My recent objective has been to help determine which runs and fills should be uploaded to the SpinDB. In previous blog entries I describe how I determined the bxing for all of the 2008 and 2009 runs.

There were a variety of problems that could make a run undesirable. There were a few fills where the ideal_mean was incorrectly calculated. Some runs had too few events (less than a few thousand), making the beam crossing histograms noisy/unreliable. And sometimes the beam crossing 7 counter was not what it regularly was (it is usually 80), although irregular bxings by themselves do not necessarily indicate a problem.

In an attempt to answer some questions about which runs were good/bad, I modified some code and wrote "BxingQAMaps.C". It works by iterating over two while loops. The first loop opens up the file "ListOfAllFillsAndRuns.txt" (a text file created by the perl script "getAllFillsAndRuns.pl", that lists every run number from "TestCleanSummary.txt" along with its associated fill) and stores that information into 3 data storing devices called maps. The next loop iterates over "summary.pass8a.dat" and reads in information about each run.

BxingQAMaps.C then analyzes the data from each run, determining if the run is a good run or a bad run. It does this by checking the bxing7 value, the ideal mean, and the number of events. If anything looks suspicious it is tagged as a bad run and added to the appropriate map. At the end of the print out, each fill is printed along with the number of bad runs and the number of good runs in that fill. This provides a ratio of bad/total runs per fill, small ratios mean that there might have been a few mistakes but the data is probably worth keep, large ratios indicate that the fill should be looked at/thrown out.

A histogram is then plotted (attached below, note that for the sake of simplicity, bxing vs "counter" is plotted, where the counter variable corresponds to the run number). The top line graph represents the ideal mean (it is scaled in order to fit on the plot). Notice the three places where it peaks, and how the beam crossing data becomes irregular there; there must be some correlation between the two.

Please note that the original files listed above can be found in the /star/institutions/uky/tjswed/submitdir/...  directory. But for your convenience, copies are attached below.