Going towards the read-out of a 160 pixel FDM system for SAFARI 76 pixels connected R.A. Hijmering R. den Hartog J. van der Kuur J.R. Gao M. Ridder A.J. v/d Linden
SPICA/SAFARI SPICA (JAXA/ESA) Infrared mission ~ 2.5m, ~8K mirror, Background limited Proposed for M5, launch in ~2028 to L2 Explores dusty area and find a route to habitable planets SAFARI instrument on SPICA (SRON + partners) 34-210μm Three band grating spectrometer S-band 2x10-19 W/ Hz 3400 TESs per band 2
FDM requirements for SPICA Pixels /channel Detector resolution Signal band 160 (22 channels) 0.2-0.4 aw/ Hz 60Hz crosstalk <0.2 10-4 Noise at SQUID input Carrier frequency range Carrier frequency spacing Carrier frequency deviation LC quality (Q) 10.5pA/ Hz 1-3MHz 12.5kHz 1.2kHz >2355f0(MHz)~7000 Using in house developed and fabricated superconducting LC filters In house developed TES arrays In house fabricated Front End Electronics and Demux Board And operated using in house developed software 3
FDM readout Circuit Each TES has its own LC filter with unique resonance frequency Multiple carries are send down at f 0 to bias the individual TESs Signal from the TES creates an amplitude modulation read out via SQUID Demodulation at the resonance frequency reveals the signal FDM read out assembly with flexible, superconducting connection to kilo-pixel TES detectors M. Bruijn (poster session 1 wednesday) The development of frequency domain multiplexing readout of TES-based X-ray microcalorimeters for Athena H. Akamatsu (today) 4
Real Circuit 2 coil Main C 3 times voltage step down ~1M - Resistive 2/400Ω - Inductive 2/2000nH - Capacitive in LC 1:9 2 LC filters in unit cell high+low f 0 L gradiometric Bias C L div. 5 5
FDM readout electronics DAC bias supplied bias carriers Summed at SQUID amplifier Amplified by LNA ADC I and Q Remodulated send to input coil by DAC fb for BBFB 6
7 First itteration of the 160 pixel FDM off on step Changing R stray Changing Rn Loading at high R Stays normal No step Varying IVmin step Varying Rn IV curves vary pixel to pixel R n varies considerably Step features Switching neighbors on big effect Saturation power varies with f 0
8 First itteration of the 160 pixel FDM off on step Stays normal Changing R stray Crosstalk! Changing Rn No step Loading at high R System not optimized against crosstalk Varying IVmin Unable to read out 160 pixels simultaneously step Varying Rn Mutual inductance: magnetic coupling between channels Common impedance: impedance outside circuit acting as voltage divider Carrier leakage: openess of filters for neighboring carriers
160V2-pixel experiment LC array TES array PTBC5 SQUID LC array GOALS ~160 pixels read out No effect of crosstalk High yield of TES+LC filters Variation of f <2kHz TES limited noise No read out effects in TES characterisation LC filter chips: 176 resonators (87% yield) Spread over 2 chips 1MHz-3.8MHZ TES bolometer array: 176 TES +8R's (87% yield) T c : 107±3 mk R N : 140 mω P sat,design :20 fw @ 65 mk NEP design : 7x10-19 W/ Hz 9
160 pixel experiment V2 Reduction in wire bonds (L com ) Reduction in inductance division coils 5µH:50nH 2µH:20nH (L com ) Mounting in stages (¼, ½, 1) coil Themometer Connection for coil Absorbing material Nb foil for shielding 10
160 pixel experiment V2 LC 2 LC chips alternating f 0 (M ) Shuffling f 0 (M ) LC filter larger unit cell 3x3mm (M ) Δf in unit cell >1MHz (M ) L com total 6nH to 4.5nH, incl. SQUID L from 2 to 3µH (CL ) } factor 2 reduction Δf from 14 to 16kHz (CL ) LC chip 1 L1/H1 L2/H2 odd rij/kolom 1 2 3 4 1 2 3 4 5 6 7 8 9 10 11 47 3 45 1 133 89 135 91 7 51 5 49 93 137 95 139 55 11 53 9 141 97 143 99 15 59 13 57 101 145 103 147 63 19 61 17 149 105 151 107 23 67 21 65 109 153 111 155 71 27 69 25 157 113 159 115 31 75 29 73 117 161 119 163 79 35 77 33 165 121 167 123 39 83 37 81 125 169 127 171 87 43 85 41 173 129 175 131 11
160 pixel experiment V2 TES Wet etch procedure Detector chip is smaller+symmetric Wiring shorter (M ) Coplanar wiring further apart (M ) Extra resistors for mapping Wiring 24mm ± 0.1mm=120Hz Development of ultra low noise TES bolometer arrays T. Suzuki (Talk Monday) 12
Measured results with one LC filter Connected 38 pixels followed by 76 pixels All TES' and R's connected found 76 Q's between 20k and 60k 76 ESR between 0.8 and 1.2 m ; <ESR> = 1.0 m 13
Resonance frequencies 7 frequencies with multiple seperation due to disconnected pixels Variation in f up to 10kHz -0.1 to 0.4% of f0 Variation in f0 caused by lithographic tolerances, first suspect is dielectric of capacitors 14
Reduction in crosstalk common impedance dl The measured common impedance is 1.5nH 4.5nH is expected and the difference is due to the screening of the input coil by the feedback coil in the SQUID 15
Crosstalk: effect of neighbors 90 80 70 60 50 100 120 140 160 180 Switching on neighbors has no effect on the IV curves or Power Plateaus No influence of neighbors, no noticeable electrical crosstalk 16
TES measurements Power Plateaus vary factor 2 due to lithographic tolerance and production issues with wet-etch DRIE process enables higher fabrication accuracy, ultra low noise ~40% No frequency dependent results, (caused by biasing circuit) Detector noise not electronic noise limited Fabrication of low noise TES array for the SAFARI instrument on SPICA M. Ridder (Poster G1.16 today) 17
Conclusions Significant reduction in electrical crosstalk Close to high yield of TES and LC filters (87%) Frequency separation up to 10kHz >2kHz Device limited noise No read-out effects in device measurements as f 0 dependent results 18