The SLAC KPiX Chip for ILC K p f GEM-Digital Hadron C l Calorimetry i t A d Andy White Whit for the GEM-DHCAL Group for the GEM-DHCAL Group (UTA – UW – CNU) ( ) With many thanks to SLAC colleagues: M. Breidenbach, G. Haller, D. Freytag, R. Herbst RD51 P RD51, Paris, October 2008 O b 2008
GEM/DHCAL active layer concept -2100V ∆ V ~400V ∆ V ~400V ∆ V ~400V 0V 0V
KPiX chip KPiX/GEM/DHCAL One channel of 1024 Dynamic gain Dynamic gain select 13 bit DHCAL A/D A/D anode pad Storage until end of train. end of train. Pipeline depth Leakage presently is 4 current subtraction Event trigger calibration
GEM-DHCAL/KPiX boards with Interface and FPGA boards
Redesigned chamber – all fishing line spacer - KPiX anode board with extra electronics protection - Better/more direct gas flow through ionization gap / d fl h h - No large dielectric spacer(s) – previously killed g p p y signal
GEM chamber with KPiX v4 – early 2008
GEM + KPiX long source run at SLAC nnel # fC f Cha Channel # Channel #
GEM + KPiX response in lab at SLAC Right under the source Right under the source Away from the source Away from the source Away from the source Right under the source
First results from new GEM chamber + KPiX fC scale fC scale
UTA 75hr Long run at SLAC
GEM/DHCAL GEM/DHCAL with KPiX readout ith KPiX d t - Due to the synchronous design of KPiX, data taking efficiency was low with (asynchronous) source. y ( y ) - KPiX v7 offers new timing flexibility. - Waiting for v7 -> study pedestal subtraction to W iti f 7 t d d t l bt ti t (eventually) extract MIP distribution - Also study channel variation/stability of calibration and pedestal feedback to SLAC KPiX developers.
Calibration studies with KPiX v4 Goa s Goals: - understand relation between pedestal distribution and “zero-charge” injected distribution zero charge injected distribution. - investigate stability of gain, response to injected charge pedestals > subtraction of noise to yield MIP charge, pedestals -> subtraction of noise to yield MIP signals. - understand factors that influence gain etc. (channel d st d f t s th t i fl i t ( h l to channel, single channel/environment, long-term fluctuations fluctuations… - ultimately – develop calibration procedure (how often, length of calibration/pedestal runs length of calibration/pedestal runs,…) )
KPiX injected charge calibration K j g Internal capacitor charged via DAC, readout through d t data path th -> measure gain from slope -> measure “zero-injected charge” response, “Y-int”
Initial Calibration Analysis Example Initial Calibration Analysis Example • Two Calib Data Sets – 19 hr run starting in evening – weekend – g g 6/28 – 24 hr run starting in morning - work week – 24 hr run starting in morning work week 7/01 • Each Channel’s Pedestal Mean Sigma; Gain • Each Channel s Pedestal Mean, Sigma; Gain, and Y-Intercept verses Run# (on order of an hour) is graphed for each channel and fit with a hour) is graphed for each channel and fit with a straight line. Work of UTA Master’s student Jacob Smith
One channel KPiX v4 Ped Gain Y-int Sig Normal Double 17 16 19 18
Normal Ped Gain Y-int Sig Double Weekend run 19hr
Normal Ped Gain Y-int Sig Double Weekday run 24hr
Normal Ped Gain Y-int Sig Double April - June
GEM/KPiX source data taking - Issue with KPiX v4 triggering mode: gg g - “forced trigger” (software) mode was used – no fixed time relation between arrival of electron from source t m r at on tw n arr a of ctron from sourc and internal timing of KPiX. - we suspect that a reset is responsible for incomplete we suspect that a reset is responsible for incomplete integration of the charge - this would distort MIP (Landau) distribution by this would distort MIP (Landau) distribution by lowering ADC values - also noise peak wider with data than for pedestal runs data than for pedestal runs -> working on understanding this effect. this effect.
Next step: KPiX v7 KPiX is a synchronous device – tied to ILC beam timing -> problem being efficient in an asynchronous -> problem being efficient in an asynchronous environment e.g. MTBF , cosmics, or source(s). -> KPiX v4 was only ≤ 5% efficient for source or > KPiX v4 was only ≤ 5% efficient for source or asynchronous beam. -> KPiX v7 allows more flexibility in timing adjustments – KPiX 7 ll fl ibilit i ti i dj t t via slowing down the master clock - should provide data taking efficiencies in 40-50% range in an asynchronous taking efficiencies in 40-50% range in an asynchronous environment, and has new reset scheme. - First version of v7 now at UTA, plus we have a new First version of v7 now at UTA plus we have a new anode board designed to work with it. - Initial tests -> calibration with injected charge I iti l t t lib ti ith i j t d h (internal) -> results.
New KPiX Board for v7 New KPiX Board for v7 1. The KPiX chip is located just next to the PAD area. 2. 310 x 310 mm. with 8 x 8cm. pad area at the center p with 0.2mm. Gaps between pads. 3. Surface mount parts (connectors...). Nothing will be on the PAD side. n th PAD sid 4. All the holes filled for gas tightness. 5 5. Leaving about 1 space on the edges for sealing.... Leaving about 1" space on the edges for sealing 6. KPiX board thicker, flatter…
KPiX v7 board layout
KPiX v7 – first calibration results
KPiX v7 chamber plans KPiX v7 chamber plans - New v7 chamber works well – stable (no trips) over several weeks so far. - Verify new software for v7 - Operate v7 with new GEM chamber (in progress) Operate v7 with new GEM chamber (in progress) - Complete calibration/understand behavior with v7 - Work with SLAC to achieve 40-50% efficiency in cosmic and source data taking. - Work at UTA and SLAC. - Take beam data at MTBF CERN? Take beam data at MTBF, CERN?
Future Plans for GEM/DHCAL - Complete GEM chamber studies with v7 - Construct ~1m x 0 3m chamber - Construct ~1m x 0.3m chamber - Construct ~ 1m 2 plane(s). - Issue with foil supplier: 3M closed plant in Missouri - Have obtained quotes from CERN for 99cm x 32cm Have obtained quotes from CERN for 99cm x 32cm foils. - Alternative: Thick-GEM chamber – if large areas are - Alternative: Thick-GEM chamber if large areas are viable - LCDRD Supplement, 2 nd year – minimal amount to LCDRD Supplement 2 nd year minimal amount to purchase foils + limited postdoc support. - Attention to 1m 2 plane design – walls/thickness, gas Att ti t 1 l d i ll /thi k 2 supply, KPiX readout… >>>
Future Plans for GEM/DHCAL Future Plans for GEM/DHCAL - Next generation KPiX (256 1024 channels) Next generation KPiX (256…1024 channels) - Note: started out to test GEM with both KPiX and DCAL chips DCAL chips – we would like to continue this…also we would like to continue this also perhaps use alternative readout from Europe - …We will have a visitor joining our group in November – W ill h i it j i i i N b much experience with detectors/readout - We also will be hiring a postdoc 50% ILC R&D, 50% % % ATLAS
Recommend
More recommend
