TileCalorimeter

 
The following files have been made by averaging the QC6 scan response at each tile/fiber position over a set of modules and over the odd and even PMT's.  The average responses are normalized to 100 in each cell-layer.  The idea is to reduce the random variation due to fibers and tiles and see what systematic variations may remain.  The Cs scans have been analysed with two different programs to extract the individual tile responses, one written by Larry Nodulman (LN) and the other is the CERN program, run by Bob Stanek.  The LED scans have been analysed by RJM.  The ntuples have the same structure as for the individual modules described on the Production page:
http://www.pa.msu.edu/~miller/atlas/optics/MSUProduction.html

---

Averages over all modules that have been scanned and analysed:

• Average of LED Scans of 22 modules instrumented at MSU:

Graphics File (ps)                        ntuple(txt)
• Average of Cs Scans of 33 modules using  LN's peak analysis:

 Graphics File (ps)                       ntuple(txt)
• Average of Cs Scans of 31 modules using CERN's  peak analysis:

 Graphics File (ps)                       ntuple(txt)

---

Averages of the last 10 modules that were instrumented at MSU and scanned with the LED source at MSU and with the Cs source at ANL:

• Average of LED Scans:

Graphics File (ps)                        ntuple (txt)
• Average of Cs Scans using LN's peak analysis:

Graphics File (ps)                        ntuple (txt)
• Average of Cs Scans with CERN's  peak analysis:

Graphics File (ps)                         ntuple (txt)

---

Averages over the last 14 Modules that were instrumented at ANL and analysed by both programs:

• Average of Cs Scans with LN's peak analysis:

Graphics File (ps)                         ntuple (txt)
• Average of Cs Scans with CERN's  peak analysis:

Graphics File (ps)                          ntuple (txt)

---

Observations:

The Cs scans have variations, typically between -5% to +5%,  that remain after the averaging.  The pattern of the variations is different for the two different analyses.

The LED scans have smaller variations after the averaging, typically +_2%.  The variations that remain are associated with special ITC tiles, and possibly with submodule structures (eg A16, layer 2).

The Cs scan patterns appear to be systematic and are probably the result of unfolding the Cs signal from neighboring tiles.

1. The same patterns exist in both ANL and MSU instrumented modules.
2. The variation is not random between neighboring tiles.
3. There are discontinuities across cell boundaries
4. The patterns are different for the LN and CERN methods.  The LN pattern is a peak at one edge of the cell and a dip at the other, alternating between odd-even layers.  The CERN pattern is periodic with the pattern spanning cells.
5. In many cases a region with high average response in one layer corresponds to a region in another layer in the same cell with a low average response .

It would be informative to do a similar analysis of the scans of the BCN and CERN modules.