Comparison of the LED and Cs Scans of the Tilecal EB Modules Instrumented at MSU

R. J. Miller
Last Update: May 23, 2001

The following table contains links to data files and graphical files showing the Cs and LED scans of EB modules instrumented at MSU (and ANL).  The data files are in the form of ntuples containing one row for each tile with the following 8 columns:

LAYER = the tile layer (hole) number 1:11

SLOT = the sequence number of the tile slot starting at the ITC end.  The sequence in each layer starts with an offset corresponding to the offset present in the module, so that all layers end at slot number 157 at the high eta end.

CELL = the cell number of the tile in the form dxx, where d = 0:3 for depth A:D, and xx is the eta bin index.

TILE = the tile sequence number within the cell.

CS = the difference between the response of this tile and the layer average in the final Cs scan at ANL. I start with the numbers given in Larry Nodulman's ntuples and renormalized those data using the following procedure.

Convert all peak heights to %'s of the cell average by dividing the original numbers (normalized to 300) by 3.

Average the numbers for the odd and even PMT's to enhance the statistical weight of the tile dependent part of the response.

Make the responses continuous across cell boundaries by constraining the two tiles straddling a cell boundary to have equal responses.  This additive correction is applied to all of the tiles in the cell.  The rationale is that the variation between neighboring tiles in the same tile packs should be small and there is no correlation between cell boundaries and tile pack boundaries.

Renormalize by subtracting the average response in that layer from all tiles.

LED = the difference between the response of this tile and the layer average in the final LED scan at MSU. I apply the same procedure as above, starting with the ntuples linked to the MSU production summary page.

PEAKDX = the distance (mm) between this tile and the previous tile, where the tile position is defined to be the maximum of a quadratic function fitted to the 7  highest points in the LED scan.  The source motion between the points is approximately 0.5 mm.  This can be used to look for correlations with submodule gaps, ITC end plate, etc.  Probably accurate at the 0.1 mm level.

PEAKW = the width (mm) of the peak in the LED scan, defined to be the positive range of the fitted quadratic function.  The fitted points have the pedestal subtracted, where the pedestal is defined to be the average of the 3+3  points at the minima on either side of the peak.  (I was hoping to measure something related to the slot width, but this probably give something related to the tile width instead.)

The graphical files show the distributions of (CS,  LED, and  CS - LED) vs SLOT for each layer.  The difference plot should remove the contribution of fiber variations and, to some extent the variation in the tile transparency.  The solid vertical lines indicate the cell boundaries and the dashed vertical lines indicate the submodules.
 
 

* PMT problem, ignore the C10 cell in the LED scan

** PMT problem, ignore the D4 cell in the LED scan

The following files are for modules that were instrumented at ANL and have not been scanned with the LED source.  Otherwise, the analysis and formats are the same as for the MSU instrumented modules.

The average module shows the average of the Cs scans for each tile slot for each layer for all modules up to ANL027, that have been instrumented since we started sorting the tile packs.
 

* Some bins overflow the plotting limits in layer 2
** Some bins overflow the plotting limits in layer 8