System Machine check abort has ocurred (firmware problem?)

From: Ricardo del Cueto <ricardo_at_r.iie.org.mx>
Date: Wed, 10 Jul 1996 10:14:15 -0600

Hello,

I have an Alpha DECStation 3000/400. As soon as I turn it on, instead of
the usual messages I get:

        System Machine check abort has ocurred

        PC=001134E0
        PS=00000000 0016F998

        MCHK+000 = 00000000 ( ... and many messages like this)


Hitting the reset button I can get the >>> prompt.

At the >>> prompt I can test the different components of the system (MEM,
SCC, NI, etc.). All of the tests run OK, all of them but the CXT test.
When I test CXT I get the system machine check error that I just wrote up
here. If I run "show conf" or "boot" I get the Machine Check error.
I can run "show device", "show error" with no problem. "init" seems to
do nothing (nothing is displayed).

Looking the manual I cannot find what CXT is for. There is an explanation
for MEM, SCSI, NI, SCC, etc., but CXT is not in the manual. Does anybody
knowns what CXT stands for?

The system was running fine. We were trying to upgrade DU from 1.3 to 3.2.
When we where at version 2.0 we updated the firmware. It was 3.2 and we
updated it to 6.1. We followed the procedure to upgrade the firmware. We
turned off the system and wait the 15 to 20 seconds recommended. We turned
the system on and nothing came out to the screen. So we followed the procedure,
we installed an alternate console (a video terminal) and with the
alternate console we where able to get the Machine Check error described at
the beggining of this message and also to run the tests at the >>> prompt.

Using the alternate console (the video terminal) the diagnostic display lights
show the hex number E1 when the Machine Check error is displayed. When we hit
the reset button, the display lights change to DD and the >>> is displayed
in the terminal.

Using the graphic monitor, instead of the alternate console (with the alternate
console switch in the up position), nothing is shown in the monitor and the
diagnostic display lights show the hex number EE. If we hit the reset button
nothing happens and the diagnostic display lights doesnt change and the >>>
prompt is not shown.

Useless to say that we are "desperados".


Thanks to all of you in advance for your help.


Ricardo del Cueto
Electrical Research Institute
delcueto_at_iie.org.mx
Received on Wed Jul 10 1996 - 17:59:32 NZST

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:46 NZDT