Fiber Tracker Meeting Minutes
December 09, 2003 - 10am, farside
- FT Status and Plans : Borcherding
- letter from Dmitri
- General
- Bias problem
- Holes in readout - 8, 7, 6 VLPC chips per SVX
- Appeared after Oct 13 warm up
- Readout error problem
- Readout errors when Z-sup = even value
- AFE boards
- Have a good number of spares
- Setting 'type' for 6 more boards -
chart
- Lowering priority for fixing MCM's
- VRBC
- Several new version boards available
- New firmware available
- Will be testing on MCH2
- CTS Status
- Operational
- Will use to qualify AFE, SEQ, VRB, ...
- Source of HOT spares
- Still requires triggers for TFW
- Working on specifying changes for Ted Z.
- CFT Status : Warchol
- talk slides
- changed "unbiased" definition to "dead"
- Dead channels list
- cftx7 170/6400
- cftx8 72/7040
- 6B6 SVX7 list and plot
- 6 alive channels spread across 2 chips
- so there is bias, we just lost the channels
- 6B6 SVX7 4 plots
- black is standard bias+1V (not red)
- Dead channels and Readout problems - 2 plots
- holes corresponding to dead and problems
- top plot is SVX readout
- bottom plot is discriminator output - readout problems do not show up
as holes in this plot
- is it a system readout problem, or something else?
- this has not happened before: take a good calibration run,
but then when running normally we see holes
- remember that "dead" is not permanent - of 9 chips listed dead back
in August, 6 are now alive again
- VLPC Cryo : Rucinski
- 149 dead in east; 21 dead in west
- picture shown of feed end without vacuum shell
- in east cryo, He input comes in below level of liquid
nitrogen due to construction error
- east was built first, so there was a learning curve
- thermal cycles have been about the same for both
- east probably a little leakier, but both kept at positive pressure
- west runs a little colder, even with same set point
- operationally
- east has a little drift - set point has to be lowered 0.5K per month
- if do a warmup to 50K or more, original set point is reestablished
- west relief valve seems to operate more often than east, but this is
probably not significant as total leak rate is very small for both
- east has only been pumped and backfield twice; west has been done 5 times
-> there might be a little bit of water left in east
- plots from George
- plots showed temperature history of cryo
- Jadzia's change seems to have happened during the long room-temperature
warmup during past shutdown
- we also had a warmup during detector studies around Oct 31;
during the period an odd cooling scheme was used -
cool with nitrogen only
- other warmup around 11/15 from power outage from wind damage -
again odd cooling scheme, using nitrogen only
- this one had some problems getting west cryo back down to desired
temperature during recovery
- remedy: warmed both up to ~20K, then both cooled back down fine
(don't know why this worked; wasn't intended to be a fix)
- could be a correspondance between some gasket leaks in lid heaters
and the problem areas
- George: this shutdown was chaotic compared to past ones -
it's hard to figure out everything that happened
- contamination issues
- contaminants: oxygen, nitrogen, water
- expected contamination since January
- oxy: (1/6)(252cm^3) -> about 1 mil (0.014in)
- nitro: (1/6)(1880cm^3) -> about 1 mil (0.014in)
- water: 12cm^3 -> about 0.0006in
- might get maximum of 10 mils on the bottom of the cryo
- don't see how contamination could get inside the copper cup
(where chip is, where contaminate would have to get to cause problems
like we're seeing
- showed picture of cold block
- could try warming up to liquid nitrogen temperature, do a couple back-fills,
and cool back down.
- could also hook up an oxygen monitor to the gas output to monitor the rate
as warmup is done
- sketch made during discussions
- Testing readout problems: Rubinov
- cranked up vref by 15 to put pedestals above threshold
- see some that were dead that are no longer completely dead,
but also see others that are worse
- just tried something: defeated the 53MHz clock gen stuff
- this did fix things
- defeat = pretty much reverted to previous setup
- problem is not easy to track down as it does not show up everywhere
- What should be done now?
- try to look at it more upstairs and get a handle on it there
- see how much delay needs to be put in to break it?
- Paul: also want to due an LED run downstairs
- Data Quality Monitoring : Corcoran
- talk slides
- might be some bugs in program,
getting some non-physical numbers
- George is interested in the dead/hot lists output of David's program to see
when channels listed as dead in January came back to life
- Run Quality Criteria
- not sure yet what fraction of data will be labeled as bad
- what should criteria be in future?
- continue to run this program? run modified program? just use recocert?
- David is putting his program into RecoCert
- plan is to run RecoCert for every run
- group being formed to take care of this and look at output
- timescales for code into RecoCert? group forming?
- longer than a few weeks
- current studies should be continued in the meantime
- Reminder comments from George
- should ask Russ to prepare in case we want a warmup
- if you are going to "try something", be thinking about it because we'd
want to do it within the week
- will we reduce thresholds on the dead channels to create "fake hits"
to send to Level 3? general concensus seems to be no
