Active Dividers

1. Active Dividers Tilecal upgrade meeting at Stockholm (3-5 June 2013) François Vazeille on behalf of Roméo Bonnefoy, Michel Crouau (Now retired) Dominique Pallin, Christian Fayard, Marie-Lise Mercier, Eric Sahuc. •  Summary of previous progresses • Main results given in the Tilecal Note • (Note not yet approved) • Missing electronics studies to do • Radiation hardness and next tests • First production for ITC scintillators • Conclusion 1

2. Summary of previous progresses • Goal: To keep the Tilecal performances (Linearity, energy resolution) for the • measurement of very high energies at the highest luminosities. • Proposed solution: Replacement of present “Passive Dividers” by • “Active Dividers” keeping stable the voltage repartition in between every stage •  Use of active components (Transistors and diodes) in the last 3 stages, • in order to balance the current stream induced by Minimum Bias events. • Preliminary results previouslyshown [Tilecal meeting, 17 June 2011]. Passive Dividers Active Dividers 2

3. Main results given in the Tilecal Note (The Note ATL-COM-TILECAL-2013-014 is not yet approved)  Two series of tests Passive Active Test Bench of batch of 10 Dividers • Pure electronics tests • using the upgraded Divider Test Bench •  Interdynode voltage comparisons • without/with injected currents of • 10 Passive and 20 Active Dividers. •  Protocol for the certifications of productions. • Full tests of Dividers associated to PMTs •  Comparisons for a given AC light (Signal) • of 20 sets ″PMT-Passive Divider-Active Divider″ • in function of the injected DC light (Minimum Bias). No other BE electronics than a scope 3

4.  Summary of specifications 4

5.  Summary of results for currents up 120 µA PMT#AB2484, Divider#108508  Examples of comparison PMT-Passive Divider for 2 current ranges • NL can reach 40%. • NL < 1% at 2 µA PMT-Active Divider for 2 current ranges - Only 5 sets show a small NL tendency. - 15 other sets have a NL  0%. PMT#9F23Q1 Divider#2 PMT#AA0061 Divider#1 5

6.  General results from the 20 sets  Passive Dividers : Results and comparison s of the means to simulations • Passive Dividers fit the ATLAS conditions: NL < 1% as requested. • Very good agreement with simulations. Active Dividers • The behavior differences are likely due to the PMT behaviors. • The Active Dividers fit all ATLAS (Crack  8µA) • and sATLAS (Tilecal 10 µA and Crack 40 µA) specifications, • with an improved PMT working and a large safety everywhere. • - Simulations at 20 µA: NL < 0.2% and measures: <NL> < 0.2%. 6

7.  Gain dispersion 20 Active Dividers coupled to the same PMT at the same HV  Gain dispersion < 0.7%: a very good working homogeneity.  Comparison of absolute gains without DC currents Active Dividers give a gain 22% higher than Passive Dividers. Precisely: (21.9  1.8) % 7

8. Missing electronics studies to do and completion of the Divider Note As proposed by comments on the Divider Note: Noise studies Comparisons of Passive and Active Dividers on the PMT working  After this Stockholm meeting. 8

9. Radiation hardness and next tests  Full application of the ATLAS radiation policy • Most pessimistic estimates (LB) • while waiting for a new ATLAS policy. TID (Gammas): Gray NIEL (neutrons): 1 MeV eq.n/cm2 SEE (protons): >20 MeV h/cm2 • Requested tests of 20 Dividers following the ATLAS protocol, • likely TID and NIEL only.  First indications • LHCb tests on 4 Dividers: no failure up to the maximum tested dose of 870 Gy • and the whole LHCb preshower still working. • - Argonne tests (See Gary’s results) on 4 Dividers: still working. 9

10.  Clermont-Ferrand set-up  Special PCB to irradiate: active components within a square 7x7 cm2. • 20 Dividers: whole last 3 stages + resistors simulating the other ones. • 16 Transistors alone + 8 Diodes alone (with direct/inverted polarities). • Everything supplied at the right HV and on-line monitored. Resistors Diodes Last 3 stages Transistors PCB to be irradiated Power Supply  Complete Test bench with on-line monitoring PCB • Hardware ready. • Software almost ready + working tests to do. • Needs only the connection to a laptop • under labview. USB interface Mux/DAQ 10

11.  Interest of having a whole on-line monitoring • Testing up to the maximum doses with a permanent monitoring • enables to know the Divider status at intermediate doses: • from ATLAS to SATLAS, • for Tilecal and ITC channels, • for softer new ATLAS radiation rules (Yesterday Liq Ar talk).  Experimental sites for tests ? • AIDA Transnational Access  support for tests in 5 European sites • (DESY, CERN, JSI, KIT, UCL)  First considered option: X-generator at Clermont-Ferrand (250 keV max) long investigations and discussion with CERN experts (Federico Faccio and Philippe Farthouat)  Big calibration difficulties in the dose estimates.  Preferred option: Gamma rays  smaller mis-estimates of the doses. But, are there possibilities in other Tilecal Institutes ? 11

12. First production for ITC scintillators • Batch of 350 Dividers soon delivered at CERN •  Clermont-Ferrand orders to selected Cies (PCB, component placements, cables) • through CERN. •  Certification will be made at Clermont-Ferrand. • Set of 10 dividers takes about Half an hour • for testing •  A quality sheet. • About 1 week for the whole production. 12

13. Conclusion  A very good agreement of electronic simulations and results, for both Passive and Active Dividers. • The Passive Dividers fit the ATLAS specifications for Tilecal PMTs, • but not fully for Crack scintillators. • The Active Dividers fit the sATLAS specifications • for both Tilecal and Crack scintillators, • within a wide range going well above the expected Luminosities. •  The linearity (and the energy resolution) will be improved /ATLAS. •  They could equip all the PMTs. • The Divider Test bench has been upgraded • and can welcome Passive/Active Dividers.  A batch of 350 Active Dividers will equip now the Crack scintillators. 13

14. Remaining studies to do • - Noise studies. • - Completion of the Divider Note. • - Radiation tests, • despite promising results (LHCb, Tilecal Argonne) •  Complete radiation tests are scheduled •  Clermont-Ferrand radiation test bench soon operational. •  Ready to provide it to other Institutes for dedicated radiation tests • with the on-line monitoring of the Divider working. 14