Geometry information for use in codes

---

---

Important!

The information on this page can be very helpful, but if you are running root (not necessarily root4star), please use the StEpdGeom class.  It has all the geometrical information you will ever need.
It is part of the StEpdUtil package (on RCAF, 'cvs co StRoot/StEpdUtil'), which also includes StBbcGeom (the geometry for the BBC, obviously), and StEpdEpFinder, which finds the Event Planes for you.

---

---

One often wants the phi and eta for a given tile. 

In the attached file TestGeometry.txt there are two Double_t arrays that you can cut and paste into your code.  It is a root macro. (drupal does not allow .C extension, so it is uploaded as .txt)

Double_t EpdTilePhi[2][12][31]

• first index is EW=0/1  - 0 means East wheel and 1 means West wheel
• second index is PP (position number)
• third index is TT (tile number)

Double_t EpdTileEta[2][31]

• first index is EW
• second index is PP

The macro also produces some pictures so you can check the validity of the geometry.  Pictures are attached below.  When you are looking to the west, be careful because -x points to the right!

---

The numbering of the supersector positions (PP) are such that it follows a clock face, when seen by particles flying into the detector.  I.e. when viewed from (0,0,0) in the TPC.  If you look at an EPD wheel from the outside, the numbering goes "counter-clockwise."  See description in the StEpd software documentation for more discussion of this.

Keep in mind the official STAR Coordinate system:

• +x points South
• +y points up (thank goodness)
• +z points West

Some very useful sketches of STAR detectors with coordinate system may be found here.  Probably the most useful image is reproduced below:

---

EpdTileCenterRadii_v4.ods is the spreadsheet which was used to calculate the geometric center of the various tiles. This center location is not trivial because the supersector design makes all but the first of the tiles asymmetric, where one side is higher than the other. An image is embedded in the spreadsheet where the basic trig calculations are done. The goal was to get these values for the psuedorapidity, so, for simplicity, the supersector was assumed to be centered on the y coordinate, which is true for no actual supersectors. Thus, if one wants the proper phi coordinate it is necessary to rotate the coordinates by (15 degrees)/2 for the first sector of the first supersector.

Additionally this spreadsheet was used for calculating the eta of the edges of each tile, as will be in the EPD NIM paper. A version of such a table was in the EPD construction proposal, but there appears to be some mistake in the calculation.

The scintillator of the EPD is assumed to start 4.6cm from the beamline, as designed. For reasons unclear to me the construction proposal has 4.5cm for this distance, but the difference is negligible. In reality there is a 1.65mm-thick epoxy gap between the tiles. This is not part of the calculation. One tile is assumed to end along the center of the gap and the next tile begins along the same line.

Hardware mapping for Run 2017

Hardware map for run 2017

It is important to maintain a record of the correspondence between (physical) supersector number (SS), position on the detector (PP) in any given run.

Also needed is the correspondence between tile number (TT) on a supersector and the QT board associated with it.

Prashanth Shanmuganathan (sprashan@kent.edu) maintains this information for the 1/8 Installation in the 2017 run.  (updated 5 Dec 2017 by Mike to show fiber bundle # also)

position on the East side (PP)	physical SS #	Fiber bundle #
4 o'clock	3	3
5 o'clock	4	4
6 o'clock	2	2

FEE cards and associated tiles

FEE #	tiles	one-wire ID
1	PP 4 odd-numbered tiles	0xFF00000007F6043A
2	PP 4 even-numbered tiles	0xFE0000003263413A
3	PP 5 odd-numbered tiles	0x0800000032C3333A
4	PP 5 even-numbered tiles	0x1100000032C0493A
5	PP 6 odd-numbered tiles	0x4800000032C04A3A
6	PP 6 even-numbered tiles	0xA700000032C3423A

Tiles in supersector at 4 o'clock

Tile	FEE# = Rx#	FEE channel
1	1	0
3	1	1
5	1	2
7	1	3
9	1	4
11	1	5
13	1	6
15	1	7
17	1	8
19	1	9
21	1	10
23	1	11
25	1	12
27	1	13
29	1	14
31	1	15
2	2	1
4	2	2
6	2	3
8	2	4
10	2	5
12	2	6
14	2	7
16	2	8
18	2	9
20	2	10
22	2	11
24	2	12
26	2	13
28	2	14
30	2	15

Tiles in supersector at 5 o'clock

Tile	FEE# = Rx#	FEE channel
1	3	0
3	3	1
5	3	2
7	3	3
9	3	4
11	3	5
13	3	6
15	3	7
17	3	8
19	3	9
21	3	10
23	3	11
25	3	12
27	3	13
29	3	14
31	3	15
2	4	1
4	4	2
6	4	3
8	4	4
10	4	5
12	4	6
14	4	7
16	4	8
18	4	9
20	4	10
22	4	11
24	4	12
26	4	13
28	4	14
30	4	15

Tiles in supersector at 6 o'clock

Tile	FEE# = Rx#	FEE channel
1	5	0
3	5	1
5	5	2
7	5	3
9	5	4
11	5	5
13	5	6
15	5	7
17	5	8
19	5	9
21	5	10
23	5	11
25	5	12
27	5	13
29	5	14
31	5	15
2	6	1
4	6	2
6	6	3
8	6	4
10	6	5
12	6	6
14	6	7
16	6	8
18	6	9
20	6	10
22	6	11
24	6	12
26	6	13
28	6	14
30	6	15

Please see attached mapping for more details.
Three Spreadsheets are included:

• Mapping Tiles
• Mapping FEEs
• Definitions

Text verision of the mapping is also attached, and epdMapReader.txt is the macro to read the map. Please change the extention to '.C', Drupal doesn't allow to upload .C extention files.


Input to 0x16 (QTC) and 0x18(QTB) are switched during access on 04/05/2017. New Mapping is attached (EPD_mapping_04052017.xlsx).

Mapping as of May 25 2017

There has been frequent mapping and swapping etc.  Things can get confusing.  This is the mapping as of 25 May 2017, and we (tentatively) plan to keep it this way at least through the Au+Au 53 GeV running.  (Edit: nope, we remapped one more time.  See drupal.star.bnl.gov/STAR/subsys/epd/mapping/mapping-53-gev-auau-run-30-may-2017)

• FEE cards:
  • Five FEE cards are the original versions.  One-wire codes 0xA700000032C3423A, 0x4800000032C04A3A, 0x0800000032C3333A, 0x1100000032C0493A, 0xFE00000032C3413A.
  • And we have one new FEE card (one-wire code 0xE700000032C03B3A) with 2.5x higher gain.  This will be the version we use going forward.
• We have four QT boards
  • Two are QT32B boards with no TAC (hence up to 32 tiles served, with ADC only).  These are addresses 0x10, 0x12.
  • One is a QT32B board with TAC (hence up to 16 tiles served, with ADC and TAC).  This is address 0x18.
  • One is a QT32C board with TAC (hence up to 16 tiles served, with ADC and TAC).  This is address 0x16.  Here, things get even a little more complicated, as regards the TACs
    • Eight of the TAC channels (4, 5, 6, 7, 28, 29, 30, 31, numbering starting at 0) are the original versions, which showed problems (spikes, etc) with the low-gain FEEs
    • Four of the TAC channels have improvements implemented by Steve in April/May.  These are channels 12, 15, 20, 23, numbering starting at 0.
    • Four of the TAC channels are disabled (for technical reasons related to fixing the channels described above).  These are 13, 14, 21, 22.

And, just for kicks, an additional complication is the signal from PP4TT15 is being split between a QT32B[with TAC] (0x18 ADC/TAC=8/12) and a QT32 C (0x16 ADC/TAC=19/23).

Steve Valentino made a very nice "cheat sheet" showing QTs 0x18 and 0x16, shown here:

I have made two maps showing these maps from the "detector point of view."  The first one shows simply the "Universal ID" which is 100*PP+TT (both numbering starting at 1), and the second one shows also the QT address and QT channels (starting from zero) of the ADC and TAC.  A TAC address of -1000 means there is no TAC.  They are below.  If you need to use these maps, you may want to download the file (see bottom of page) in pdf or png, and blow them up.

  

Mapping for the 53 GeV Au+Au run - 30 May 2017

For the final push of the 2017 run, we measure Au+Au with a mapping slightly different than the 25 May mapping.  Here it is:

PSSST!  Wanna know how this nice graphic was generated?  The macro is attached at the bottom of this page, or you can just click here