SPIRou Early Performances
Last update: October 2017.

The numbers below show the early-performance measurements performed in the lab from June to November 2017. They will be updated after re-integration at CFHT and technical commissioning. Note that in the lab, an engineering detector with 2k x 2k, 18-micron pixels has been used. Detector characteristics come from early images obtained at TIS, before the detector was delivered to the SPIRou team. See the footnotes below for explanations.

Throughput in JHK (1) 5-8, 8-14, 8-14%
Peak SNR in 1h on H=11, 3100K star 100-130 per pix at 1.65micron
Magnitude for S/N=10 in 1h (2) mH=14-14.5
Magnitude for 1m/s in 900s (3) mH~7.0
Thermal background at 2.35 microns ~40 ph/s/A between at 0-5C
Sensitivity of the guiding (4) TBD
On-sky aperture 1.29"
Wavelength Domain (5) 1.14-2.49 microns (measured, see 5)
0.98-2.35 microns (nominal)
Spectral Resolution from calibration lines (6) 61,000 (measured and predicted with H2RG)
Spectral Resolution from the broadening of stellar spectra (6) 73,000 (predicted with H4RG)
Pixel size 2.3 km/s
Radial-velocity accuracy (7) 0.2 m/s (internal)
TBD (on RV standards)
Method spectral calibration: use of simultaneous FP
Polarimetric sensitivity in stellar lines (8) TBD
Polarimetric crosstalk (9) TBD
Type H4RG
Dimensions 4k x 4k
Pixel size 15 microns
Red cut-off (50% QE) 2.45
Quantum Efficiency over JHK > 80%
Cosmetics (clusters) < 0.1% bad pixels
Readout mode Ramp fitting
Dark current < 0.01 e-/s
Readout Noise (CDS) 20 e-
Readout Noise (ramps) (10) < 10 e-
Overheads Slew and start guiding (not charged): 90 s
charged time per sequence: 30 s (TBC)
Optical setup (not charged) per sequence: 10 s (TBC)
Number of available hours per night (11) 7 h (estimate)
Precipitable water vapour 50% of time (observing airmass) (12) 2.85 mm
Precipitable water vapour 80% of time (observing airmass) (12) 5.40 mm
Calibrations RV and telluric standards: ~30min per night in 18B
Pipeline processing 18B: first release of pipeline


1. Throughput is difficult to measure in the lab. In the case of SPIRou, mean values in JHK were calibrated from on-sky calibrations using the guide camera, in addition to combining individual transmission curves from optical elements. Another source of uncertainty is the gain of the detector used for these lab tests. These numbers reflect the best of our knowledge and can be considered precise to within a factor of 1.5. They will be refined after first light of SPIRou at the telescope.

2. Limiting magnitude in 1h shows the H magnitude of a 3100K star where a SNR per pixel of 10 is achieved in median conditions (0.6" seeing in H) in a total exposure time of 1 hour. In the lab, this quantity is obtained from throughput measurements and using an exposure time calculator that includes the H4RG noise properties. Throughput measurements will be better defined during technical commissioning.

3. Limiting magnitude for a 1m/s accuracy in 900s is the typical H magnitude of a 3100K star where simulations predict an RV uncertainty of 1 m/s under median conditions (0.6" seeing in H). This uses the noise estimates and the RV content of such a star (Figueira et al, 2016 and Artigau et al, 2017), assuming i) that the star rotation is unresolved and ii) that the telluric lines are corrected for. The RV uncertainty is expected on the same star to be up to 20 m/s if the rotation profile is 10km/s and telluric lines are just masked out.

4. Sensitivity of the guiding is the limiting magnitude below which guiding delivers an rms stability of the input image averaged over a typical exposure better than 0.05". It is in the process of being evaluated from lab tests and will be refined in real observing conditions during the technical commissioning.

5. The current wavelength domain has been measured in the lab using the H2RG which covers only part of the focal plane (88%). One additional red order is present, but the bluest orders covering 0.98 to 1.14 micron are not on the test chip. In the red, 17% gaps are present on the test chip but will be covered by the final detector.

6. The spectral resolution relevant for radial velocity work, Rrv, relates to the gaussian broadening that the spectrograph induces onto the spectrum of a typical M dwarf. In the lab, Rcal, the resolution estimated from calibration lines, has been measured. Both resolutions are not equal when the spectrograph point-spread function (PSF) is not Gaussian. For SPIRou, Zemax predicts Rcal =61.6 K (within specification) and Rrv =71.2K for the H2RG detector, and Rcal =63.8K and Rrv =73.4K for the H4RG detector.

7. Internal RV accuracy has been measured in the lab on continuous series of Fabry-Perot exposures. They have shown internal relative stability at a level of 0.2m/s over 24h, measured on the science fibers and corrected for the absolute drift as measured on the calibration fiber. Light illumination effects and their impact on the RV precision are being evaluated in the lab, and will be better estimated in realistic on-sky conditions following commissioning and science verification.

8. Polarimetric sensitivity in stellar lines is to be measured on a solar spectrum observed with SPIRou in the lab. More to come.

9. A residual polarization crosstalk is observed, especially at low external temperatures.

10. The readout noise in the ramp sampling decreases with the number of exposures, until it reaches a minimum. These values and the way the noise decreases will be documented in December 2017.

11. The observing efficiency is based on ESPaDOnS efficiency and takes into account specific known overheads. Time loss due to instrumental problems may arise in the beginning of the instrument lifetime but are not included here.

12. The amount of precipitable water vapour is expected to have an impact on the spectra and radial velocity measurement. The fraction of time which minimizes the precipitable water vapour may be better adapted for extreme PRV requirements.