Re: ALPHA 255, in the intensive care unit
From: Miguel Angel Usón Finkenzeller (mauf2_at_unizar.es)
Date: 10/27/04
- Next message: David Froble: "Re: Question about Descriptors"
- Previous message: Keith A. Lewis: "Re: ALPHA 255, in the intensive care unit"
- Maybe in reply to: Miguel Angel Usón Finkenzeller: "ALPHA 255, in the intensive care unit"
- Next in thread: David Mathog: "Re: ALPHA 255, in the intensive care unit"
- Reply: David Mathog: "Re: ALPHA 255, in the intensive care unit"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Wed, 27 Oct 2004 19:22:33 +0200
>Miguel Angel Usón Finkenzeller wrote:
>
>>
>> The famous 2.2 microFarad capacitor next to the mouse and keyboard outlets
>> seems allright,
>
>Tested or just visual inspection?
>
>> the power supply puts out the expected +-12, +5 and 3.3
>> volts, all the connectors a were and are again firmly attached, and yet...
>
>I've recently had a series of failures on Athlon systems that were very
>similar to this. Each power supply seemed ok, but when it was swapped
>out with a new unit, the systems stopped crashing. Apparently
>some power supplies can fail by passing too much noise or just
>glitching from time to time, but these always test at the correct
>voltage with a voltmeter (or the system's own sensors, if these are
>present.) Admittedly this is a bit unusual, more typically power
>supplies just die (sometimes spectacularly). In any case, the power
>supply is far and away the most likely component to fail in a way
>that would produce these sorts of symptoms. (The disk is probably
>more likely to fail but VMS is usually pretty clear about that
>and it shows up in the error logs - power supply failures just
>crash the system instantly.)
>
>>
>> One thing I find strange is that the CPU cooling fan receives only 7.0
>> volts. Is that what it should be? As said, the 12v reach the motherboard.
>
>Your voltmeter may be confused trying to measure the DC voltage across
>the running fan motor. Does the fan appear to be spinning at a normal
>speed? If you're concerned try replacing the fan (fans are cheap).
>Also, you should check that the heat sink on the CPU isn't packed up
>with dust. I'm not familiar with the heat sink in a 255 but sometimes
>one must remove the fan on top to get a good look at the heat sink.
>
>>
>> Any hints are appreciated.
>
>It doesn't sound too much like a heat problem if the system
>crashes quickly from a cold start. But you didn't say that
>exactly, only that it it crashes on restart, and the system
>may have been (too) hot at the time. Take off the top of the
>case, aim a big floor fan into it, turn the system off
>and let it cool down for 15 minutes. Then start it up with
>the case still open and the external fan running and see if runs
>without crashing. Caveat: what's the ambient room temperature?
>Caveat 2: don't plug the floor fan into the same circuit as the computer!
>
>Another possibility - noise or spikes through the power lines.
>You're at a university and who knows what hideous equipment is
>plugged in nearby. Who knows when said equipment may be running
>or not. Try plugging the Alpha into an uninterruptible
>power supply or some other device that will feed it clean power.
>If that isn't available try it elsewhere on campus or at home
>where the power is less likely to be contaminated by heavy
>electrical equipment.
>
>
>Regards,
>
>David Mathog
>mathog@caltech.edu
>If you're afraid the CPU might get too hot, you can check it with:
>f$getsyi("temperature_vector")
>
>I can't remember under which OpenVMS version this became available, but I
>beclieve it's in the 7.x somewhere.
>
>--
>Syltrem
>
Thanks a lot for your hints. However, I did not update over VMS v7.1
(because of TCPIP services) and that lexical is unrecognized (and is not
documented in HELP Lexicals, nor any alternative lexical).
Besides, I would not know what the proper temperature of the CPU should be
like...
The capacitor was visually inspected and, the tested by a technician (I do
not how). The 7.0 volts were measured with the fan unplugged: before
startup, the supply is 0.6v (!), and after some 1.5 minutes it raises to
those 7.0, when the heat sink (which has a sensor attached) gets warmer. I
do not know if a "second speed" (higher voltage) should set in.
As suggested above, I will try another fan. It is not only cheap but also
easy: the fan lies aside and in front of the CPU; the heat sink is clean
(all the inside of the box has been thoroughly cleaned on a "regular" 9-15
months basis).
Sometimes the system crashes after a "cold start" (say, immediately after 3
hours powered off and with the power cord unplugged).
Room temperature is 22 +- 2 degrees (although, for the first two years out
of eight, this machine has had to suffer temps ranging from +10 to +35). I
have set the room fancoil down to 20 +-1 (pretty cold now, I will have to
wear somehow heavier clothes than a short sleeved shirt). The top cover is
still off. An UPS is out of question, as is carrying the Alpha home (I'd
love having an Alpha at home, but the marshals would catch me long before I
leave the Campus).
Thank you, again
Miguel Ángel Usón
>>------------
>Dept. of Inorganic Chemistry
>University of Zaragoza
>E-50009 SPAIN
- Next message: David Froble: "Re: Question about Descriptors"
- Previous message: Keith A. Lewis: "Re: ALPHA 255, in the intensive care unit"
- Maybe in reply to: Miguel Angel Usón Finkenzeller: "ALPHA 255, in the intensive care unit"
- Next in thread: David Mathog: "Re: ALPHA 255, in the intensive care unit"
- Reply: David Mathog: "Re: ALPHA 255, in the intensive care unit"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
