Conference 7.286::atarist

If you've been in the machine before then just open it up and push down FIRMLY
on all of the socketed chips.

If you think it may be heat related then try running the ST for a while (over
4 hours it sounds like) with the it wide open (ie: with the case and metal off
of it).  If it is a heat problem, this will either elliminate the problem or
at least greatly extent the on time before failure.

	Ray

Boy, I hate those RAM upgrades.  I'd be my lunch the problem would go
away if you removed the RAM upgrade.

    
    Dont forget the possibility of VIRII, my son brought one home on a disk
    he got from school that would do just what you describe. Despite my
    telling him to check for virii before running foreign disks!
    
    After a random time programs would bomb, or screen would go "spotty" or
    freeze. (or any combination)
    
    I had to virus kill all the floppys that had been thro the m/c since
    the time he ran the game on it, (within the same power on period, if you
    get my meaning), since I didnt know exactly which ones, I had to check
    the lot. Its a REAL PAIN!. If I could get my hands on the Virii writers
    I would find somewhere to put the infected floppies!
    
    
    PJK

    I've got a 520ST upgraded to 2.5 MB.
    
    It'll run fine for a couple of months, then I'll get two bombs or
    three bombs in the middle of some random process.
    
    When this happens, I take off the covers, and firmly press down on
    all the chips, but especially the MMU.
    
    Then it's usually good for another six months of heavy use.
    
    When I get rich, I'm going to buy a TT, and no more memory upgrades
    for me!
    

    I've not experienced the problem, but if it is a solderless upgrade...
    ie the kind where you have a multi-pin connector piggy-backed on the
    MMU... and if the ST is kept in a humid/dusty/whatever environment
    then could some kind of oxidisation be affecting the pins such that
    when you wiggle them about and reseat them a good connection is
    re-established (boy that was a long sentence... pause for a deep breath
    8^)
    
    I use a Frontier solderless board and as said earlier have had no
    problems with it.
    
    Just FWIW
    
    
    Jonathan

    Yes, it's exactly that kind of upgrade - there's a multi-pin connector
    that is inserted into the MMU socket, then the MMU chip is plugged into
    the upgrade board.
    
    I ran the memory diagnostic for 48 hours. On pass number 6586 there
    was an error in location 0027BBD4, which contained 5555 when it should
    have contained 1868.
    
    Then it ran fine for another 48 hours.
    
    While running the diagnostic that failed, the temperature in the room
    varied from 55 degrees to 70 degrees and back again - twice.
    
    Is there a way to turn a Solderless board into a soldered board?
    

Just to let any interested parties know;

I reseated the chips and have run for extended periods of time since then with
no problems. Apparantly I'll just have to regard this as routine maintenance
whenever the problem pops up from now on. I can live with that :^)

Steve

I do the same thing occasionally. I have one of the original ST's with a
solderless ram upgrade. When reseating chips, avoid touching any pins. If you
must touch a pin, make sure you're grounded. This is to prevent Electro
Static Discharge ( ESD ) from frying your components.

							John.

    Yeah, seating problems seem to be endemic with the early ST's - a
    friend and I certainly had them galore on our old 520STm's.  However,
    I've discovered a simple remedy that appears to fix them permanently!

    After about the third time that I had to take my ST apart to reseat
    things, I sat down with it and had a good hard look at both the QIP
    sockets and the chips themselves.  I was expecting to find poor quality
    sockets, or oxidation on the contacts and/or chip pins.  Well, I didn't
    find that; in fact I was pleasantly surprised by both the quality of
    the materials and the standard of construction in my ST.

    What I *did* find was that the chips were sitting too low in the
    sockets.  Instead of the raised parts of the socket contacts pressing
    firmly on the flats of the chip pins they were riding up on the curved
    section of the pins where they bend to go into the body of the chip. 
    Naturally, this greatly reduced the contact pressure and left the chips
    free to shift downwards in response to vibration and thermal
    expansion/contraction.

    To fix that I made some square cardboard spacers about 2mm thick out of
    an old shoebox and put them in the sockets, underneath the chips. 
    This also made the tops of the chips nicely flush with the tops of the
    sockets instead of recessed as before, so the spring retainer clips
    held the chips much more firmly.

    I've done this to four STm's now and none of them has ever acted up
    again, not even during some pretty serious thumping and dropping I
    subjected them to as part of checking them out after doing the fix.

    So give it a try, it's easy to do.  Just remember those ESD precautions!
    I'd strongly recommend that you borrow a conductive anti-static mat and
    wrist strap, if at all possible.  Perhaps some obliging Customer Service
    person will lend you one and show you the right way to use it.


    Bret.

Great tip, thanks.

The next time it acts up I'll do it that way. I've already got the wrist strap.

Steve

	Note .8 brought back memories for me...I remember many years ago, in the
days of wire-wrapped backplanes, CSS in Reading bought loads of backplane
sockets for their special projects...A similar problem was experienced in that
the harder the modules were pushed into the backplanes, the worse the contact
became because the contact pins were pushed away by the module...That cost us
a lot of money in rebuilds...Nice to know it happens to others as well, but
sorry it has to be Atari...

    One last follow-up:
    
    I've been using the machine extensively since the fix and have not had
    one reoccurence of the problem. I guess it is fixed!
    
    Steve