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

Re: [Altair Computer Club] Re: Altair 8800 Fan

Expand Messages
  • Craig Landrum
    ... During the late 70 s and early 80 s our company developed Z-80 based Multibus boards that drove peripherals for our microfilm/microfiche scanning systems.
    Message 1 of 17 , Sep 10, 2012
    • 0 Attachment
      On 9/8/12 at 11:32 AM, eberhard@... (mfeberhard) wrote:

      >:-) I discovered that boards *usually* survived if you pulled
      >them out when the power was on. usually...
      >
      >Martin
      During the late 70's and early 80's our company developed Z-80 based
      Multibus boards that drove peripherals for our microfilm/microfiche
      scanning systems. As the head of the firmware team, we were constantly
      yanking boards and putting them back into the bus to test our latest
      EPROMs. We had our own low-level "OS" that had common routines for
      doing all bus-related I/O and board-to-board communications and
      I recall
      actually coding fault-tolerance into those boards to the point
      that we
      could yank or insert a board into the bus and it would not
      disrupt the
      normal flow of data or kill the machine. Of course Multiubus was a
      regulated bus and S-100 was unregulated so that may have made a big
      difference. In any case, we never fried a board in the 6 years that
      I worked on those systems.

      We also (almost) never fried a chip from static (we were in VA
      and static
      can be a problem in winter). Fried one or two of those over the years,
      but it wasn't enough to make us wear the wristbands, which we considered
      to be much more trouble than they were worth. That situation may have
      been different if we had been working extensively with CMOS
      parts, but
      we weren't.

      Killed a lot of EPROMS though. We would cycle through a 2716 or
      2732 EPROM
      anywhere from 50 to 100 times, and eventually we would either
      break off a
      pin or just wear the thing out. Once an EPROM failed to program
      we would
      stick it into the drop-ceiling tiles directly over the
      development system
      we were using - sort of an EPROM graveyard. We used Genrad
      Futuredata 2600
      systems which were built like tanks, but were terrific
      development mchines.
      I believe I have the only currently operational Futuredata
      system left
      in existence. 32K RAM and dual 8-inch floppies. Still boots
      and I can still
      program EPROMS with it which I used to when playing with my
      IMSAI :-)

      --
      Craig Landrum
      Chief Technical Officer
      mindwrap, inc.
      Phone: (540) 347-2552 x 229
      Fax: (540) 347-2556
      email: craigl@...
    • steve
      Same story here, Craig. MITS techs and assemblers didn t normally wear wrist straps, but no real static damage was evident, even with NM s dry air. Probably
      Message 2 of 17 , Sep 10, 2012
      • 0 Attachment
        Same story here, Craig. MITS techs and assemblers didn't normally wear wrist straps, but no real static damage was evident, even with NM's dry air. Probably because most of the floors were bare concrete.

        The main reason that you shouldn't pull S-100 cards out of their sockets is because the +18 and -18 volt pins are directly opposite each other, and quickly pulling the board out may allow those 2 socket contacts to spring together and touch, which will weld the contacts together, smoke the power supply, and probably vaporise a few traces on the bottom side of the mother board. There are some old discussions about this in the message archives. The Multibus is better laid out and the sockets are different, which I guess is why you could get away with hot swapping in your Z80 system. I wonder if there is such a thing as an S-100 ZIF socket.

        Steve
        ================================

        --- In altaircomputerclub@yahoogroups.com, Craig Landrum <craigl@...> wrote:
        >
        > On 9/8/12 at 11:32 AM, eberhard@... (mfeberhard) wrote:
        >
        > >:-) I discovered that boards *usually* survived if you pulled
        > >them out when the power was on. usually...
        > >
        > >Martin
        > During the late 70's and early 80's our company developed Z-80 based
        > Multibus boards that drove peripherals for our microfilm/microfiche
        > scanning systems. As the head of the firmware team, we were constantly
        > yanking boards and putting them back into the bus to test our latest
        > EPROMs. We had our own low-level "OS" that had common routines for
        > doing all bus-related I/O and board-to-board communications and
        > I recall
        > actually coding fault-tolerance into those boards to the point
        > that we
        > could yank or insert a board into the bus and it would not
        > disrupt the
        > normal flow of data or kill the machine. Of course Multiubus was a
        > regulated bus and S-100 was unregulated so that may have made a big
        > difference. In any case, we never fried a board in the 6 years that
        > I worked on those systems.
        >
        > We also (almost) never fried a chip from static (we were in VA
        > and static
        > can be a problem in winter). Fried one or two of those over the years,
        > but it wasn't enough to make us wear the wristbands, which we considered
        > to be much more trouble than they were worth. That situation may have
        > been different if we had been working extensively with CMOS
        > parts, but
        > we weren't.
        >
        > Killed a lot of EPROMS though. We would cycle through a 2716 or
        > 2732 EPROM
        > anywhere from 50 to 100 times, and eventually we would either
        > break off a
        > pin or just wear the thing out. Once an EPROM failed to program
        > we would
        > stick it into the drop-ceiling tiles directly over the
        > development system
        > we were using - sort of an EPROM graveyard. We used Genrad
        > Futuredata 2600
        > systems which were built like tanks, but were terrific
        > development mchines.
        > I believe I have the only currently operational Futuredata
        > system left
        > in existence. 32K RAM and dual 8-inch floppies. Still boots
        > and I can still
        > program EPROMS with it which I used to when playing with my
        > IMSAI :-)
        >
        > --
        > Craig Landrum
        > Chief Technical Officer
        > mindwrap, inc.
        > Phone: (540) 347-2552 x 229
        > Fax: (540) 347-2556
        > email: craigl@...
        >
      • Jack Rubin
        ... Yes - I have one on one of my extender cards - makes life very easy, but I still power down (unless I forget). And just powering down doesn t instantly
        Message 3 of 17 , Sep 10, 2012
        • 0 Attachment
          >
          > The main reason that you shouldn't pull S-100 cards out of their
          > sockets is because the +18 and -18 volt pins are directly opposite each
          > other, and quickly pulling the board out may allow those 2 socket
          > contacts to spring together and touch, which will weld the contacts
          > together, smoke the power supply, and probably vaporise a few traces on
          > the bottom side of the mother board. There are some old discussions
          > about this in the message archives. The Multibus is better laid out
          > and the sockets are different, which I guess is why you could get away
          > with hot swapping in your Z80 system. I wonder if there is such a
          > thing as an S-100 ZIF socket.
          >
          > Steve


          Yes - I have one on one of my extender cards - makes life very easy, but I
          still power down (unless I forget). And just powering down doesn't instantly
          discharge those beer-can size caps. Jerry Pournelle claimed his cat once
          knocked the power plug of his CompuPro system (Ezekial?) out of the wall and
          he was able to plug it back in before the computer powered down. Of course,
          Pournelle was a fiction writer...

          Jack
        • Craig Landrum
          ... Multibus was also a regulated bus - i.e. the voltages were regulated before they hit the bus instead of the S-100 way where they were regulated on each
          Message 4 of 17 , Sep 11, 2012
          • 0 Attachment
            On 9/10/12 at 5:54 PM, alltare@... (steve) wrote:

            >
            >Same story here, Craig. MITS techs and assemblers didn't
            >normally wear wrist straps, but no real static damage was
            >evident, even with NM's dry air. Probably because most of the
            >floors were bare concrete.
            >
            >The main reason that you shouldn't pull S-100 cards out of
            >their sockets is because the +18 and -18 volt pins are directly
            >opposite each other, and quickly pulling the board out may
            >allow those 2 socket contacts to spring together and touch,
            >which will weld the contacts together, smoke the power supply,
            >and probably vaporise a few traces on the bottom side of the
            >mother board. There are some old discussions about this in the
            >message archives. The Multibus is better laid out and the
            >sockets are different, which I guess is why you could get away
            >with hot swapping in your Z80 system. I wonder if there is
            >such a thing as an S-100 ZIF socket.
            >
            >Steve


            Multibus was also a regulated bus - i.e. the voltages were
            regulated before
            they hit the bus instead of the S-100 way where they were
            regulated on each
            individual board. Also, Multibus had the advantage of learning
            from S-100's
            experience :-)

            --
            Craig Landrum
            Chief Technical Officer
            mindwrap, inc.
            Phone: (540) 347-2552 x 229
            Fax: (540) 347-2556
            email: craigl@...
          • steve
            I believe that MITS thinking was that it was cheaper and more reliable to have small inexpensive regulators on each board than it was to have a large common
            Message 5 of 17 , Sep 11, 2012
            • 0 Attachment
              I believe that MITS' thinking was that it was cheaper and more reliable to have small inexpensive regulators on each board than it was to have a large common regulated supply that powered everything.

              Switching power supplies were not common back then, and they were expensive. It was all done with discreet components, too, which made them a lot more complex and harder to troubleshoot. For that matter, I wouldn't want to have to troubleshoot a modern-day switcher for a PC either.

              Steve
              ==================================

              --- In altaircomputerclub@yahoogroups.com, Craig Landrum <craigl@...> wrote:
              >
              > ...
              >
              > Multibus was also a regulated bus - i.e. the voltages were
              > regulated before
              > they hit the bus instead of the S-100 way where they were
              > regulated on each
              > individual board. Also, Multibus had the advantage of learning
              > from S-100's
              > experience :-)
              >
              > --
              > Craig Landrum
              > Chief Technical Officer
              > mindwrap, inc.
              > Phone: (540) 347-2552 x 229
              > Fax: (540) 347-2556
              > email: craigl@...
              >
            • Tom at Home
              Yes, less expensive for each card to provide its own regulator, and to spread heat from regulator loss (minimize hotspots inside the box) And yes, switchers
              Message 6 of 17 , Sep 12, 2012
              • 0 Attachment

                Yes, less expensive for each card to provide its own regulator, and to spread heat from regulator loss (minimize hotspots inside the box)

                And yes, switchers were expensive, complex, used lots of discrete parts (no switcher ICs then) and not widely implemented in commercial grade products in 1975.  Mostly implemented in larger, higher power regulators.

                 

                Tom Durston

                 


                From: altaircomputerclub@yahoogroups.com [mailto: altaircomputerclub@yahoogroups.com ] On Behalf Of steve
                Sent: Tuesday, September 11, 2012 1:55 PM
                To: altaircomputerclub@yahoogroups.com
                Subject: [Altair Computer Club] Re: Altair 8800 Fan

                 

                 



                I believe that MITS' thinking was that it was cheaper and more reliable to have small inexpensive regulators on each board than it was to have a large common regulated supply that powered everything.

                Switching power supplies were not common back then, and they were expensive. It was all done with discreet components, too, which made them a lot more complex and harder to troubleshoot. For that matter, I wouldn't want to have to troubleshoot a m odern-day switcher for a PC either.

                Steve
                ==================================

                --- In altaircomputerclub@yahoogroups.com, Craig Landrum <craigl@...> wrote:

                >
                > ...
                >
                > Multibus was also a regulated bus - i.e. the voltages were
                > regulated before
                > they hit the bus instead of the S-100 way where they were
                > regulated on each
                > individual board. Also, Multibus had the advantage of learning
                > from S-100's
                > experience :-)
                >
                > --
                > Craig Landrum
                > Chief Technical Officer
                > mindwrap, inc.
                > Phone: (540) 347-2552 x 229
                > Fax: (540) 347-2556
                > email: craigl@...
                >

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