[SUMMARY] Misbehaving DEC 3000-500X

From: Sean O'Connell <sto_at_stat.Duke.EDU>
Date: Mon, 2 Dec 1996 09:13:03 -0500 (EST)

Greetings-

I recieved several responses to my query regarding a DEC 3000-500X in
the department which is exhibiting some stability problems. It seems
that responses fell into two categories: 1) buy a new 3000-800X
motherboard (there is a flaw in the 3000-500X) or perform a sort of
exorcism involving pushing on all socketed chips and unplugging and
resetting all connections. In conclusion, I will probably pick a
weekend and try the latter first, as the former is the $5000.00+
solution. If the exorcism is sucessful, I will drop a note.

Many thanks to: Geert Jan Bex <gjb_at_luc.ac.be>
                Jo Knox <fxjwk_at_aurora.alaska.edu>
                Marc Thoelen <mthoelen_at_luc.ac.be>

Their responses were:

from Geert
> When we bought our central server something like three years back, we
> were among the first in Belgium to have one. About half a year later the
> machine got unstable. After doing some tests it turned out that it
> crashed due to heavy memory load (192 MB RAM installed, a 100 MB application
> consistantly crashed the machine).
>
> Digital tried to replace everything (powersupply, motherboard, SIMMs,
> you name it). Just before the waranty period ran out, they got word from
> headquarters that the 500X motherboard had a design error and they had
> to upgrade the machine to a 3000/800X.

from Jo
> That the problem is related to the temperature is a telling point! It
> suggests your problem is either a poor connection or a chip going bad,
> both of which could be affected by the temperature. If it's the latter,
> it'll only get worse over time, and you'll eventually bite the bullet
> and replace the board; if it's the former, you may be able to fix---very
> carefully! Get a grounding strap and keep yourself well grounded at
> all times, and firmly depress every plugged in chip on the board---only
> chips which are plugged in to a socket, NOT soldered on chips! If possible,
> support the back of the board while you're pressing on the front; you
> don't want to crack the board, but you do want to get some pressure on
> the chips. You can often do one side of the chip at a time.
>
> Also pull and replug any plug wires (if there are any; I've never been
> inside a 3000!)

from Marc
> We had similar problems 2 years ago with a 3500X. However the crashes
> only occurred when we tried to access the complete memory (i.c. 192MB).
> After numerous testing and board swapping, DEC admitted that something
> was wrong with the 3500X motherboard (the 3500 board we installed
> temporarily worked well), and eventually we received a 3800 motherboard,
> which we have now been using for 2 years without problems.

*************************************************************************
* Sean O'Connell *
* Computer Projects Manager *
* Duke University Institute of Statistics and Decision Sciences *
*************************************************************************
* Phone: (919) 684-5419 *
* Fax: (919) 684-8594 *
* Email: sto_at_stat.Duke.EDU *
* Mail: 220 Old Chemistry Building *
* P.O. Box 90251 *
* Durham NC 27708-0251 *
*************************************************************************
Received on Mon Dec 02 1996 - 15:36:27 NZDT

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:47 NZDT