COMPTON POLARIMETRY Collected data Cavity power Status on

COMPTON POLARIMETRY • Collected data • Cavity power • Status on counting methods • Systematic errors and hardware issues

Collected Data Good Runs Ebeam (Ge. V) k’Max (Me. V) <A>int (%) AMax (%) 4 He ~450 2. 76 129 2. 3 4. 8 LH 2 ~450 3. 18 171 2. 7 5. 5

Cavity Power • Laser power supply changed in June, tested OK: 1200 W • Down to 300 W few days before the experiment • Incident laser power still nominal • Laser beam waist measured in agreement with simulation • Alignment optimized • Back to 500 W after optics cleaning • Slow exponential decay of cavity power seen during part of the run • Cavity mirrors? Two finesse measurements so far, differ by a factor 2 to be done again…

Counting Analysis Differential 650 mm strips 4. 5 mm gap Pb. W 04 Semi-integrated

Calibration e- Detector Ee- Position of the Compton edge calibrates the detector and determine the vertical gap Ydet. DPe/Pe ~ 2 DY/Y Can use the rate or the asym spectra. Discrepancy taken as syst error. DY=200 mm DPe/Pe = 3. 5% @ 3 Ge. V Ee- a Bdl, lever arm after D 3 (~4. 1 m)

Calibration e- Detector Comtpon+Background S/B Compton edge position sensitive to beam motion (coil pulsing) Extracted Ydet is always underestimated

Calibration e- Detector • Fit of the asymmetry spectra of the 4 planes is more robust because it uses information from all strips. • Systematic error could be pinned down by simulation • Stringent cross-check available on line…

Calibration cross-check • The response function of the g detector is determined using the e- det as an energy tagger reference run. • When fitting the photon spectrum, l corrects for gain drift and error in the e- det calibration. expect l=1 for the fit of the reference run itself l-1 = 10 -4 !! Had to correct for the fact that mstrip planes are 1 cm appart and use survey data for the lever arm after dipole 3 (4. 11 m instead of 4. 10)

Photon Analysis No dependence on the software threshold over a wide range Systematics of the response function under control.

Comparison e- - g 03/08/05 19/09/05

Comparison e- - g 15/10/05 01/11/05

Laser Polarization

Laser Polarization

Laser Polarization e- and g data are not compatible origin found to be electronic pickup: ADC pedestal changes by 4 channels depending on the cavity state! Wasn’t there last year… Could be due to logical signals of cavity ON/OFF, Left/Right present in several electronic crates? Easy to test.

Laser Polarization After pedestal correction, all analysis are compatible at 1 s level. Compton polarimetry is a tedious way to check pedestal correlations… The correction doesn’t affect the mean value, but reduces the Left Right discrepancy. Left laser polarization is ~0. 6% lower than Right polarization?

Laser Polarization La ser CIP Tran sfer Ex it Monitoring of the laser polarization downstream the cavity Li ne l/4 plates Fun ction

l/4 Scans Integrating Sphere S 1 TM pol. state 8. 86 9. 86 TE pol. state x Slow Wollaston Prism b x y Rotatable l/4 plate Incoming Polarization Ellipse Integrating Sphere S 2 y Stokes Parameters: P 0, P 1, P 2, P 3

Laser Polarization • 30 l/4 scans performed during helium run. • We do find a 0. 5% difference in laser polarization states but with opposite sign Need to check the Left-Right definitions in both analysis… Accumulated cavity power: Left state = Right state at few % level (~500 W)

Conclusion ØSignificant improvements in the counting analysis Best agreement between the two methods ever achieved On track for 2% accuracy or better. ØHardware issues: compton is getting old… • Should revise all cabling at some point • Need cross-talk study before GEn starts • Numerous « end of run failed » remove 600 Hz part of the aquisition? • Cavity power is low? How long before very low?

Systematic Errors Counting photon Source Last Year To Do Expect. Pg 0. 70% Open cavity? 0. 7%-1% Resp. Func. 1. 25% Ana. +Simul. 1. 1% Dead Time 1. 00% - ~0% Pile up 1. 00% Simulation 0. 5% Rad. Corr. 0. 25% - ~0% TOTAL 2. 0% 1. 4%-1. 6% Counting electron Expect to reduce syst. error as well because of better calibration. (3. 5% for now)