Implementation of frequency domain multiplexing in imaging arrays of microcalorimeters

Transcription

1 Implementation of frequency domain multiplexing in imaging arrays of microcalorimeters Jan van der Kuur, P.A.J. de Korte, P. de Groene, N.H.R. Baars, M.P. ubbers SRON National Institute for Space Research The Netherlands M. Kiviranta VTT Microsensing, Espoo, Finland TD-10 Genova 7 11 July

2 Overview Frequency domain multiplexing Non-ideal behaviour 1. summing point impedance 2. magnetic coupling between filters Bias source multiplexing topologies Conclusions TD-10 Genova 7 11 July

3 Frequency domain multiplexing Operating principle voltage biased with AC source Amplitude modulation by Signals separated in frequency space signals f 1 f 2 f 1 f 2 I Z c SQUID in summing point Noise blocking filters required High count rate low bias source impedance signal power f 1 f 2 frequency [Hz] TD-10 Genova 7 11 July

4 Summing point impedance Z c Mechanism (2 channels) I b Voltage accross Z c current through parallel branches I p(arasitic) /I bias Z c Consequences: f f = Z c Q C 1 I parasitic 1. increased impedance bias circuit 2. cross talk by modulation of I p (power effect) V 1 Z c SQ1 tuning C Z c 0 I p /I b Z c /2 independent f i and f C 2 C n TD-10 Genova 7 11 July

5 Summing point impedance Z c Numerical simulations (32 channels) c / = 5%, f = 200 khz, f = 10 MHz, Z c = 0 at f i Extra series impedance: upto 20% independent of Q = f i / f Also: shift of resonance frequencies Our setup: c < 2 nh, = 90 nh so c / = 2.2%. ow information bandwidth applications: less effects voltage [% of Vbias) voltage due to common impedance (Common inductance 4 nh) Vtes fc= 10 MHz, df= 200 khz Vtes fc= 100 MHz, df= 1MHz channel no TD-10 Genova 7 11 July

6 Magnetic coupling between filters Mechanism (2 channels) I 1 I parasitic Coils form transformer Effect similar to common impedance (no tuning) C 1 k C 2 I p /I c 0.5kf i / f = 0.5kQ Too complicated to tune with capacitor Minimised by geometric separation of filter coils for nearest frequencies kω 1 I 1 f 1 f 2 SQ1 TD-10 Genova 7 11 July

7 Magnetic coupling between filters Numerical simulations (5 channels) f = 200 khz, f = 10 MHz, Q = 50 Extra Z bias scales strongly with k Cross talk less prominent (power effect) No scaling with # of channels (nearest pixels dominant) Coupling > 1% problematic (low resistances s) Vtes [% Vbias] voltage and cross talk as function of "k" 5 coupled coils f 0 = 10 MHz, df = 200 khz Vtes x-talk-2 x-talk-1 x-talk+1 x-talk+2 0 0,01 0,02 0,03 0,04 0,05 coupling coefficient "k" TD-10 Genova 7 11 July

8 Bias source multiplexing topologies Bias per column One bias source per column Multiple frequencies per source Cross talk equivalent to highly coupled filters filter accuracies moderate applicability depends on carrier separation (Q-factor) f 1 f 2 SQ1 SQ2 SQ3 f f f TD-10 Genova 7 11 July

9 Bias source multiplexing topologies Bias per row One bias source per row Filter center frequencies highly uniform Frequency errors can lead to instability No additional cross talk effects applicability dependent on filter reproducibility f 1 f 1 f 2 f 2 f 1 f 1 f 2 f 2 SQ1 SQ2 SQ3 TD-10 Genova 7 11 July

10 Conclusions Summing point impedance lower count rate and cross talk Minimised by tuning capacitor Effects largest for high countrate applications Magnetic coupling between filters similar to summing point impedance Effects scale with operating frequency and band separation Effects largest for low countrate applications Bias source multiplexing feasible, preferred topology is application dependent Simulations required for optimal solutions TD-10 Genova 7 11 July