Camilo Carrillo
High Energy Physics 2006 - 2013
Software
Links
While learning CMSSW I spent some time finding bookmarks with information. This short collection may be useful for newcomers.
- CMS convention (Requires sign in with your CERN account)
- CMS for newcomers (public INFN Bari site)
- C++ Language Syntax Reference
- CMSSW repository
- LXR (The LXR cross reference for CMSSW)
- CMSSW and more (Giuseppe Zito Home Page)
- CMS Main Page
- RPC Main Page page (Requires sign in with CERN account)
RPC Detector Id
The RPCDetId is a class written in CMSSW. It identifies a detector unit in the RPC system. The inheritance diagram and more information here.
In the following paragraphs more information about how is constructed.
::region
The region is an index that could be –1,0 or 1.
- -1 Negative Endcap
- 0 Barrel
- 1 Positive Endcap
::ring
In the barrel the ring could be -2,1,0,1,2 referring to each wheel in the barrel system.
While in the endcap it could be 2 or 3, (in the future 1 as well that most probably will be GEMs)
The next picture shows a disk with rings 2 and 3 installed and with the space devoted for the ring 1.
::station
In the barrel the stations could be 1,2,3,4 they are shown in the yellow part of the following plot. And represent each cavity in between the iron yoke layers.
In the endcap there are just 3 stations 1,2,3 they represent the disk (for the upgrade 2013-2014 there will be a forth disk in the endcaps RE4)
::sector
There are 12 sectors in the barrel, sector 10 is show in the image below. One wheel has 12 sectors
In the endcap there are 6 sectors with no clear segmentation in the hardware of the detector it is just a set of 6 chambers in every ring, since each disks is composed by 36 chambers each sector, at the same time, is composed by 6 chambers.
::subsector
The subsector in the endcap goes from 1 to 6, it refers to one single chamber in a given sector. The combination of sector and subsectors gives the Chamber number in the endcap.
::roll
The roll identifies an extra segmentation along z in the barrel and along r in the endcap. In total there are 2316 rolls in CMS 1020 in the Barrel and 1296 en el endcap (+432 from RE4) .
The rolls have a name according to their position inside a chamber. The conventions are the following
Barrel
- Backward roll
- Middle roll
- Forward roll
EndCap
- A
- B
- C
Strips
There are 109608 strips in the RPC system (not counting RE4)
- 68136 in the Barrel
- 41472 in the Endcap
In the barrel the number of strips per roll changes according to the type of chamber. All these rolls in the endcap have 32 strips (including RE4) with RE4 we will have 144 x 32 = 4608 strips more for a total of 114216 strips (channels).
Layer
The layer is not a function of the RPCId, but can be found in the RPCGeomtryServices class.
The layer does not make sense in the endcaps so it is always set to 1. In the barrel however everywhere is 1 except in the stations 1 and 2 where there is an extra layer in the station called out for the out layers the index layer value in the detector id is 2.
In total in the barrel we have 6 layers:
- Station 1 in, RB1in
- Station 1 out, RB1out
- Station 2 in, RB2in
- Station 2 out, RB2out
- Station 3, RB3
- Station 4, RB4
Chamber
The chamber is not a function for RPCId but can be found in the RPCGeometryServices class. It is very useful in the endcap where the number runs from 1 to 36 indexing every chamber.
Studies with RPC Geometry
The first and natural step that should be done in CMSSW is to understand the geometry, unfortunately there is not much documentation about this, so here we have some plots that I did with gnuplot in order to understand what the indices means. In geometry contains all the information about the position and shape of all the detectors in CMS, you
can ask for the coordinates of the central points of a set of strips. For example next image shows all the central points in the region 0 (the barrel) and ring 1 (Wheel +1)
This one shows all the central points of those strips in the barrel in the first sector.
And finally all the central points sector 1 and sector 6
It is clear that there are 6 layers in the barrel from this very last image
RPC Efficiency
My task in this field was check the efficiency of the RPCs in the Barrel using the information from DTs, so in this picture you can find a sketch of the idea.
We got from the event the information about de DT Segments, that is basically a point and a direction, then an extrapolation is done following a straight line from the point until the RPC surface (The magnetic field is very log outside the yoke). (in MB1 and MB2 the extrapolation is done forward and backward). We get a “prediction” (the shadowed strip in the image) a histogram is filled for this strip, after the prediction we check in the same event if we have some RPC data (the yellow strip) in the same strip, if we got a coincidence, another histogram is filled for the same strip, after many iterations we got the efficiency of each strip. This method works for MB1, MB2 and MB3, in the case of MB4 we have a limitation due that the DT Segments there are 2-Dimensionals, then it is necessary to extrapolate from MB3 in order to know the extrapolated point. In this case it is necessary to keep in mind that the extrapolation should not be done in a straight line due the existence of the magnetic field.
This fact does not affect the measurement of the efficiency due to the fact we ask for a correlation between the directions of the MB3 segment and the MB4 segment to do the prediction. But this affect the statistics of the measurement, then in order to improve the statistics we introduce an assertion that let the prediction continue, if the 2 points in MB3 and the 2 points in MB4 are co-circulars (the 4 points could be contained in a circle) then we can do the extrapolation and check for the coincidence. Here you can see some preliminary results from MTCC2.
A work that should be done in the future is similar to this but this time with the CSC in the end caps.
Noise Polluter
Our first work done simulating noise for RPCs in CMS consisted in: choose randomly a set of points in CMS’ volume, as you can see in the next image.
And then those points that are contained in the parallelepipeds surrounding a strip are understand as a successful hit doing by randomly noise. At the end we can show with iguana an event example showing which strip was turned on following this criteria.
Here other image with IGUANA, showing the strips turned on by the noise.
Summary
During the second half of 2006 I started my work in software. Working in building 40.
Here you will find key information, almost random, about CMSSW specially related to Muon/RPCs.