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

1651Re: Gremlins!

Expand Messages
  • john_crane_59
    Feb 10, 2009
    • 0 Attachment
      It's a Seattle computer 16K static card. Since the little white
      blocks jump around randomly, I don't think it's a problem with a bad
      chip or socket. I think it's some sort of timing issue. When I get
      around to it, I'll solder in a couple more slots and put in a bus
      terminator. "Normal" apps work just fine. Except for the 8K BASIC
      problem which I'll discuss in another post. But that problem
      appeared before I installed the dazzler and additional motherboard
      plate and wires, etc.

      But you do have an otherwise excellent comment. Everytime I get a
      new card, I reseat the chips and sand away any "black plague"
      corrosion I see.

      I experimented with the new "DeoxIT" chemical. The residue is
      appearently a protectant/lubricant and is non-conductive. So I
      tried some on an old HP-25C calculator that had light corrosion on
      the PC traces. It seemed to work, but after 2 days it started
      acting flakey. I flushed and scrubbed the circuit card with alcohol
      and all was fine. It's a good chemical cleaner, but I wouldn't
      recommend leaving the residue on circuits with contacts. Anyone
      familair with the HP woodstock line will recall the many
      gold "fingers" connecting the keyboard to the logic board.
      Apparently, some of the residue seeped onto those contacts through
      time, resulting in erratic behavior.

      The best solution to the corrosion problem I've found is to run a
      dehumidifier in the room where you store your old goodies. If you
      need to work on something, it's best to remove the unit to another
      room, as the lowered humidity increases static potential.

      -John


      --- In altaircomputerclub@yahoogroups.com, "Steve" <alltare@...>
      wrote:
      >
      > That sounds like the kind of behavior that one would often see
      with
      > dynamic boards having refresh problems.
      >
      > Is it a MITS 16K static board? The memory chips on those boards
      had
      > a tendency to lift out of their sockets. LAy the board on a flat
      > surface and push down on each RAM chip to be sure they're all
      fully
      > seated. If you've never done this before, you'll be surprised at
      how
      > many are loose.
      >
      > Always put the memory board(s) as close as possible to the CPU
      board
      > to avoid buss ringing problems. Dazzler should work OK without
      buss
      > termination as long as all the boards are close together.
      >
      > steve
      > ==================================================
      >
      >
      > --- In altaircomputerclub@yahoogroups.com, "john_crane_59"
      > <john_crane_59@> wrote:
      > >
      > > I wondered when something like this would happen...
      > >
      > > I'm running a Dazzler in a 8800 with a 16K static RAM card.
      > > Kalideoscope works fine (apparently), but Life gives me
      trouble.
      > Of
      > > the dozens of times I've tried to run it, it worked properly
      only
      > > once. All the other times the lifeforms stop with a mixture of
      > blue
      > > and white. But the Life colors are red, green, blue, and
      black.
      > > Not white.
      > >
      > > As an experiment, I wrote a Dazzler program to repeatedly copy a
      > > block of known stable data (in ROM) down to the Dazzler mapped
      area
      > > for display. I saw little white blocks flitting about the
      colored
      > > quilt pattern like fireflies. I used my monitor to verify the
      two
      > > blocks of data, and I found errors: the white blocks were small
      > > strings of FFh 2-4 bytes in length. Evidently the copy operation
      > > fails randomly. But I knew it wasn't the code as I used a copy
      > > routine from my assembly library of known good code, well
      tested. I
      > > then changed the code to turn off the Dazzler before the memory
      > copy
      > > operation, then after the copy turn it on with a delay for
      > viewing.
      > > Now the screen flashed off an on very quickly, but I could still
      > see
      > > the pattern was perfect - no white fireflies, no changes at
      all.
      > So
      > > the memory card was not to blame, and the chips are well under
      the
      > > 1ms access time stated in the Dazzler manual. Apparently, the
      > memory
      > > copy operation is unreliable when the Dazzler is turned on.
      > >
      > > I have heard of the infamous bus "noise" problemns in the early
      > > 8800s. As it happens, I had to install another bus board to fit
      > the
      > > Dazzler into my system. Yes - with all those tiny wires! Could
      > > this be the culprit? When I installed the new bus board
      addition,
      > I
      > > did some basic testing of the system - ran several programs (non-
      > > Dazzler) and all looked good. But the Dazzler uses a high speed
      > DMA
      > > ( 1M bytes per sec.) Could this be asking too much from a
      working
      > > (but weak) bus design? Enough to push it over the edge to
      failure?
      > >
      > > If this is the problem, would a bus termintor card help the
      > > situation? I also have a one-piece MITS motherboard I could
      > > install, but I would rather not do that as it wasn't original
      spec
      > > for this model. (And I don't want any more nightmares of thin
      white
      > > spiders crawling up my soldering iron to bite my fingers!)
      > >
      > > And if this is a common enough problem in Altairs, I'm surprised
      > > this sort of thing isn't mentioned in the Dazzler manual.
      Cromemco
      > > is usually more thorough than that. But then, I could see Ed
      > > Roberts going ballistic over someone criticizing "his computer"
      in
      > > print. So I guess the silence makes sense in a weird sort of
      way.
      > >
      > > -J
      > >
      > > "All my Life is Dead!"
      > >
      >
    • Show all 6 messages in this topic