Run 8 Spin Database QA 

This page details the effort to extract information about fills to upload to the spin database for Run 8 in 2008.

 

Method:

 The method used for Run 8 is basically the same method used in 2006, see details here. In 2006, the cdev database (which is the recording of MCR broadcast of many machine parameters) was updated ~4 min, whereas for Run 8 the cdev database is updated every ~5 min. The script used to analyze the fills checks for discrepencies between the start of a fill as recorded by cdev and as recorded by the runlist. In the 2006 version of the code, discrepencies greater that 250 seconds were flagged as an error. The script used for Run 8 now allows discrepencies of up to 310 seconds in line with the 5 min cdev update interval. The version of the script used for Run 8 can be seen here.
 

Results:

I ran the analyzation script on the following run list. This produced a set of good fills which had no discrepencies and can be loaded into the database and a set of bad fills which had problems and need additional examination.

 

Bad Fills:

Here I list what was wrong with the bad fills and the steps I took to correct them.

F9817: This fill has a discrepency in spin data between beginning and end of fill. Fill contains four listed runs. All are Balewski_test/Physics and all are marked as questionable. First run occurs when there are no collisions(12:50:06) while other three runs occur while there are. See RHIC scalars. Spin pattern is likely not defined for period with no collisions leading to the discrepency. Can probably remove this fill from run list.

 

F9820: This fill has a discrepency between the fill number recorded in the database and found in the run list at the end of the fill. Fill contains one listed run. Run is FMStest and is marked as successful, however online run log has this run marked as a run crash. Can probably remove this fill from run list.

 

F9831: This fill has a discrepency in spin data between beginning and end of fill. Fill contains 14 listed runs. All are testPP2008-aL2 / Physics and all are marked as bad by the shift leader. Spin discrepency probably arises because first run occurs when there are no collisions(7:16:25) while the last run occurs while there are collisions. See RICH scalars. Can probably remove this fill from run list.

 

F9847: This fill has a discrepency in spin data between beginning and end of fill. Fill contains 2 listed runs. Both are EmcCheck and both are marked as questionable by the shift leader. Both runs appear to be during time with colliding beam(20:43:47 and 20:46:58) so not sure why there is a spin discrepency. See RICH scalars. Since these events are tests and questionable, can probably remove this fill from run list.

 

F9849: This fill has a discrepency between the fill number recorded in the database and found in the run list at the end of the fill. The RICH scalars seem to go dead at around 23:30 on 02-11-2008 and do not come back on again until 6:30 on 02-12-2008. The cdev database has F9850 starting between 23:41 and 23:46 on 02-11-2008, so when the scalars went out, the change of fill number must have not gotten recorded somewhere. According to the cdev database, runs 9042128-9042131 belong in F9849 while runs 9043001-9043062 belong with F9850. Will need to do more work to get these fills where they belong.

 

F9861: This fill has a discrepency in spin data between beginning and end of fill. Fill contains two listed runs. Both are ppProd2008 and both are marked bad by the shift leader. Both runs occur during period when there are no collisions(16:41 & 16:45), so spin data is probably not stable. See RICH scalars. Can probably remove from run list.

 

F9885: This fill has a discrepency in spin data between beginning and end of fill. Fill contains 16 listed runs of various kinds and completion status. The first two runs, 9047021 & 90947024 are marked as bad and occur at 2:27 and 3:00. This is before the collision rate becomes stable and is greater than 1h 45min before the next run(9047028) begins at 4:46. See RICH scalars. Removing these two runs from the run list takes care of the spin discrepency for this fill. So this fill has stable spin data between unix time 1203155195 to unix time 1203175726 (beginning of run 9047028 to end of run 9047047).

 

F9917: This fill has a discrepency in spin data between beginning and end of fill. Fill contains 10 listed runs. The first nine are all streamtest and are listed as questionable. The last run is ppProd2008 and marked as bad, but it comes more than 2 hours after the previous run. Can probably remove this fill from the run list as all runs are questionable tests or marked as bad.

 

F9938: This fill has a discrepency in spin data between beginning and end of fill. Fill contains 21 listed runs of various type and completion status. All runs occur during what appears to be a good fill with good, stable collision rates. The last run in this fill is a successful FMStest which starts at 3:26. See RICH scalars. Removing this run from the run list takes care of the spin discrepency. So this fill has stable spin data from unix time 1203733012 to unix time 1203754931 (beginning of run 9053128 to end of run 9054013).

 

F9948: This fill has a discrepency in spin data between beginning and end of fill. Fill contains 31 listed runs of various type and completion status. The first two runs(9056054 & 9056055) are both successful FMStest and occur at 11:23 and 11:28. This is before there are any collisions and over 2.5 hours before the next run(9056062) starts at 14:01. See RICH scalars. Removing these first two runs will fix the spin discrepency for this fill. So this fill has stable spin data from unix time 1203966094 to unix time 1203995033 (beginning of run 9056062 to end of run 9056104).

 

F9964: This fill has a discrepency in spin data between beginning and end of fill. Fill contains one listed run. The run is an FMStest and was marked as successful. Run starts at 18:09 which is before we have collisions. See RICH scalars. Can probably remove this fill from the run list as it only has one test run during a time with no collisions.

 

F9804: This fill has a discrepency between the fill number recorded in the database and found in the run list at the beginning of the fill. Fill contains one listed run. Run is Balewski_test/Physics and was marked as questionable by the shift leader. Run starts at 12:29 which is before we have collisions. See RICH scalars. Can probably remove this fill from the run list as it only has one test run during a time with no collisions.

 

F9811: This fill has a discrepency between the fill number recorded in the database and found in the run list at the beginning of the fill. Fill contains one listed run. Run is Balewski_test/Physics and was marked as a run crash. Can probably remove this fill from the run list as it only has one test run that crashed and has no counts.

 

F9827: This fill has a discrepency between the fill number recorded in the database and found in the run list at the beginning of the fill. Fill contains one listed run. Run is FMStest and is marked as bad by the shift leader. Can probably remove this fill from the run list as it only has one test run and it was marked as bad.

 

F9876: This fill has a discrepency between the fill number recorded in the database and found in the run list at the beginning of the fill, also has discrepency in spin data between beginning and end of fill. Fill contains four listed runs. First two runs(9045033 & 9045041) start at 11:06 and 11:35 before we start getting many collisions and both are marked as questionable. Last two runs(9045045 & 9045047) start at 12:04 and 12:22 and are both marked as bad. See RICH scalars. Can probably remove this fill from the run list as it has two questionable runs during times that probably have unstable spin information and two runs which are marked as bad.

 

F9877: This fill has a discrepency between the fill number recorded in the database and found in the run list at the beginning of the fill, also has discrepency in spin data between beginning and end of fill. Fill contains one listed run. Run was ppProduction2008/Physics and the shift leader marked it as bad. Can probably remove this fill from the run list as it has only one run which is bad.

 

F9904: This fill has a discrepency between the fill number recorded in the database and found in the run list at the beginning of the fill. Fill contains one listed run. Run was EmcCheck/Physics and was marked as questionable by the shift leader. Run starts at 12:56 in region with very few collisions. See RICH scalars. Can probably remove this rill from the run list.

 

F9951: This fill has a discrepency between the fill number recorded in the database and found in the run list at the beginning of the fill. Fill contains 16 listed runs of various type and completion status. The first run(9057060) is ppProduction2008/Physics and is marked as successful. This run starts at 12:41 which seems to put it in the previous block of collisions. The next run is 9057068 and is an EmcCheck/Physics marked as successful and starts at 14:15. See RICH scalars. Removal of run 9057060 from the run list fixes the discrepency. So this fill has consistent spin data from unix time 1204053320 to unix time 1204069433 (beginning of run 9057068 to end of run 9057087).

 

F0: This was a strange event. The fill listed as F0 contains one run(9062012) which starts at 2:07 on 03-02-2008(m-d-y). The run list has F0 being flanked before and after by run F9975. The 21 runs before run 9062012 and the 7 runs after run 9062012 belong to fill F9975. The RICH scalars show that F0 occurs during the middle of a good fill and the run log indicates that run 9062012 was a successful ppProduction2008/Physics. The shift log does not show any abnormal activity with the beam. It appears some kind of glitch occured making it seem as if a new fill was starting for just one run. All evidence points to run 9062012 belonging with F9975 and the fill F0 not being a true, seperate fill. Putting run 9062012 in with fill F9975 and deleating F0 in run list fixes this problem.

 

Conclusion:

After applying all the changes detailed above, I am left with this run list. This leads to a final list of good fills and a list of bad fills. We see that the steps taken above resolve every bad fill except F9849 which contained the dead RICH scalar time.