1. Dataset

• 2005 p+p HighTower-1 data
• Single-particle Monte Carlo simulation
• Default cluster finder
• EMC points used in this study are what we call "photon candidates" in the analysis, passed the TPC veto cut

2. SMD energy data/simulation

2.1 SMDE and SMDP in the data

2.2 SMDE and SMDP in the simulation

2.3 SMDE and SMDP in data vs. simulation

3. SMD energy scale as a function of pseudorapidity

3.1 SMDE and SMDP data/simulation per eta bin

3.2 Constant fit per eta bin

4. SMD cluster widths in data and simulation

5. Conclusions

Based on this study (figures 2.3, between 1 and 15 GeV), I assign a 15% systematic uncertainty on the BSMD energy calibration.

Being an in situ measurement, this estimate already includes the uncertainty due to the atmospheric pressure changes.

Steve's documentation on the pressure corrections to the SMD energies is here (reports 3% gain variation in ppLong 2005). My study of pressure variations is here (report 5–6% in 2006–2009).

In this study I do not observe the strong eta dependent divergence between BSMDE and BSMDP gains reported by Jan here (observes up to 27% difference between the two planes at eta=1 in the 2008 data). Differences between this analysis and Jan's are: (a) different dataset; (b) different cluster finder; (c) Jan operates on the raw energy deposition in the BSMD layers while the database tables have that corrected with the sampling fraction to get the full particle energy, these sampling fractions may correct the eta dependence.