1 / 29

HBD Software Status Update

HBD Software Status Update. By Sky Rolnick UC Riverside. Outline. HBD Local Coordinate Issues. Status on Gain Calibrations. HbdEventDisplay Hub n’ Spoke Clustering Algorithm HbdAnalyzer EventAccumulator. HBD Local Coordinates. Current Issues.

brownjanice
Download Presentation

HBD Software Status Update

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. HBD Software Status Update By Sky Rolnick UC Riverside Sky Rolnick, UC Riverside

  2. Outline • HBD Local Coordinate Issues. • Status on Gain Calibrations. • HbdEventDisplay • Hub n’ Spoke Clustering Algorithm • HbdAnalyzer • EventAccumulator Sky Rolnick, UC Riverside

  3. HBD Local Coordinates Sky Rolnick, UC Riverside

  4. Current Issues • Multiple coordinates defined in code which are inconsistent and incompatible. • Almost no documentation exists to define coordinate system used. • Need to discuss weather we should keep things “as is”, or if its worth cleaning this up a bit. Sky Rolnick, UC Riverside

  5. HbdFinalSimSupport • Defined coordinates are Left Handed!!! • There are 24 individual local coordinates defined. • Many of these coordinates are related through a parallel transport. • Transformations require unnecessary shift depending on module. • Coordinates are different than what is defined on PCB coordinates. Sky Rolnick, UC Riverside

  6. HbdFinalSimSupport Sky Rolnick, UC Riverside

  7. HbdFinalSimSupport Technically coordinates are consistent due to two cancelling effects. If one goes from glob2loc and then loc2glob everything is okay since shifts cancel out. But if someone uses PCB coordinates and uses loc2glob then coordinates are shifted. Result of not using standard PCB coordinates in definition of local coordinate system. Sky Rolnick, UC Riverside

  8. Define “new” Local Coordinates Sky Rolnick, UC Riverside

  9. New Coordinates much Simpler New coordinates much simpler to use. Will be completely compatible with all code that goes from glob2loc and loc2glob and also compatible with PCB coordinates. Will require minor modifications to FindPad, and DistanceFromPoint in HbdFinalSimSupport. Seems to work fine for me, but should be tested by others. Sky Rolnick, UC Riverside

  10. hbdDetectorGeo • Uses database to fetch local coordinates. • Has advantage that module mapping does not depend on cable connections. • Would like to add row, col coordinates for pad, sector, as well as padkey to database. hbdDetectorGeo hbdgeo; hbdgeo.fetch(215034); hbdgeo.LocToGlob(PosY, PosZ,x_glob,y_glob,z_glob,sector); Sky Rolnick, UC Riverside

  11. What to do??? • Remove one of these coordinate systems to eliminate confusion or keep them both for backwards compatibility? • Redefine coordinates to be consistent with PCB coordinates. • Include (row, col) functionality to allow clustering across boundaries. • I have worked out my own mapping that eliminates “boundary effects” but I could use help in how to add multimap to data base. Sky Rolnick, UC Riverside

  12. HbdEventDisplay cvs co offline/analysis/HbdEventDisplay Sky Rolnick, UC Riverside

  13. HbdEventDisplay This is original event display as created by Weizmann Institute but packaged in Fun4All framework allowing display of multiple data types including DST & PRDF for looking at raw data or sim data. run275586 pp 500 GeV in (+-) field configuration Sky Rolnick, UC Riverside

  14. HBD Gain Calibrations Sky Rolnick, UC Riverside

  15. HBD Gain Calibrations Without calibrations we see very high charge >120 fC. After calibrations this settles down by ~factor of 5 depending on module, padnum, P/T, and gain settings. Sky Rolnick, UC Riverside

  16. HBD Gain Calibrations Gains vary slightly among modules depending on characteristics of GEM. Also pad by pad variations due primarily to irregularities of GEM holes. We can compare Module by Module gain calibrations to Pad by Pad gain calibrations. Typically, analysis will use both calibrations to achieve greatest uniformity. Are we satisfied with what is in the database, or do these need to be updated further? Sky Rolnick, UC Riverside

  17. Hub n’ Spoke Clustering Algorithm cvs co offline/packages/hbd/HbdHnSClusterizer.C cvs co offline/packages/hbd/HbdHnSClusterizer.h Sky Rolnick, UC Riverside

  18. Hub n’ Spoke Clustering • Apologies for taking so long… • Complicated pad mapping. • Issues with coordinate system. • Lots of memory leaks and other bugs. • Multiple versions not always updated. • Differences between simulated and real data. • My poor programming skills. cvs co offline/packages/hbd/HbdHnSClusterizer.C cvs co offline/packages/hbd/HbdHnSClusterizer.h Sky Rolnick, UC Riverside

  19. How it works • Uses track projections to locate hub, where hub is defined as three-tuple consisting of three closest pads to track. • Searches for spoke defined as the three-tuple surrounding the hub with the largest charge. • Currently search region includes a spoke radius of 2, which means at most spoke can be located a distance of two hub centers away. Sky Rolnick, UC Riverside

  20. Hub n’ Spoke Clustering Sky Rolnick, UC Riverside

  21. Hub n’ Spoke Clustering Sky Rolnick, UC Riverside

  22. Hub n’ Spoke Clustering Sky Rolnick, UC Riverside

  23. Limitations • As you can see, very often electrons are detected which have little or no HBD response. These are likely to be conversion pairs on outside of HBD. • Many of the HnS cluster shapes do not agree well with actual cluster shape and are therefore pick up extra scintillation. Is there a way around this? • It is very difficult to distinguish between “single”, “double”, and “hadron” just by looking at distributions. • Higher multiplicity events will pick up lots of scintillation light, skewing results. Unavoidable??? Sky Rolnick, UC Riverside

  24. HbdAnalyzer cvs co offline/analysis/HbdAnalyzer Sky Rolnick, UC Riverside

  25. HbdAnalyzer • Allows one to apply hub n’ spoke parameters in analysis. • HbdHnSClusterizer must be added before HbdAnalyzer otherwise track variables will not be filled. • Implemented new version of PHCentralTrack, up to v24 now. Sky Rolnick, UC Riverside

  26. EventAccumulator cvs co offline/analysis/EvntAccumulator/ Sky Rolnick, UC Riverside

  27. EventAccumulator • Uses some of the fancy stuff in Fun4All to accumulate tracks. • Has the ability to accumulate tracks over multiple events as way of stacking events. • Currently using a buffer to store tracks but this is the dumb way, the much smarter way will be just to use PHCentralTrack as the container. Accumulates over previous 5 events. Starting new buffer of Accumulated events Ncalls = 96 There are 0 tracks in PHCentralTrack Total number of accumulated tracks is 0 Ncalls = 97 There are 2 tracks in PHCentralTrack Total number of accumulated tracks is 2 Ncalls = 98 There are 2 tracks in PHCentralTrack Total number of accumulated tracks is 4 Ncalls = 99 There are 2 tracks in PHCentralTrack Total number of accumulated tracks is 6 Ncalls = 100 There are 3 tracks in PHCentralTrack Total number of accumulated tracks is 9 ************************************ There are 9 tracks in MEGACentralTrack Sky Rolnick, UC Riverside

  28. EventAccumulator • Does it make sense only to accumulate events that have a similar vertex? • There could be some funny side effects of tracking with events from multiple vertex points. • Maybe I should create vertex pools similar to what is done in Cabanaboy? Sky Rolnick, UC Riverside

  29. The End Sky Rolnick, UC Riverside

More Related