Loading ...
Sorry, an error occurred while loading the content.

EconoRAM VIIA Issues

Expand Messages
  • Kyle Owen
    I recently got a few MK4104 4k by 1 bit SRAM chips to replace a few bad MM5257s on my CompuPro EconoRAM board. I also wanted to fill out the entire RAM board,
    Message 1 of 7 , Oct 23, 2012
    • 0 Attachment
      I recently got a few MK4104 4k by 1 bit SRAM chips to replace a few bad MM5257s on my CompuPro EconoRAM board. I also wanted to fill out the entire RAM board, so I needed 10 total chips, 8 for the 4k bank, and 2 to replace the bad 5257s. After putting it all back together, I noticed that my 4k bank at $0000 is testing bad. The odd thing is, it is very intermittent and the same bytes that test bad one time may test good next. I wrote the RAM test utility as a part of the ROM monitor that I'm using. It writes $A6 and $6A respectively, to test all bits for 1 and 0. All other RAM tests just fine. The board itself is divided into 4 banks: two 4k banks and two 8k banks. Each can be positioned anywhere in RAM so long as it starts at 

      Here's the weird thing. The same bank of RAM, when positioned at anything but $0000, works just fine. The other 4k bank in position $0000 has the same issues as the first 4k bank.

      But things get weirder: either 8k bank at $0000 works just fine. So all in all, I'm not sure what the heck is going on. I can't really tell from the schematic what component would only fail when the card has either 4k bank jumpered to $0000. 


      Any ideas would be very much appreciated!

      Thanks,

      Kyle
    • B. Degnan
      You may have bad jumper switches for the 8K bank, causing a conflict. See if you have similar problems with other RAM combos that include the same jumper used
      Message 2 of 7 , Oct 23, 2012
      • 0 Attachment
        You may have bad jumper switches for the 8K bank, causing a conflict. See
        if you have similar problems with other RAM combos that include the same
        jumper used to set for 0000 (are they really off then they are set to off).


        It could be that the replacement RAM is not as fast as the original, but I
        assume you have ruled that out.

        My guess. I think I have one or two of these cards, they're usually
        reliable.

        Bill

        -------- Original Message --------
        > From: "Kyle Owen" <kylevowen@...>
        > Sent: Tuesday, October 23, 2012 8:37 PM
        > To: midatlanticretro@yahoogroups.com
        > Subject: [midatlanticretro] EconoRAM VIIA Issues
        >
        > I recently got a few MK4104 4k by 1 bit SRAM chips to replace a few bad
        > MM5257s on my CompuPro EconoRAM board. I also wanted to fill out the
        entire
        > RAM board, so I needed 10 total chips, 8 for the 4k bank, and 2 to
        replace
        > the bad 5257s. After putting it all back together, I noticed that my 4k
        > bank at $0000 is testing bad. The odd thing is, it is very intermittent
        and
        > the same bytes that test bad one time may test good next. I wrote the
        RAM
        > test utility as a part of the ROM monitor that I'm using. It writes $A6
        and
        > $6A respectively, to test all bits for 1 and 0. All other RAM tests just
        > fine. The board itself is divided into 4 banks: two 4k banks and two 8k
        > banks. Each can be positioned anywhere in RAM so long as it starts at
        >
        > Here's the weird thing. The same bank of RAM, when positioned at
        anything
        > but $0000, works just fine. The other 4k bank in position $0000 has the
        > same issues as the first 4k bank.
        >
        > But things get weirder: either 8k bank at $0000 works just fine. So all
        in
        > all, I'm not sure what the heck is going on. I can't really tell from
        the
        > schematic what component would only fail when the card has either 4k
        bank
        > jumpered to $0000.
        >
        > The manual for this RAM board can be found here:
        >
        http://www.hartetechnologies.com/manuals/CompuPro/CompuPro%20Ram%20VIIa_2.pd
        f
        >
        > Any ideas would be very much appreciated!
        >
        > Thanks,
        >
        > Kyle
      • Mike
        A couple of thoughts 1) some types of ceramic caps used for decoupling can lose their effectiveness with age. They can be brought back to life with an
        Message 3 of 7 , Oct 24, 2012
        • 0 Attachment
          A couple of thoughts

          1) some types of ceramic caps used for decoupling can lose their effectiveness with age. They can be brought back to life with an appropriate applicate of heat.
          2) I created an SRAM test fixture for my Apple 2. Makes testing SRAMs relatively painless. You can find the design on my superproto paper in the experiments, section. So far I've only done up to 1k SRAMs, but it could be extended.
          www.willegal.net/superproto

          Regards,
          Mike w

          --- In midatlanticretro@yahoogroups.com, Kyle Owen <kylevowen@...> wrote:
          >
          > I recently got a few MK4104 4k by 1 bit SRAM chips to replace a few bad
          > MM5257s on my CompuPro EconoRAM board. I also wanted to fill out the entire
          > RAM board, so I needed 10 total chips, 8 for the 4k bank, and 2 to replace
          > the bad 5257s. After putting it all back together, I noticed that my 4k
          > bank at $0000 is testing bad. The odd thing is, it is very intermittent and
          > the same bytes that test bad one time may test good next. I wrote the RAM
          > test utility as a part of the ROM monitor that I'm using. It writes $A6 and
          > $6A respectively, to test all bits for 1 and 0. All other RAM tests just
          > fine. The board itself is divided into 4 banks: two 4k banks and two 8k
          > banks. Each can be positioned anywhere in RAM so long as it starts at
          >
          > Here's the weird thing. The same bank of RAM, when positioned at anything
          > but $0000, works just fine. The other 4k bank in position $0000 has the
          > same issues as the first 4k bank.
          >
          > But things get weirder: either 8k bank at $0000 works just fine. So all in
          > all, I'm not sure what the heck is going on. I can't really tell from the
          > schematic what component would only fail when the card has either 4k bank
          > jumpered to $0000.
          >
          > The manual for this RAM board can be found here:
          > http://www.hartetechnologies.com/manuals/CompuPro/CompuPro%20Ram%20VIIa_2.pdf
          >
          > Any ideas would be very much appreciated!
          >
          > Thanks,
          >
          > Kyle
          >
        • Bill Sudbrink
          Are you seeing garbage when you get an error or $A6 when you expect $6A (or vice-versa)? This might clue us in to the level of the problem. Garbage would
          Message 4 of 7 , Oct 24, 2012
          • 0 Attachment
            Are you seeing "garbage" when you get an error or $A6
            when you expect $6A (or vice-versa)? This might clue
            us in to the "level" of the problem. Garbage would
            indicate a complete read or write failure, while the
            reversal error would be more likely to indicate an
            address decode error. I would look very carefully for
            bent pins. I have seen board where the pin was
            crumpled directly beneath the chip, sometimes making
            electrical contact with the socket, sometimes not. I
            don't have time to look at the schematic now, I'll see
            if I can make time at lunch.

            Bill S.
          • Kyle Owen
            Here s something I ve noticed: if I have the board configured as 8k at $0000, 8k at $2000, 4k at $4000, and 4k at $5000, everything works fine. It doesn t
            Message 5 of 7 , Oct 24, 2012
            • 0 Attachment
              Here's something I've noticed: if I have the board configured as 8k at $0000, 8k at $2000, 4k at $4000, and 4k at $5000, everything works fine. It doesn't matter which 4k or 8k bank is where; all configurations like the above work very well.

              When I configure it with either 4k bank at $0000, things begin to fail. Only the 4k bank at $0000 has problems, and again, that's either 4k bank. When either bank is positioned anywhere else, it works just fine. When either 8k bank is positioned at $0000, it works fine. But when any 4k bank is at $0000, writes (or maybe reads) become messed up. 

              And by messed up, I mean sometimes it'll write, and sometimes it won't. I believe the problem is solely based on writing, as the reading back shows some $A6s and some $6As after my RAM test, where good RAM will show all $6As. There are no signs of bent pins, and all chips have been reseated. And again, in the 8k-8k-4k-4k configuration starting at $0000, everything is working fine. Only in a 4k-4k-8k-8k or similar approach, with any 4k bank at $0000, do things begin to not work. 

              I'll take some screenshots of the RAM tests when I get back from work. 

              Kyle

              On Wed, Oct 24, 2012 at 6:52 AM, Bill Sudbrink <wh.sudbrink@...> wrote:
               

              Are you seeing "garbage" when you get an error or $A6
              when you expect $6A (or vice-versa)? This might clue
              us in to the "level" of the problem. Garbage would
              indicate a complete read or write failure, while the
              reversal error would be more likely to indicate an
              address decode error. I would look very carefully for
              bent pins. I have seen board where the pin was
              crumpled directly beneath the chip, sometimes making
              electrical contact with the socket, sometimes not. I
              don't have time to look at the schematic now, I'll see
              if I can make time at lunch.

              Bill S.


            • B Degnan
              Crazy. -- Sent from my PDP 8/e.
              Message 6 of 7 , Oct 24, 2012
              • 0 Attachment
                Crazy.
                --
                Sent from my PDP 8/e.
              • Bill Sudbrink
                Sorry I was not able to look at the schematics and I only have a minute or two to write this email. Hopefully, tomorrow will be less crazy. Anyway, let me
                Message 7 of 7 , Oct 24, 2012
                • 0 Attachment
                  Sorry I was not able to look at the schematics and I only
                  have a minute or two to write this email. Hopefully,
                  tomorrow will be less crazy. Anyway, let me understand
                  your memory test program. Are you writing all $A6 and then
                  turning around and writing all $6A? I had the (clearly
                  wrong) impression that you were writing alternating $A6
                  and $6A. What I would suggest is that you write using
                  the following algorithm:

                  <byte value> = <address high byte> + <address low byte>

                  So the memory at $0000 to $00FF would contain $00 to $FF.
                  $0100 to $01FF would contain $01,$02,$03...$FD,$FE,$FF,$00
                  $0200 to $02FF would contain $02,$03,$04...$FE,$FF,$00,$01

                  etc.

                  What this pattern will allow you to do is to determine, to
                  some degree, whether blocks of memory are being misaddressed
                  in some manor, whether responding to reads when they should
                  not or recording writes when they should not. I forget if
                  you mentioned what sort of chassis this card is in. If an
                  IMSAI, you should be able to examine specific addresses with
                  the front panel and the address lines will be held. With
                  the machine in this state, you can check the various bank
                  select and other address decode circuitry on the card with
                  a logic probe (even just an LED and a current limiting
                  resistor) and get some idea of what is going on. I'm not
                  sure if an Altair will do the same (unfortunately, I don't
                  have a full front panel Altair... I have a couple of turn-key
                  units). I have to go, hopefully tomorrow will be calmer.

                  Bill S.
                Your message has been successfully submitted and would be delivered to recipients shortly.