Parameters Number of lines HansGnther Moser MPI fr

Parameters Number of lines: Hans-Günther Moser MPI für Physik DEPFET Valencia Sept 29 -Oct. 1 1 readout time/line > 100 ns use switcher efficiently stay with 32 channels/switcher easy PLL clock synthesis active length (Ladder) gap at centre half length parallel switcher channels (chip) number of switchers (module) pitch in z 1 90 1 44. 5 4 32 7 49. 67 2 90 1 44. 5 4 30 7 52. 98 3 90 1 44. 5 4 28 7 56. 76 4 90 1 44. 5 4 26 7 61. 13 5 90 1 44. 5 4 36 6 51. 50 6 90 1 44. 5 4 34 6 54. 53 7 90 1 44. 5 4 32 6 57. 94 8 90 1 44. 5 4 30 6 61. 81 switcher channels (module) lines/module time/line PPL : denominator PLL: nominator 224 896 89. 29 40 7 210 840 95. 24 128 21 196 784 102. 04 320 49 182 728 109. 89 640 91 216 864 92. 59 160 27 204 816 98. 04 320 51 192 768 104. 17 20 3 180 720 111. 11 64 9 168 672 119. 05 ns 160 21 closest round PPL : denominator PLL: nominator lines/module time/line pitch/Z unused channels 224 40 7 896 89. 29 49. 67 0 200 32 5 800 100. 00 55. 63 10 192 20 3 768 104. 17 57. 94 4 160 8 1 640 125. 00 69. 53 22 200 32 5 800 100. 00 55. 63 16 200 32 5 800 100. 00 55. 63 4 192 20 3 768 104. 17 57. 94 0 160 8 1 640 125. 00 69. 53 20 160 8 1 640 125. 00 ns 69. 53 µm 8 768 Lines 104 ns 58 µm pitch 9 90 mm 1 mm 44. 5 mm 4 mm 28 6 66. 22 µm

Injection Noise Hans-Günther Moser MPI für Physik DEPFET Valencia Sept 29 -Oct. 1 2

Rolling Shutter Picture composed of consecutive stripes Hans-Günther Moser MPI für Physik Shutter slit DEPFET Valencia Sept 29 -Oct. 1 3 Jacques Henri Lartigue – Car Race, France 1912

Injection Veto Data not read before next noise bunch z Hans-Günther Moser MPI für Physik Data spoiled by noise bunch T/2 Injection noise T t Injection noise Data not read before next clear z DEPFET Valencia Sept 29 -Oct. 1 4 Data spoiled by noise bunch T/2 Injection noise + clear Read out every 5µs Still no complete frame without noise T Injection noise + clear Read out every 5µs + common clear 50% of frames ok t

Consequences 50 Hz Injection: 400 µs noisy out of 20 ms (20%) Hans-Günther Moser MPI für Physik Ganging of 4 regions => 5µs frame rate global clear in time with veto 50% efficient during injection veto cycles 10% overall inefficiency (gain 10%) However: unganged readout: ½ % dynamic inefficiency (read/clear cycle) 4 regions ganged: 2% dynamic inefficiency more power (4 times!) larger pedestals (dynamic range) mode switching (no global clear during normal cycles) (pedestals? ) (however injection noise disappears gradually) quadruple effects of noisy/dead pixels On the other hand: less bandwidth due to better trigger rejection (2 x) DEPFET Valencia Sept 29 -Oct. 1 5 Recommendation: keep double ganging as option!