
    The "1" in "PARITY CHECK 1" means that the parity error happened on the motherboard.
    The "?????" indicates that the NMI handler was unable to determine the address.
    The "20000" indicates an address in segment 20000.
    The "(S)" simply indicates that the address shown is a segment address.


    When seeing PARITY CHECK messages, one has to consider that the cause can be something other than RAM, e.g. faulty refresh, faulty parity calculation circuitry, etc. In fact, recently there was a thread in which a 5170 was randomly presenting "PARITY CHECK" (that and nothing else), and the cause was a "PARITY CHECK" virus.

    Also to be considered is that, per [here], the POST does not enable parity checking (the 'ENABLE NON MASKABLE INTERRUPTS' step) until near the end of the POST - just before it boots to BASIC or disk.

    I think that it would be a good idea for you to run the Supersoft/Landmark Diagnostic ROM. Details at [here]. Note its particular video card requirements. If you like, I can post a ROM to you.

  3. #3
    Join Date
    Jan 2015
    Location
    Sydney, Australia
    Posts
    27

    Default

    Hi,

    Many thanks for your reply. I downloaded the Supersoft code and burned it into a 27256, and installed it into the XT.

    It's running through the tests now. It has done 5 passes at the moment, but no memory fails yet. I expect the floppy test to fail as I have not put in a diskette.

    I am getting a fail on the following....
    - U28 System BIOS
    - U30 Basic
    - U31 Basic

    U29 Basic and U32 Basic pass.


    I will let it keep cycling for the memory error to occur.

  4. #4
    Join Date
    May 2006
    Location
    Melbourne, Australia
    Posts
    5,775

    Default

    Quote Originally Posted by river View Post
    I am getting a fail on the following....
    - U28 System BIOS
    - U30 Basic
    - U31 Basic
    The same version of the Supersoft/Landmark Diagnostic ROM is used for both the IBM 5150 and IBM 5160.
    However, the diagnostic's reference to chip sockets is for sockets in the IBM 5150.

    Refer to the diagram at [here]. For example, when the Supersoft/Landmark Diagnostic ROM refers to socket U29, that corresponds to the top quarter of socket U19 in the IBM 5160.

    Re the failure of "U28", which corresponds to a block in socket U19. Did you remove the U19 chip ?

    Re the failure of "U30" and "U31, which corresponds to blocks in socket U18. When you created a Supersoft/Landmark Diagnostic ROM, did you use a 27256, and use a 32 KB sized image made of the Supersoft/Landmark Diagnostic ROM image (8 KB sized) repeated four times ?

  5. #5
    Join Date
    Jan 2015
    Location
    Sydney, Australia
    Posts
    27

    Default

    Okay, got an error in RAM.

    If fails on Critical Memory, bit 6. So, I changed the relevant chip in Bank 0 and ran the test again. Same failure. Replaced chip again, retested and same failure. So, swapped with the chip beside it, and still failed on bit 6.

    Quote Originally Posted by modem7 View Post
    Re the failure of "U28", which corresponds to a block in socket U19. Did you remove the U19 chip ?

    Re the failure of "U30" and "U31, which corresponds to blocks in socket U18. When you created a Supersoft/Landmark Diagnostic ROM, did you use a 27256, and use a 32 KB sized image made of the Supersoft/Landmark Diagnostic ROM image (8 KB sized) repeated four times ?
    No, I left U19 in place.

    I used a 27256 EPROM and the image I downloaded, from the link you supplied, is a 32KB image "5150 or 5160_27256_32KB". I burnt the EPROM on a QB 4x4 burner and verified the image. I burnt 2 EPROMS in case one didn't work.

    Something interesting.... prior to running the Test EPROM, when the XT is first powered up, the screen goes blank but nothing happens. If you switch it off and then on again it works - that is, it shows the cursor at the top l/h corner and commences counting the RAM. Twice this has happened with the Test EPROM in place. The screen goes blank but nothing is displayed or happens. Turn it off, wait a couple of seconds and re-apply power and it works.

    I think this machine may have deeper issues

  6. #6

    Default

    Quote Originally Posted by river View Post
    ...when the XT is first powered up, the screen goes blank but nothing happens. If you switch it off and then on again it works - that is, it shows the cursor at the top l/h corner and commences counting the RAM. Twice this has happened with the Test EPROM in place. The screen goes blank but nothing is displayed or happens. Turn it off, wait a couple of seconds and re-apply power and it works.
    This might be due to the PSU, the power good signal might not be coming up. Have you got another one/ something else you can power it with? If not you can try to meter the power good signal.
    Looking for: OMTI SMS Scientific Micro Systems 8610 or 8627 ESDI ISA drive controller, May also be branded Core HC, Please PM me if you want to part with one.

  7. #7
    Join Date
    May 2006
    Location
    Melbourne, Australia
    Posts
    5,775

    Default

    A 'flakey' power supply can also cause random parity errors, so it would be good to rule out the power supply. In case you are unaware, adapters are available (photo [here]) that allow the use of an ATX power supply.

    But a 'flakey' power supply would not cause the errors described in post #3, if those errors always appear at each running of the diagnostic.

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
  •