LowNoise Readout of TES Detectors with Baseband Feedback Frequency Domain Multiplexing
Authors
 First Online:
 Received:
 Accepted:
DOI: 10.1007/s1090901205778
 Cite this article as:
 den Hartog, R., Audley, M.D., Beyer, J. et al. J Low Temp Phys (2012) 167: 652. doi:10.1007/s1090901205778
Abstract
SRON is developing an electronic readout system for an array of transition edge sensors (TES) based on the techniques of frequency domain multiplexing (FDM) and baseband feedback (BBFB). The astronomical applications of our system are the readout of soft Xray microcalorimeters in a potential instrument on the European Xray missionunderstudy Athena and farIR bolometers for the SAFARI instrument on the Japanese mission SPICA. In this paper we demonstrate the simultaneous locking of up to 51 BBFB loops. While locked, the inband readout noise of the loops is shown to reach the \(\mbox{10~pA/}\sqrt{\mathrm{Hz}}\) level required for these missions.
Keywords
Transition edge sensors Multiplexed readout Bolometers Microcalorimeters Frequency domain multiplexing1 Introduction
Requirements on the FDM readout system for the Athena XMS and Spica Safari instruments
Athena XMS 
SPICA Safari  

# pixels / channel 
16–32 
160 
total # pixels 
1024 
4100–6400 
readout mode 
triggered 
continuous 
resolution 
2.5 eV FWHM at 6 keV (\({\equiv} 10^{17} \mbox{~W/}\sqrt{\mathrm{Hz}}\) NEP) 
\(\mbox{24} \times 10^{19} \mbox{~W/}\sqrt{\mathrm{Hz}}\) optical NEP 
noise at SQUID input 
\({<}14\mbox{~pA/}\sqrt{\mathrm{Hz}}\) 
\({<}10.5\mbox{~pA /}\sqrt{\mathrm{Hz}}\) 
signal bandwidth 
5 kHz 
<60 Hz 
detector power plateau 
10 pW 
10 fW 
crosstalk 
<2×10^{−4} 
<1×10^{−4} ^{a} 
dynamic range density 
\({<} 6.4\times 10^{5}\ \sqrt{\mathrm{Hz}}\) 
\({< }7\times 10^{4}\ \sqrt{\mathrm{Hz}}\) 
carrier frequency range 
1–4 MHz 
1–3 MHz 
carrier freq. spacing 
100 kHz 
12.5 kHz 
carrier freq. accuracy 
±10 kHz 
±1.2 kHz 
LC filter quality Q 
1400–5600 
>7000 
2 Experimental SetUP
3 Measurements
Missing from these openloop noise measurements is the contribution from the feedback path and the actual confirmation that the noise in the signal band around the carrier also stays below the required level. The right panel shows therefore a measurement of the noise in a ±25 kHz band around one of the demodulation frequencies (which correspond usually oneonone to LC resonance frequencies), taken at the location of the DAC that generates the baseband feedback signal (point C, Fig. 1). The bath temperature in this measurement was set to 100 mK, which is above the T _{C} of the TES detectors. Since then no carriers are necessary to keep the detectors in a setpoint, it was possible to reduce them, either individually or simultaneously, to zero, after the BBFB loops were closed. The peak in the noise spectra around the resonance frequency contains information about the inband readout noise, which is equal to the peak level corrected for the Johnson noise of the resistor. The normal resistance was measured to be 66±3 mΩ in an openloop Q measurement. Taking into account a 3 mK uncertainty in the bath temperature, this results in an estimated Johnson noise of \(9.0 \pm 0.5\mbox{~pA/}\sqrt{\mathrm{Hz}}\). In Fig. 3 the derived level of the noise peak is \(12.4\mbox{~pA/}\sqrt{\mathrm{Hz}}\), which corresponds, after correction for the Johnson noise, to a readout noise level of \({\sim} 8.5\mbox{~pA/}\sqrt{\mathrm{Hz}}\) at the SQUID input. It confirms the result obtained in the openloop noise measurements.

Fabrication of LC filter chips with more robust voltage division,

Multiplexed lownoise readout of TES bolometers in a setpoint,

Demonstration of low optical NEPs in multiplexed readout.
Open Access
This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.