The Simons Observatory: Studies of Phase Drift in RF Transmission Lines from the First Large-Scale Deployment of Microwave Frequency Multiplexing for Cosmology

Fulfilling the science goals of the Simons Observatory, a state-of-the-art cosmic microwave background (CMB) experiment, has required deploying tens of thousands of superconducting bolometers. Reading out data from the observatory’s more than 67,000 transition-edge sensor (TES) detectors while maintaining cryogenic conditions requires an effective multiplexing scheme. The SLAC microresonator radio frequency (SMuRF) electronics have been developed to provide the warm electronics for a high-density microwave frequency multiplexing readout system, and this system has been shown to achieve multiplexing factors on the order of 1,000. SMuRF has recently been deployed to the Simons Observatory, which is located at 5,200 m on Cerro Toco in Chile’s Atacama Desert. As the SMuRF system is exposed to the desert’s diurnal temperature swings, resulting phase drift in RF transmission lines may introduce a systematic signal contamination. We present studies of phase drift in the room-temperature RF lines of the Simons Observatory’s 6 m large-aperture telescope, which hosts the largest deployment to date of TES microwave frequency multiplexing to a single telescope. We show that these phase drifts occur on time scales which are significantly longer than sky scanning, and that their contribution to on-sky in-transition detector noise is within the readout noise budget.

💡 Research Summary

The Simons Observatory (SO) operates more than 67,000 transition‑edge sensor (TES) bolometers across four telescopes on Cerro Toco, Chile. To read out this enormous detector count while keeping the detectors at 100 mK, the project employs microwave frequency multiplexing (µMUX) using the SLAC Micro‑resonator Radio Frequency (SMuRF) warm‑electronics system. Each SMuRF unit drives two Advanced Mezzanine Cards (AMCs) that generate and demodulate probe tones in the 4–6 GHz band, achieving multiplexing factors of order 1,000 per pair of RF coaxial lines.

For the 6 m Large‑Aperture Telescope (LAT), the SMuRF crates and associated cabling are mounted directly on the receiver cabin (LA TR), which is kept isothermal with the telescope optics. Consequently, the warm electronics and the room‑temperature RF transmission lines are exposed to the diurnal temperature swings typical of the Atacama desert (tens of degrees Celsius over a day). Temperature‑dependent changes in cable length and dielectric constant can cause a systematic phase drift in the RF signals, potentially contaminating the TES data if not properly characterized.

The authors address this risk by exploiting “off‑resonance” (pilot) tones that SMuRF can generate in the gaps between resonator frequencies. Because these tones do not interact with the cryogenic resonators, any measured phase change originates solely from the warm coaxial cables. During a 22.6‑hour observation on 2 June 2025, 290 off‑resonance tones from 30 AMCs were streamed continuously, while a thermometer in the receiver cabin recorded ambient temperature. The phase of each tone, expressed as θ = atan2(Q, I), was converted to a time delay τ = θ/(2πf). The data show a clear correlation between τ and cabin temperature, with a slope of roughly 0.5–1 ps °C⁻¹. Over a full day‑night cycle the total drift is only a few picoseconds, far smaller than the timescale of sky‑scanning (seconds).

To assess whether such a drift could masquerade as a detector signal, the authors consider the calibrated I‑Q response of a resonator. After SMuRF’s firmware‑level rotation, perturbations around resonance appear purely as changes in the Q component. A small phase rotation θ therefore induces a Q‑shift δQ = I_offset_res · tan θ, where I_offset_res is the post‑calibration distance of the resonance point from the I‑axis (median ≈ −3.3 × 10⁴ Hz). Converting this Q‑shift to an equivalent TES current requires the slope ⟨df/dI_TES⟩, measured from SQUID‑TES flux‑ramp curves. The authors find a median ⟨df/dI_TES⟩ ≈ 5.71 × 10⁻² Hz pA⁻¹. Using this conversion, the phase‑drift‑induced current noise was computed for a typical 40‑minute constant‑elevation scan. The resulting amplitude spectral density (ASD) lies well below the ASD of an on‑sky, in‑transition 150 GHz detector and is comparable to the “open SQUID” channel that contains only readout noise. In other words, the phase‑drift contribution is a negligible fraction of the total detector noise budget.

The study concludes that, despite being exposed to unregulated temperature swings, the chosen True Blue 205 ruggedized SMA coaxial cables (four 2 m runs per SMuRF) and the SMuRF tone‑tracking firmware together keep phase drifts at the picosecond level. This drift is both temporally slow and amplitude‑small enough that it does not compromise the scientific performance of the LAT. Consequently, the first large‑scale deployment of µMUX readout for cosmology demonstrates that room‑temperature RF line stability is not a limiting systematic for current CMB experiments, even in harsh, high‑altitude environments. Future upgrades (e.g., the Advanced Simons Observatory) can therefore rely on similar warm‑electronics architectures without needing elaborate thermal control of the RF cabling.