Path: utzoo!utgpu!water!watmath!clyde!att!pacbell!ames!umd5!hans
From: hans@umd5.umd.edu (Hans Breitenlohner)
Newsgroups: comp.sys.atari.8bit
Subject: Re: 130 XE Service problem
Keywords: 130XE service problem
Message-ID: <2760@umd5.umd.edu>
Date: 20 May 88 17:41:30 GMT
References: <670@lakesys.UUCP>
Reply-To: hans@umd5 (Hans Breitenlohner)
Organization: University of Maryland, College Park
Lines: 48

In article <670@lakesys.UUCP> rich@lakesys.UUCP (Rich Dankert) writes:
>[]
>
>	I have a 130 XE machine, that will not allow one to do 
>	anything except watch the RAM test in the self help 
>	diag's text. It goes there right off. Once in a while one 
>    ....
>
Fixing small computers 101:  (also applies to any other electronic
equipment):  Check the power supply.  Make sure it puts out 5 volts and
not much ripple.  A scope is great for this, but a VOM will do.
Actually this one does not sound like a power supply problem, but you
never can tell for sure.  And you always feel like kicking yourself when
you discover, after much searching, that the solution was that simple.

When doing a coldstart, one of the first things the XL/XE OS does is a
quick memory test.  If that fails, it dumps you into the memory diagnostics,
and it sounds just like that is what is happening to you.  The fact that you
get that far indicates that, in all likelyhood, the cpu, antic, gtia, parts
of the pia, and the rom are all working.  The RAM must be working at least
somewhat, since the memory test must use it for scratch storage, screen, etc.

You did not indicate if the memory test seemed to be running, and how it 
thought that the RAM and ROM areas checked out.  If you see some failures
there, you should have some idea of what is broken.  From the symptoms I
strongly suspect that your problem is bad RAM.

Assuming that RAM is your problem, the next step is to identify the
bad IC.  Unfortunately the diagnostics go to much detail identifying the
1k region containing the error, but neglect to identify the bit.  If your
display is somewhat garbeled, the pattern may help to identify the bad bit.
If your memories are socketed, swap away (you have a set of spares right 
there).

If your memories are soldered in, here is a trick I once used to fix just
such a beast:  take 3 memory chips, and bend their legs inward slightly, so
that you can stack them on top of each other, and have them hold together
tightly.  Put this stack on top of a memory chip, and see if the problem
goes away.  If not, move it to the next one, and so on.  Be sure to remove
power whenever you are moving the chips, be careful not to short anything,
and all usual precautions about handling static-sensitive components apply
here, too.

When I did this, it turned out that two memories were enough to "outvote"
the bad one, but one was not enough.  I wold use three in case your
error is more persistent.

I hope this helps!