SAT detector counts, NET, and overall efficiency from DSR
Colin Bischoff, 2019-11-12
As an input for time-domain sims, I reproduce below the detector counts and per-detector NET from Table 3-1 of the Decadal Survey Report.
IMPORTANT NOTE: In the arXiv version of the DSR, we accidentally swapped the per-detector NET for 95 and 145 GHz. This has been fixed in the Github repo (thanks Ben!) and it is noted in this google doc. The table below has correct values.
|Frequency||# of optics tubes||total # of detectors||NET per-detector [μK sqrt(s)]||beam FWHM [arcmin]|
The DSR also contains Table 2-1 (reproduced below), which lists the map depth and total survey weight (Q and U) for the ultra-deep survey. These are the numbers that are used for parameter forecasts, map-based Data Challenge simulations, etc.
We can use the detector counts and per-detector NET from the table above to calculate an idealized version of total survey weight (for a seven-year survey):
τ = 7 years × 365 days / year × 86400 seconds / day = 2.2e8 seconds sw_ideal [μK-2] = τ × Ndet / NET2
If we take the ratio of the Table 2-1 forecast survey weight to this idealized survey weight, we obtain a value for the end-to-end efficiency that is being assumed in forecasting. The performance-based forecasting method uses the difference between idealized and achieved performance from BICEP/Keck, with SAT 30/40/85/95 GHz efficiency extrapolated from BK 95 GHz, SAT 145/155 GHz efficiency extrapolated from BK 150 GHz, and SAT 220/270 GHz efficiency extrapolated from BK 220 GHz, which accounts for the pattern of efficiency values shown below.
|Frequency||Q/U rms [μK arcmin]||Total survey weight [μK-2]||End-to-end efficiency|