[23562] in Hotline Meeting
e40-008 AC alarm
daemon@ATHENA.MIT.EDU (David Krikorian)
Sat Jul 30 06:03:51 1994
Date: Sat, 30 Jul 94 06:03:42 -0400
From: David Krikorian <dkk@MIT.EDU>
To: hotline@MIT.EDU
Cc: burke@MIT.EDU, jis@MIT.EDU, cec@MIT.EDU
Cc: op@MIT.EDU, larugsi@MIT.EDU
This message contains a combination of Lou's experiences and mine
during recent hours. It raises some issues of Phys Plant's dealings
with us (ie: shutdown notification). The problem described below has
not been resolved. Managers: this message is long, so you might want
to skim or skip down to the "ISSUES" section.
------------
CHRONOLOGY
Lou discovered that the nearer air conditioning unit in the e40-008
machine room (serial number 708073) had a high humidity alarm. (He
also discovered something else unusual, which I'll describe
separately, below.)
At 11:45pm last night (Friday, 7/29/94) he called Physical Plant to
investigate the problem. No one had responded by 12:15am today
(Saturday, 7/30/94 -- 30 minutes after the first call) so Lou called
Plant again. Soon afterward, a maintenance person named Ron showed up
and began investigating. (I don't know his last name.) Lou paged Ops
at about 12:30am. (Due to some confusion in the pager schedule
change-over, it took an extra 15 minutes or so for notification to
reach me. Sorry Anne!)
I returned to campus and joined Lou at about 1am. Lou stayed until
2am (when it appeared the problem would be resolved), and checked back
in from home at 3am.
------------
TROUBLE
Lou determined that the AC unit had no heat. Heat is used to reduce
the humidity. (Ron, from Plant, said the heat is used only to raise
the temperature of the air, so the amount of water stays the same as
without the heat, but sincw the air's capacity for holding water goes
up, the humidity goes down.) The troublesome AC unit in 11-135 also
had had no heat, and kept tripping the alarm as a result.
The AC problem in e40-008 bore enough of a resemblance to the one in
11-135, that Lou had Plant check on the status of the heat supply to
E40. They determined it was shut off. (They said it was a regular
seasonal shutoff.) The Plant men finally agreed (with the
intervention of a supervisor) to turn on the low pressure steam, and
thus the heat to the AC unit. That was when Lou left for the night.
(2am) I left Ron my pager number and went upstairs to get something
done while he worked.
At 2:16am I was paged by Ron Beaulieu (a different Ron from Plant),
who explained that they couldn't turn on the heat because there was
some shutoff in a manhole outside the building, and they needed a
confined space permit (or some such) to work there. They said they
would file a work order for the heat to be turned on by the people who
do things like work in manholes (who wouldn't be in until Monday), and
that they would check the machine room periodically to assure that
nothing further had gone wrong with the AC.
(As of this writing, there is no work order. Linda of the Operations
Center said the work order would be filed near 8am today.)
I'll be on or near campus all weekend (except when I find time to
sleep :-), and will check (in addition to Plant) that e40-008 is
running smoothly.
There was another symptom of trouble that Lou noticed at the beginning
of all this. When he walked into e40-008, he felt a mist blowing or
settling on face. The mist lasted for perhaps 10 minutes afterwards,
and covered the area just inside the main door and partially up the
ramp. A vent up on the right wall about 10 feet into the room seems
the most likely source. The mist effect did not return while we were
in the room, and we couldn't find any more clues about its nature.
------------
ISSUES
This incident brought up several issues which Lou requested I include
in this report. One of those is the central monitoring of equipment
alarms, whereby Plant should have already known that the AC unit was
alarming before Lou chanced upon it. Did they not get an alarm
notification? Did they choose to ignore it?
Another issue is the notification of shutdowns. Every time Plant is
going to shut down a fan for an hour, they send us email. But they
shut off heat and cooling to whole buildings on a seasonal basis
without notifying DCNS, either before or afterward. Two examples of
heat shutdown are buildings 11 and e40, where AC units appear to rely
year-round on the heat supply for proper operation. In both cases,
the heat is shut down for the summer without the knowledge of DCNS.
Even air-conditioning of areas with workstations and servers has been
shut down seasonally or daily. I don't know if this still happens,
but years ago Plant used to shut off the AC on E40-3 at night, even
though we had many computers (and some people) working through very
hot nights.
The upshot of the notification issue is that Lou and I both feel that
Plant should be notifying us of seasonal shutdowns of utilities, since
we rely on some of them.
