A DIRC Water Plant Digest for Commissioners
If you're looking for some information on the procedure to be followed after
a SOB dump, just go here.
This document does not aim at describing in a complete and detailled way the
DIRC Water plant. It just summarizes a few information gathered here and
there which may be useful for DIRC oncall experts. For a more precise and
technical view of this essential DIRC component, the water plant experts,
Jerry Va'Vra and Matt MacCulloch, should be contacted.
The DIRC water plant is a complex and sensitive hardware system: therefore,
DIRC oncall experts should not attempt to fix problems or
operate any of its components without specific instructions. Our main
jobs are the following:
- to acknowledge the pages related to this system;
- to make sure that the problem is real;
- to dispatch it to the water plant experts (Matt and Jerry) via page
or e-mail depending on its severity and on the required time response.
The most urgent problems which may occur are the following:
- a water dump (which would trigger epics alarms, automatic page
and a beta alarm on the pilot top left console);
- a clear sign of smoke (normally caught up by the hall VESDA system);
- a water leak.
Such event may be difficult to diagnose for non-experts.
If you have any doubt, page the ops. manager first and
then a water expert if needed.
Otherwise, nothing really serious should occur on a day timescale: the water
quality (transmission, bacteria level) deteriorates very slowly and the system
can run with the pump off during several hours.
In particular, if we get a page outside 'normal hours' (6am → 9pm on week
days and 9am → 9pm on weekends and holiday periods), we should estimate
whether the problem needs to be fixed urgently or if it can wait the next day
-- which is likely.
Other sections of the ops. manual are dealing with the pumps of the DIRC
water system. Read them carefully if one of these pumps stops working and
strictly follow the instructions they provide.
The water tank
Ultra-pure water is permanently circulating in the SOB whose volume is about
6000l. Filters are used to stop dust particles and to remove ions from water;
UV-light is used to to kill bacteria.
One pump makes the water circulating and the other one is degasing the water.
The nominal water level is around 5.4m -- this number is indirectly measured
via a Bernouilli-like relation. If the level would reach 5.3m/5.2m, it would
trigger a yellow/red alarm on the ALH.
A gauge is glued on the tank to show the interval in which the water level is
expected to fluctuate during normal operations. The water level is expected
to fluctuate for various reasons: temperature variations, BaBar doors being
opened or closed etc.
If the water level is found outside this range, the DIRC
water experts should be contacted. Before doing this, check for any
sign of water leak, either on the IR2 floor (if accessible!) or in the
moisture sensors. Those devices will likely trigger some alarm in case of a
real leak.
DIRC water tank
If the circulating pump stops, the water will start
decreasing in the SOB. This should trigger a red alarm in the ALH and the DIRC
oncall expert should be paged automatically. After the pump gets restarted,
the water level should come back to its nominal value in a few minutes.
To have a better understanding of the water level evolution with the
circulating pump off, a test was setup mid-November
2005
before the startup of Run 5b. The circulating pump was stopped manually and
the evolution of the water level and of other water plant parameters were
monitored during 5 or 6 hours. A detailled summary of the test can be found
here.
Its main results are the following:
after an initial drop of ~3-4 cm (the water level was 5.38m when the test
started), the water went down slowly at a rate of ~7-8 cm/h. The yellow alarm
level was reached after 50 minutes. At that level, the decrease rate was slowly going down. About half an hour later, the rate increased again; 20 minutes
later, the water level reached a plateau at ~5.22m, i.e. two centimeters (!) above the red alarm level. Then, the water level stayed stable during the next
three hours before the test was ended.
The water level evolution is thought to reflect the geometry on top of the
SOB. The plateau should correspond to the moment when the water reaches the
top of the main SOB volume -- in that area, you need to drop a large volume of
water to see a change in the measured level.
Therefore, if the circulating pump breaks during the night, the water
level should remain quite stable after an initial drop. Unless the red alarm
level is reached, fixing the problem should wait for the next morning.
The degasing pump is connected to
a SIAM as there is no other way to detect that it stopped working.
Therefore, its failure would immediately trigger a red alarm on the ALH.
The dump system
The infiltration of water (SOB water leaking in some bar box or moisture
condensation) could damage some components of BaBar, in particular the
EMC crystals. Therefore, the DIRC has designed and implemented a complex
system to avoid humidity (nitrogen flow in the bar boxes and on top of
the SOB). If a significant leak is detected, the water in the SOB would
be dumped in a few minutes -- to be compared to the refill time, about
10 hours without including the investigation required to understand and
fix the origin of the leak.
Actions on or involving the dump system should only be
done by the DIRC water experts.
Yet, if you're oncall and if you're wondering whether the water dump has
occured, there are two quick ways to check that it was not the case...
- Check the dump valve position. If it is the same as on the picture
below, no dump has occured. After a dump, the valve should have moved by
90 degrees.
DIRC dump valve
- To be more convinced, you could also touch the dump pipe and make sure
it is perfectly dry.
DIRC dump pipe
Water epics panel
The information on the water system is summarized in a dedicated epics panel,
reachable from the main BaBar epics panel: 'BaBar Online Detector Control'
→ 'DRC' button → 'Water'.
DIRC water epics panel
Its data reach the
drc-mon IOC through
VSAM boards located in this crate. A dead VSAM would have a red light on its
front panel and would prevent the IOC from booting. When you change a VSAM,
make sure that you're not disturbing the other VSAM connections which are
somewhat loose. If a connector gets unplugged, the information displayed on the
corresponding epics panel will be crazy.
The list of epics variables related to the water system (as of 2006/06/22) can be here.
Nicolas Arnaud
Last modified: Thu Jun 22 13:27:19 PDT 2006
