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

Re: [Altair Computer Club] Example Altair 680 Troubleshooting

Expand Messages
  • Dan
    good job There is usually a systematic approach in troubleshooting systems. You certainly followed a logical approach. Given a few known good conditions(eg.
    Message 1 of 9 , May 26, 2007
    • 0 Attachment
      good job
      There is usually a systematic approach in troubleshooting systems.
      You certainly followed a logical approach.
      Given a few known good conditions(eg. power, reset, etc),
      you can skip a few steps and it helps to have sockets too :)
      Also LEDs never fail to offer clues.

      =Dan


      [ My Corner of Cyberspace http://ragooman.home.comcast.net/ ]



      Grant Stockly wrote:
      >
      > I don't really know the correct way, or if there is one, to diagnose
      > these
      > old computers but I have been lucky with the last few Altair boards
      > that I
      > copied. : ) When I copy a board there is the chance that a bridge is left
      > out, inserted where there was none, etc.
      >
      > This is about a real vintage Altair 680. When powered up there is no
      > monitor prompt. The computer does not work. A5 and D2 are stuck low.
      >
      > I'm going to write this message as I do it so I don't forget anything or
      > the order. : )
      >
      > I first removed the 1702 ACIA EPROM to see if it was the cause. It
      > isn't. : (
      >
      > I'm going to remove the 6850 now. Nope. : (
      >
      > I'm going to REPLACE the 6800 now. Nope. : (
      >
      > I just noticed that A5 lights up the Altair is NOT halted. D2 (BD2)never
      > lights.
      >
      > The scope shows BD2 is pretty hard to ground. A continuity meter says 15
      > ohms. Maybe a buffer or memory chip is to blame. When trying to deposit
      > with the front panel there is "data" on the line that only comes up to 1v
      > or so.
      >
      > RP3 ties some lines including BD2 to ground, so I removed it. Nope. : (
      >
      > BD2 goes to IC WW next. Removed it, short is gone! A continuity check
      > when desoldered on the IC from pin 8 (GND) to 6 (BD2) shows 13.5 ohms
      > resistance. : )
      >
      > Forgot to resolder RP3 before checking it.
      >
      > BD2 can now be deposited. All ram chips check out (walking 1s test).
      >
      > Removing board to install RP3, 6850, and ACIA.
      >
      > Crossing my fingers... ; )
      >
      > WOOHOO!!! :
      >
      > >From my terminal:
      >
      > .M 0000 00 10
      > .M 0000 10
      >
      > In conclusion, the 680 was actually pretty hard to debug compared to the
      > Altair 8800. The lack of status lines on the 6800 compared to the 8080
      > made it a little harder for me. Also, the failure modes of the 6800 are
      > VERY strange. For example, it can get in some kind of loop where where
      > the
      > HLT line doesn't actually halt the processor any more!?! The processor
      > seems pretty unpredictable compared to the 8080. Maybe its just me...
      >
      > Grant
      >
      >
      > ------------------------------------------------------------------------
      >
      > No virus found in this incoming message.
      > Checked by AVG Free Edition.
      > Version: 7.5.472 / Virus Database: 269.8.0/819 - Release Date: 5/26/2007 10:47 AM
      >
    • Grant Stockly
      ... That is one thing that wasn t quite straight forward. Once the 6800 gets an invalid opcode all bets are off on what happens next. The reset and halt
      Message 2 of 9 , May 26, 2007
      • 0 Attachment
        At 07:06 PM 5/26/2007, you wrote:


        >good job
        >There is usually a systematic approach in troubleshooting systems.
        >You certainly followed a logical approach.
        >Given a few known good conditions(eg. power, reset, etc),
        >you can skip a few steps and it helps to have sockets too :)
        >Also LEDs never fail to offer clues.

        That is one thing that wasn't quite straight forward. Once the 6800 gets
        an invalid opcode all bets are off on what happens next. The reset and
        halt inputs to the 6800 didn't seem to do anything at times!!! I was
        watching the halt line while watching the data lines... Nothing! Because
        B2 was stuck low it was getting invalid opcodes I guess. The MITS
        documentation has a "fix" for the front panel board that will allow it to
        reset the 680 if it gets an invalid opcide. It involves cutting one trace
        on the back (non switch side) and two on the front. The addition of a
        10ohm resistor and a .1u cap. I'm not sure of the details yet...

        I was very lucky to have both the 680 Turnkey and "regular" front
        panels. The turnkey is worthless for troubleshooting.

        I'm troubleshooting another 680 right now, but it seems to be VERY
        sick. I'm going to hook up the scope and check the 2 phase clock... Its a
        digital storage scope so I will upload the results...

        Grant
      • jack99rubin
        ... diagnose these ... that I ... Since this is such an important topic, let me replay it and ask for clarification in a few places where I lost you in the
        Message 3 of 9 , May 27, 2007
        • 0 Attachment
          --- In altaircomputerclub@yahoogroups.com, Grant Stockly <grant@...>
          wrote:
          >
          > I don't really know the correct way, or if there is one, to
          diagnose these
          > old computers but I have been lucky with the last few Altair boards
          that I
          > copied.

          Since this is such an important topic, let me replay it and ask for
          clarification in a few places where I lost you in the narrative - I'm
          just a solder-monkey hoping to advance to tech!

          First, let me make sure I understand the problem based on symptoms
          seen - A5 and D2 are "stuck low" and you get no monitor prompt. So
          you've already verified internal power levels and distribution. Do
          you know A5 and D2 are low based on the front panel - you can't get
          the LEDs to light? Or are you already actually scoping (with logic
          analyer?) the address and data lines? Are you attaching your leads at
          the CPU or at the front panel?


          > I'm going to write this message as I do it so I don't forget
          anything or
          > the order. : )
          >
          > I first removed the 1702 ACIA EPROM to see if it was the cause. It
          isn't. : (
          >
          > I'm going to remove the 6850 now. Nope. : (

          ... and since it is really one of several I/O options, not necessary
          for operation with the front panel? Was this the logic in removing
          the ACIA prom as well?

          >
          > I'm going to REPLACE the 6800 now. Nope. : (

          did you mean to say you replaced the 6850, or did you remove the 6800
          prior to starting the process (step zero?)?

          >
          > I just noticed that A5 lights up the Altair is NOT halted. D2
          (BD2)never
          > lights.

          OK, so something changed here? A5 was not lighting before but now
          does, while D2 remains low? If so, was that because the 6800 was
          replaced while the 6850 remained out?


          >
          > The scope shows BD2 is pretty hard to ground. A continuity meter
          says 15
          > ohms. Maybe a buffer or memory chip is to blame. When trying to
          deposit
          > with the front panel there is "data" on the line that only comes up
          to 1v
          > or so.

          This is really interesting - "pretty hard" sounds very "analog" - I'd
          been thinking that a line is either high or low - seeing a voltage in
          the "undefined" range would be unexpected - is this a common? I don't
          think this would even trigger a LED-based logic probe.


          > RP3 ties some lines including BD2 to ground, so I removed it.
          Nope. : (

          I'm following along on the schematic here - RP3 is a 4.7K resistor
          pack connected to Vcc - so wouldn't it be pulling lines high?

          >
          > BD2 goes to IC WW next. Removed it, short is gone!

          OK, so I may be really anal here but I'm trying to fully grok this -
          when did you move from D2 to BD2? D2 also goes through one of the
          resistor packs - did you remove them all? Is an IC more likely to
          fail than the resistor pack or does it make sense to follow the path
          of least resistance (hahaha) by removing socketed ICs instead of
          soldered resistor packs?

          > when desoldered on the IC from pin 8 (GND) to 6 (BD2) shows 13.5
          ohms
          > resistance. : )

          WW is a 74367 hex bus driver - so out of the circuit, all pins should
          be isolated from each other, while in circuit, pin 6 and 7 will
          either be open or connected based on the state of pin 1? There should
          _never_ be a low impedance pathway between Vcc (pin 16), GND (pin 8)
          and any other pins, right? With a "regular" multimeter in continuity
          mode will 13.5 ohms show as open or closed?

          >
          > Forgot to resolder RP3 before checking it.
          >
          > BD2 can now be deposited. All ram chips check out (walking 1s
          test).
          >
          > Removing board to install RP3, 6850, and ACIA.
          >
          > Crossing my fingers... ; )

          I can do this!

          >
          > WOOHOO!!! :
          >

          and I look forward to doing that!

          Thanks for your generosity and patience in sharing your work! I'll be
          bringing my 680b to VCF next month - hope to exchange notes with Bill
          Degnan. Will you be there too or is that just an urban myth?

          best,
          Jack
        • Dan
          Yea, I noticed that weirdness in the Reset and Halt circuit. Basically it s wired in synchronous mode (versus the proper method which should be asynchronous).
          Message 4 of 9 , May 27, 2007
          • 0 Attachment
            Yea, I noticed that weirdness in the Reset and Halt circuit.
            Basically it's wired in synchronous mode (versus the proper method which
            should be asynchronous).
            That mod from MITS that you mentioned is just a basic method of wiring
            the reset line.
            I also have the reset modified on mine from back then.

            It a creates an RC time constant together with the resistor and cap.
            When the switch is pressed it will shunt the cap to ground forcing the
            reset signal to Lo.
            The time constant begins when the switch is released and then the cap
            begins to chargeup.
            Then the reset input of the cpu will remain Lo (due to the rise time)
            for a certain amount of clock cycles to allow it to propagate throughout
            the entire chip. --There is a minimum time period required, about 20
            cycles I think. Also when the system is first turned on, to allow the
            power supply to settle.

            I always design any microprocessor with a minimum of this circuit, then
            there's additional reset features that can be added to this.

            =Dan

            [ My Corner of Cyberspace http://ragooman.home.comcast.net/ ]



            Grant Stockly wrote:
            >
            > At 07:06 PM 5/26/2007, you wrote:
            >
            > >good job
            > >There is usually a systematic approach in troubleshooting systems.
            > >You certainly followed a logical approach.
            > >Given a few known good conditions(eg. power, reset, etc),
            > >you can skip a few steps and it helps to have sockets too :)
            > >Also LEDs never fail to offer clues.
            >
            > That is one thing that wasn't quite straight forward. Once the 6800 gets
            > an invalid opcode all bets are off on what happens next. The reset and
            > halt inputs to the 6800 didn't seem to do anything at times!!! I was
            > watching the halt line while watching the data lines... Nothing! Because
            > B2 was stuck low it was getting invalid opcodes I guess. The MITS
            > documentation has a "fix" for the front panel board that will allow it to
            > reset the 680 if it gets an invalid opcide. It involves cutting one trace
            > on the back (non switch side) and two on the front. The addition of a
            > 10ohm resistor and a .1u cap. I'm not sure of the details yet...
            >
            > I was very lucky to have both the 680 Turnkey and "regular" front
            > panels. The turnkey is worthless for troubleshooting.
            >
            > I'm troubleshooting another 680 right now, but it seems to be VERY
            > sick. I'm going to hook up the scope and check the 2 phase clock... Its a
            > digital storage scope so I will upload the results...
            >
            > Grant
            >
            >
            > ------------------------------------------------------------------------
            >
            > No virus found in this incoming message.
            > Checked by AVG Free Edition.
            > Version: 7.5.472 / Virus Database: 269.8.0/819 - Release Date: 5/26/2007 10:47 AM
            >
          • Grant Stockly
            ... I did both. There are actually two D2 s in the Altair. One for the processor and another called BD2 , the buffered one. I used the front panel to
            Message 5 of 9 , May 27, 2007
            • 0 Attachment
              >First, let me make sure I understand the problem based on symptoms
              >seen - A5 and D2 are "stuck low" and you get no monitor prompt. So
              >you've already verified internal power levels and distribution. Do
              >you know A5 and D2 are low based on the front panel - you can't get
              >the LEDs to light? Or are you already actually scoping (with logic
              >analyer?) the address and data lines? Are you attaching your leads at
              >the CPU or at the front panel?

              I did both. There are actually two "D2"s in the Altair. One for the
              processor and another called "BD2", the buffered one. I used the front
              panel to quickly narrow in on what the problem was and then investigate it
              with the scope.

              >... and since it is really one of several I/O options, not necessary
              >for operation with the front panel? Was this the logic in removing
              >the ACIA prom as well?

              Yes, the front panel in the 680 needs no memory or I/O devices installed to
              work. If there is not at least RAM or ROM then you can't actually examine
              anything, it would be all FF.

              The 6850 itself couldn't cause the address lines (other than A0) to go bad
              because they all go through some LS04s and LS30s first. But it does
              connect to the buffered data bus.

              > > I'm going to REPLACE the 6800 now. Nope. : (
              >did you mean to say you replaced the 6850, or did you remove the 6800
              >prior to starting the process (step zero?)?

              I replaced the 6800 with a new one. I left the 6850 out. I don't think
              the front panel will work without a processor. I'll try........ Nope.

              > > I just noticed that A5 lights up the Altair is NOT halted. D2
              >(BD2)never lights.
              >OK, so something changed here? A5 was not lighting before but now
              >does, while D2 remains low? If so, was that because the 6800 was
              >replaced while the 6850 remained out?

              Nope. This was a weird one. A5 would not light up doing an examine while
              halted but when I flipped the run switch it would glow. D2 would
              NEVER. So I decided to focus on D2 first and then figure out why A5 was
              acting up. : )

              > > The scope shows BD2 is pretty hard to ground. A continuity meter
              >says 15 ohms

              >This is really interesting - "pretty hard" sounds very "analog" - I'd
              >been thinking that a line is either high or low - seeing a voltage in
              >the "undefined" range would be unexpected - is this a common? I don't
              >think this would even trigger a LED-based logic probe.

              Well, the prohlem was kind of analog. It wasn't grounded in the sense that
              it was bridged or connected directly to ground. The buffers on the front
              panel were strong enough to create a little 1v or so blip for depositing to
              BD2, but thats it (the blim was on the input to the buffer, the buffer was
              not buffering). The other data lines make it up to 4.5-5v during a
              deposit. The short inside the 74367 was later measured at 15ohms in
              circuit, and 13.5 ohms out of circuit. That is what I meant by "pretty hard".

              > > RP3 ties some lines including BD2 to ground, so I removed it.
              >Nope. : (
              >
              >I'm following along on the schematic here - RP3 is a 4.7K resistor
              >pack connected to Vcc - so wouldn't it be pulling lines high?

              Yep...brain fart...but I have a nice desoldering pump so it only takes me
              30 seconds to pop an IC off with no damage. : ) The resistor pack wasn't
              a big deal. I can't even tell it was ever off.

              > > BD2 goes to IC WW next. Removed it, short is gone!
              >
              >OK, so I may be really anal here but I'm trying to fully grok this -
              >when did you move from D2 to BD2? D2 also goes through one of the
              >resistor packs - did you remove them all? Is an IC more likely to
              >fail than the resistor pack or does it make sense to follow the path
              >of least resistance (hahaha) by removing socketed ICs instead of
              >soldered resistor packs?

              D2 as shown on the front panel, BD2 on the buss. I thought I had edited
              the e-mail to say D2 as written on the front panel and BD2 on the bus. D2
              was also getting shorted because of the bad buffer, but from the scope I
              could tell that it was BD2. I guess I didn't write everything I did. : (


              > > when desoldered on the IC from pin 8 (GND) to 6 (BD2) shows 13.5
              >ohms
              > > resistance. : )
              >
              >WW is a 74367 hex bus driver - so out of the circuit, all pins should
              >be isolated from each other, while in circuit, pin 6 and 7 will
              >either be open or connected based on the state of pin 1? There should
              >_never_ be a low impedance pathway between Vcc (pin 16), GND (pin 8)
              >and any other pins, right? With a "regular" multimeter in continuity
              >mode will 13.5 ohms show as open or closed?

              All of the other pins measured as you would think. The Fluke 111 that I
              have does pass voltage to measure resistance, but in this case it was only
              6 and 8 that were shorted together. I just tested the chip again and it
              measured 13.7 ohms with the positive lead on 8 and negative on 6 AND with
              the leads switched around. I don't know if that means something important
              or not. Maybe it means that its not a silicon short.

              >Thanks for your generosity and patience in sharing your work! I'll be
              >bringing my 680b to VCF next month - hope to exchange notes with Bill
              >Degnan. Will you be there too or is that just an urban myth?

              Yep. : )

              Grant
            • Derek J. Lassen
              I can t remember the opcode, but there was a 6800 instruction that I gave the mnemonic HCF for halt and catch fire. Motorola didn t document it. If executed,
              Message 6 of 9 , May 27, 2007
              • 0 Attachment
                I can't remember the opcode, but there was a 6800 instruction that I gave the mnemonic "HCF" for halt and catch fire. Motorola didn't document it. If executed, and if power wasn't dropped within a few seconds (20 -30 ?), The processor would turn into a 16 bit down counter forever more.

                (s) Derek

                At 08:07 PM 5/26/2007 -0800, you wrote:

                At 07:06 PM 5/26/2007, you wrote:

                >good job
                >There is usually a systematic approach in troubleshooting systems.
                >You certainly followed a logical approach.
                >Given a few known good conditions(eg. power, reset, etc),
                >you can skip a few steps and it helps to have sockets too :)
                >Also LEDs never fail to offer clues.

                That is one thing that wasn't quite straight forward. Once the 6800 gets
                an invalid opcode all bets are off on what happens next. The reset and
                halt inputs to the 6800 didn't seem to do anything at times!!! I was
                watching the halt line while watching the data lines... Nothing! Because
                B2 was stuck low it was getting invalid opcodes I guess. The MITS
                documentation has a "fix" for the front panel board that will allow it to
                reset the 680 if it gets an invalid opcide. It involves cutting one trace
                on the back (non switch side) and two on the front. The addition of a
                10ohm resistor and a .1u cap. I'm not sure of the details yet...

                I was very lucky to have both the 680 Turnkey and "regular" front
                panels. The turnkey is worthless for troubleshooting.

                I'm troubleshooting another 680 right now, but it seems to be VERY
                sick. I'm going to hook up the scope and check the 2 phase clock... Its a
                digital storage scope so I will upload the results...

                Grant

              • Grant Stockly
                Another tip with the 680 is watch out for the power switch. While plugging in the serial cable I gave myself a little shock. I m fairly grounded where I am
                Message 7 of 9 , May 27, 2007
                • 0 Attachment
                  Another tip with the 680 is watch out for the power switch. While plugging
                  in the serial cable I gave myself a little shock. I'm fairly grounded
                  where I am and it hurt!

                  Be safe guys! : )

                  Grant
                • billdeg@aol.com
                  In a message dated 5/27/2007 9:07:53 AM Eastern Daylight Time, ... Jack, Speaking of Bill Degnan I have been working on my 680 for the past two days. I found
                  Message 8 of 9 , May 28, 2007
                  • 0 Attachment
                    In a message dated 5/27/2007 9:07:53 AM Eastern Daylight Time, 
                    jack.rubin@... writes:
                    
                    
                    > Thanks for your generosity and patience in sharing your work! I'll be > bringing my 680b to VCF next month - hope to exchange notes with Bill > Degnan. Will you be there too or is that just an urban myth? > > best, > Jack
                    Jack, Speaking of "Bill Degnan" I have been working on my 680 for the past two days. I found a copy of VTL (Very Tiny Language) for the Altair 680, and I am working to enter the entire thing into my computer manually, for the experience. The program fits on a 1K 680 with ACIA loaded as well. I understand how to enter values into an address, but I am a little shaky on how to read the code listing and translate some of the shorthand notations. I will be bringing my 680 to VCF E, along with my TRS 80 Mod 1 exhibit. Bill


                    **************************************
                    See what's free at http://www.aol.com.
                  Your message has been successfully submitted and would be delivered to recipients shortly.