Image Map Image Map
Page 1 of 5 12345 LastLast
Results 1 to 10 of 50

Thread: Altair 8800 restoration question

  1. #1
    Join Date
    Apr 2013
    Location
    Albuquerque
    Posts
    106

    Default Altair 8800 restoration question

    Hello again,

    So a few years on down the line, and I'm still working on my Altair 8800 in my spare time, and I've been able to fix all the issues I've had with it except for one which frankly has stumped me for quite some time. I have a theory about it that I was hoping someone can confirm for me. Upon power on, I do a hard reset (flipping the stop/run switch and the reset switch at the same time), and then flip the switches to send it to address 0 and hit deposit. However it does not deposit anything. The lights don't change, and it gives no indication that it accepted the command. now when I toggle Deposit next, examine, examine next, and single step it starts going through random data on the memory card, and when I hit run, some of the lights change, and a couple of them get dimmer and kind of flicker. I just cannot deposit anything. Now my theory is that since I don't have a second 8v supply for the display board due to the dual secondary transformer being gone when I inherited the thing, there is not enough juice for varying things to happen like depositing commands because of the voltage drop on the regulators. I am running the display board, and the s-100 mobo from the same 8v transformer. I have the Stancor 6138 for the 8v supply to the rail, a Stancor 8392 for the -16v supply and a Stancor 8397. for the +16v supply. So my theory is that because I'm running the display board and the other boards from the same 8v source, that one or more boards may be drawing more power than the others and therefore I don't have enough headroom for it all to operate properly. I know that the original transformer that I replaced with an 8397 was a dual secondary that provided the second 8v that was dedicated to the display board only. So I was wondering if the fact that I don't have a second 8v going only to the display board, that this may be my problem. I have a video attached showing my issue, and any advice would be appreciated.





    Thanks,

    -k1atn

  2. #2
    Join Date
    Jan 2007
    Location
    Pacific Northwest, USA
    Posts
    32,661
    Blog Entries
    18

    Default

    You're probably onto something. The original PSU in the 8800 was very wimpy. It was good enough for the CPU, SIO-1, and a couple of memory boards, but I think that if you went much beyond the 4 slot default with the original, you needed more power. Also, are your filter caps original or new? The originals have probably long dried out.

  3. #3
    Join Date
    Apr 2013
    Location
    Albuquerque
    Posts
    106

    Default

    The capacitors are new. Every component on the PS board is new in fact. I bought a replica board from Klyball a while back and added the components myself. The 8800 CPU board I built myself from a replica bare board that I ordered on Ebay as well as I did not have one and the original board that was in it was a z80. Furthermore, because it was wired for a z80 CPU, it was wired differently on the bus to work with the z80. That coupled with the fact that the old wires were quite brittle, I took the liberty of tearing all of the old wiring out of it and putting new wiring in as well as following the assembly manual and wiring it for the original 8080 CPU board. Needless to say, it is all stock now and back to original spec minus a dual secondary transformer that provides a separate 8v for the display board. The only capacitor I haven't changed is the one on the back of the display board. In fact, aside from a few worn out switches that I replaced with new c&k ones and replacing all of the old wiring for all of the signal lines and data lines, I did not touch the display board. I'm not sure how to test all of the TTL logic chips in their sockets as I've never done that before, but as far as I know they're good. I have a logic probe if need be, but if any are bad, I tend to just buy new ones since they're cheap online.

    IMG_20190810_134625207.jpg
    IMG_20190810_134707580.jpg
    IMG_20190810_134713590.jpg

  4. #4

    Default

    I'll walk you through trouble shooting if you like.
    First measure the front panel 5V line and the 8V line with a voltmeter ( report what you find ).
    I can see that it is not loading the address. The front panel uses the 8080 cpu to supply the address to the S100 board. It does this by jamming a JUMP instruction followed by the address switches. The JUMP instruction is C3H. Every thing else is handled by putting a NOP instruction on the CPU so that it will increment the address as needed. Clever isn't it!
    We can see that it looks like the NOP is getting to the CPU. This instruction is generated by 4 inverters in ic E and 4 in ic D. ( look at the front panel schematic ). The missing C3H instruction is created by the pull up resisitors for one bits and inverters pulling down on the center 4 bits to be 0's.
    Next with the power off make the following measurements with an ohm meter:
    Front panel ICs to CPU card upper right connector
    C-4 to CPU data 2
    C-6 to CPU data 3
    D-4 to CPU data 4
    D-2 to CPU data 5

    After this, report back and we may make some sign tracing with your logic probe.
    Dwight

  5. #5

    Default

    Actually, it looks like you may be in better shape than you think. The reset operation looks correct, the status lights look correct, and the Examine operation works, as can be seen at 0:31 in the video. Right now I’d suspect your RAM board. It appears to be the notoriously unreliable Altair DRAM board. Do you have another RAM board to use?

    Mike

  6. #6

    Default

    Hi Mike
    I took another look at the video, and I see that he used the examine but I did not see enough to ensure it was wired correctly. I find that this is a common error point on the Altairs. I always check it before examining other problems. He is missing the A1 light. That could be a problem in the display or a short on the back plane ( a common problem on these home built machines ).
    It looks like the front panel is not accessing the S100 bus for the data read. That could be in many places. It is unusual for a memory to read constant 1s after reset. DRAM can do that with a failure to refresh. Since he is using a built CPU card, it may not be providing a refresh signal. I have some check we can do for this but first things first.
    There are a number of things we can check.
    Dwight

    I agree that another memory board would be a good idea. Using a static memory is always a good idea.

  7. #7

    Default

    Dwight,

    You are right about A1. Watching the front panel during reset, it looks like A15, A13, A7, and A1 all need to be checked to see if their failure to illuminate is due to front panel/LED problems, bus shorts, or a bus driver issue on the CPU board. An additional quick check would be to remove the DRAM board and watch the LEDs during reset to verify the RAM board is not shorting the address lines.

    The CPU board k1atn built is a replica of the original Altair 8080 CPU board. It does not generate any sort of refresh signal for system RAM. Based on experience with this board, it’s likely the 2-phase clock timing will need to be adjusted on this board before it will operate properly with a running program, but front panel operations typically work without tweaking the clock.

    The Altair CPU has pull-ups on all data lines, so if no memory is in the system (or in this case, if the RAM board is not selecting and therefore driving the data bus), then all data LEDs will be lit as we’re seeing.

    As you note, there are many things to cause the primary symptom of memory deposit/examine failure. Hopefully k1atn has ready access to a static RAM board know to be compatible with front panel machines in order to continue testing.

    Mike
    Last edited by deramp5113; August 11th, 2019 at 12:54 PM.

  8. #8
    Join Date
    Jan 2007
    Location
    Pacific Northwest, USA
    Posts
    32,661
    Blog Entries
    18

    Default

    Do these "reproductions" still use the 74L00 in the front panel circuitry? Those have to be hard to find nowadays.

  9. #9

    Default

    The board itself doesn’t come with parts, though a seller on eBay offers a parts kit and another seller offers the board fully assembled. I think both provide the several “L” parts that were supplied with the original CPU board.

    MITS used a handful of L parts on a few of their early boards. In all cases (as far as I’ve encountered), the L part was not required for power or as an additional delay source. My guess is MITS got a good deal on surplus L parts as they fell out of favor in response to the release of the much better LS parts.

    Mike

  10. #10
    Join Date
    Apr 2013
    Location
    Albuquerque
    Posts
    106

    Default

    Ok, so the 8v line measures 9.8v with no boards in. the 5v line measures 5.1v at pin 3 of the 7805 regulator.

    Also, with the power off
    C-4 to CPU data 2 measures 0.50 on the ohm meter
    C-6 to CPU data 3 measures 0.49 on the ohmmeter
    D-4 to CPU data 4 measures 0.36 on the ohmmeter
    D-2 to CPU data 5 measures 0.49 on the ohmmeter
    Now as for sign tracing, I'll have to dig up my logic probe from storage as I think I accidentally put it there when I moved recently because I can't find it.

    A few side notes to answer a few of your questions, no, the replica 8080 board does not lave an 74L00, it uses a 74F00N made by Motorola. Same chip, different era. As for all of the address lines not lighting up, they do. I reseated the CPU board and fired it up, and they all light up with the memory board out as you can see in the video I just posted here. I put the CPU board in the other 3 remaining slots to make sure they didn't have a short either, and they all light up in the same way, so I think the backplane is good. As for an alternate RAM board to try, I do have one more, but its not a MITS board, and its a 16k DRAM board. So in other words, I have no static RAM board to try. In the video, you'll see some wires attached to a couple switches on the display board that normally wouldn't be there, but that's because there were some destroyed traces as my dad tried to replace a couple of switches with non pc mounting pinned ones, tried to drill out the board to make the pins fit and took the traces with it. So I had to run wires from the now proper switches to the remains of the traces since I couldn't solder the pins to the board anymore. LOL
    Last edited by k1atn; August 11th, 2019 at 04:21 PM.

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
  •