[comp.unix.ultrix] How to Cook a DECstation

ddr@flux.isr.alaska.edu (Donald D Rice) (06/22/91)

After the building air handlers failed yesterday, the temperature in my
office got up into the 90's overnight, primarily due to the VR299 monitor,
though the attached DS5000+peripherals no doubt helped.

The machine kept running with no apparent problems, but I got to wondering
about the rated operating temperatures.  In the DS5000 Operator's Guide,
under System Unit Operating Conditions, the temperature range is given as
10-40 C, 50-104 F.

Does this imply that the machines can operate indefinitely at 104 F with no
ill effects, or simply that they will not Halt and Catch Fire in this range?
Folklore says that operating temperatures above room temperature, ~ 70 F,
reduces the MTBF even if it doesn't crash the machine outright, and I've
seen the operators shut down the main campus machines if they can't keep the
temperature below about 90 F.  But from an engineering standpoint, it isn't
clear that a 30 F rise translates into a significant MTBF impact.

So, I'd like to know if DEC has anything more definitive on the operating
temperature ranges of DECstations.  Is the entire 50-104 F range really
thought to be satisfactory, or is running these guys at the high end of the
range seriously frowned upon?

It's summer in Fairbanks when it is 90 F during the day and they play baseball
at midnight without any artificial lighting, all the stores have sales that
start at midnight, and when you go biking at midnight you see people sitting
out on their lawns reading the newspaper...

Don
-- 
Don Rice                                  Internet: ddr@flux.isr.alaska.edu
Geophysical Institute                     E-mail:   fnddr@alaska.bitnet
University of Alaska                      Phone:    (907) 474-7569
Fairbanks, AK 99775                       Loran:    64.86N 212.16E

jg@crl.dec.com (Jim Gettys) (06/22/91)

I believe the specs are correct as written...

Our machines are often used in harsh environments; they get tested
in test chambers at elevated temperature extensively.  Think about
a factory floor sometime.  Machine room machines are often more
sensitive, and often when your machine room airconditioner allows
the temp. to get to 90, it is broken, and this forces the shutdown
before the machine room gets too hot.  The other thing to be avoided
which stresses machines is rapid temperature change; again, if a machine
room airconditioner fails, the temperature often rises quickly, and
disk drives can become unhappy.

Most of our desktop machines have thermostatically controlled fans these
days; the fans run faster when the machine gets hotter, to keep the
operating temperature of the components relatively constant.
(i.e. hot conditions causes you machine to get noisier!).

Just keep the airflow to the machine unrestricted, and check that
everything you have is similarly rated, but I believe we mean
what we say in the specs...
				- Jim

ajc@thendara.pa.dec.com (AJ Casamento) (06/22/91)

	Don,

	  As Jim Gettys has already pointed out, the system boxes on our
	DS5000/xxx machines have thermostatically controlled fans. This
	allows them to speed up when the temperature in the box gets higher.
	It also allows them to run very slowly/quietly when they are not
	needed. However, I posted your inquiry to the Reliability Engineers
	just out of curiousity, and Jim was also correct in that we really
	do mean what the specifications say.

	  To paraphrase my colleagues, and with NO intent to imply any 
	warranty other than that explicitly in the documentation that is
	provided with the system (ie. this discussion is ONLY that, just
	a discussion):


	* The DS5000/xxx reliability numbers are generated at a target of
	  25 C but are verified at 40 C.

	* While the system will probably crash at ~50 C (122 F) it is not
	  likely that the componenets will be permanently damaged below 
	  ~55 C (131 F).

	* The other potential for generating such damage is if you block 
	  the ventilation of the system (say by placing it on it's side
	  or putting books along it).

	* For the system to be in danger of catching fire, it would have
	  to be heated to well over 200 C (400 F). The plastics in the
	  system box, keyboard and monitor are flame retardent and are
	  not intended to support combustion once the flame has been removed
	  (ie. if you hold a torch to it you can make it burn, but if you
	  take the torch away, the flames should go out).


	  While I am not a reliability engineer myself, I would be surprised
	if the MTBF numbers (Mean Time Between Failures) remains constant 
	when the machine is run at 40 C on a continual basis. That is why
	ruggedized versions of the systems are made, to support the application
	needs of industrial/military systems.

	  So, if I've bored you enough, I hope this information satisfies
	your curiousity.

				Thanx,
				  AJ


      **********************************************************************
      * AJ Casamento			"The question is not whether or    *
      * Digital's TRI/ADD Program	 not the opinions are mine; but    *
      * 100 Hamilton Ave. UCO1-B	 rather, which of my personalities *
      * Palo Alto, CA 94301-1616	 do they belong to?"		   *
      * 415.853.6744							   *
      * ajc@decwrl.dec.com						   *
      **********************************************************************

ddr@flux.isr.alaska.edu (Donald D Rice) (06/25/91)

In article <AJC.91Jun22104209@thendara.pa.dec.com> ajc@thendara.pa.dec.com (AJ Casamento) writes:
>
>	Don,
>
>	  As Jim Gettys has already pointed out, the system boxes on our
>	DS5000/xxx machines have thermostatically controlled fans. This
>	allows them to speed up when the temperature in the box gets higher.
>	It also allows them to run very slowly/quietly when they are not
>	needed. However, I posted your inquiry to the Reliability Engineers
>	just out of curiousity, and Jim was also correct in that we really
>	do mean what the specifications say.

Thanks, both to you and Jim Gettys.  I saw Jim's message on Friday, but yours
came along after a network hub overheated and quit.  Some PC clones also lost
their marbles, but the DECstations did fine.  One person on the sunward side
of the building shut down his DECstation when the room temperature got near
104 F, but the rest of us survived by turning off the monitors to keep the
room temperature under 100.  The network hub stayed dead all weekend and
finally resumed operation this morning (not DEC equipment; I didn't buy it).

I guess the moral is that even in Alaska it pays to buy equipment with a
decent operating temperature range...especially when you work in a building
with a paleolithic air handling system.

Thanks again.
Don
-- 
Don Rice                                  Internet: ddr@flux.isr.alaska.edu
Geophysical Institute                     E-mail:   fnddr@alaska.bitnet
University of Alaska                      Phone:    (907) 474-7569
Fairbanks, AK 99775                       Loran:    64.86N 212.16E