Queued Service Observations - QSO-SNR mode

The QSO-SNR mode (using condition-driven exposure times with a SNR goal rather than fixed exposure times) has been tested on ESPaDOnS and MegaCam for a few semesters now. Following a SAC recommendation, the QSO-SNR mode will be the default operational mode for ESPaDOnS and MegaCam, starting in 2017A.

This mode is NOT used for non-sidereal observations, very short exposures of 30 seconds or less, or observations that require fixed exposure times designed to ensure a specific spacing in time between observations.

On ESPaDOnS, our experience also shows that the QSO SNR mode is not optimal for (1) very low SNR goals (SNR of 30 and below), (2) targets with unknown temperature and SNR wavelength goal different from 730nm, (3) targets withtemperature < 3000K, and for (4) targets which are intrinsically variable (with timescales ~ exposure).

To opt out of the SNR mode, PI must include a justification in their proposal. All proposals, whether they will be performed with the SNR mode or not, must request the number of hours needed to reach the scientific goals under specific sky constraints, and justify the exposure times and corresponding SNR. In particular, please ensure you have added calculations from the relevant Exposure Time Calculator.

In addition to the material presented below, interested users may consult the ESPaDOnS poster from the UM2016, which covers the SNR mode, and a 2016 SPIE paper by Devost et al. The latest SNR report for ESPaDOnS, SNR-QSO with ESPaDOnS: Tests, impacts, and feasibility is also available. An early paper on QSO-SNR for MegaCam was published in the proceedings of ADASS 2013 by Cuillandre et al.

Please contact the qsoteam (qsoteam -=at=- cfht.hawaii.edu) if you have questions or concerns.

1. What is the QSO-SNR mode?

In SNR mode, we integrate on the sky until the SNR requested by the PI of a program is reached, or until the original exposure time is reached, whichever happens first.

2. Is the SNR mode used on all instruments?

No. The SNR mode is not used on WIRCam or SITELLE. The SNR mode is used on MegaCam and ESPaDOnS only.

3. Which programs are not suitable for the SNR mode?

Programs with very short exposures (30 seconds or less) are not worth executing under SNR mode. MegaCam programs that use non-sidereal tracking or guiding are not suitable for SNR mode. Programs that have exposure times designed to ensure a specific cadence of observations, or a specific time interval between exposures, are not suitable for SNR mode.

On ESPaDOnS, our experience also shows that the QSO SNR mode is not optimal for (1) very low SNR goals (SNR of 30 and below), (2) targets with unknown temperature and SNR wavelength goal different from 730nm, (3) targets withtemperature < 3000K, and for (4) targets which are intrinsically variable (with timescales ~ exposure).

If you are writing a MegaCam or ESPaDOnS proposal for a project that falls under those categories, please make sure you mention this in your proposal.

The SNR mode is the default mode for MegaCam and ESPaDOnS, starting in 2017A. Unless the proposal has a justification for not using the SNR mode, the program will be performed under the SNR mode.

4. At night, is the SNR mode always enabled?

Programs are only executed if the current sky conditions meet (or exceed) the constraints specified by the in PH2, whether the QSO-SNR mode is used or not.

On MegaCam, if the measured Image Quality is the same as the one used to compute exposure times in PH2, the full exposure time will be used. If the Image Quality and other parameters are better than those used to compute exposure times in PH2, a shorter exposure time will be used. (If the measured IQ is higher than the one indicated in PH2, the program is not executed.)

On ESPaDOnS, if the sky conditions are not stable (variable Image Quality or variable extinction), the SNR mode is not used for polarimetric observations.

5. How is the exposure time determined for ESPaDOnS SNR mode observations?

ESPaDOnS has an exposure meter (EM) that picks up a few percent of the object's light and records the flux in real time (see red line in the figure below). The EM has been calibrated in such a way that we know when we have reached the SNR (blue curve in the figure below) indicated by the PI at the wavelength requested (green text and horizontal line in the figure below).

The exposure is automatically stopped when the EM has reached the target SNR (the blue curve intersects the horizontal green line), or when the exposure time entered in PH2 has been obtained, whichever happens first.

An exposure can not be lengthened.

6. How is the exposure time determined for MegaCam SNR mode observations?

For MegaCam, an initial measurement of the sky conditions (Image Quality, extinction, airmass) is performed, and those values are used in the MegaCam ETC along with the indicated requested SNR. The exposure time is calculated right before obtaining an observation.

Depending how the program is setup, the initial measurement can be obtained with a "snap" short exposure or during the first exposure of a dithering pattern. When a dithering pattern is used, individual exposures can be shortened, or the number of exposures in the dithering pattern can be decreased.

For a dithering pattern with a minimal number of exposures specified by the PI, the exposure time computed at the beginning of the dithering pattern is divided by the (minimal) number of exposures.

During the dithering pattern execution, the (quadratically) cumulated signal-to-noise obtained so far is estimated, and the remaining signal-to-noise goal (that would add quadratically up to the overall signal-to-noise goal for the target) is used to compute the remaining total exposure time. Again, this exposure time is divided by the number of remaining mandatory exposures of the dithering pattern (if more than two are remaining), or defines the exposure time of the next exposure otherwise.

Two safeguards are in place to avoid runaway situations: the number n in the chosen (L)DPn dithering pattern is considered as a maximum number of exposures.

A maximum cumulated exposure time is also required, and in practice should not exceed the exposure time given by the ETC under observing conditions expected for the program by more than say 20%.

The software will ensure that those numbers are not exceeded even if the signal-to-noise goal is not achieved.

This situation can happen in degrading observing conditions, and is a good indication for the observer to switch to another program.

7. What if the flux unexpectedly drops during an exposure?

A MegaCam exposure taken under SNR mode will not be validated if the requested SNR is significantly far from the requested SNR (due to the seeing blowing up, or to the arrival of a cloud). Since the exposure time was calculated for different sky conditions, the SNR will not be reached. The exposure will be repeated (either right away, or once conditions have improved).

An SNR mode ESPaDOnS exposure that doesn't reach its requested SNR will not be validated (and will be repeated) if it differs by more than ~20%.

If conditions degrade significantly during a polarimetric sequence, the SNR mode will be disabled and the polarimetric sequence restarted.

8. What are the advantages of the SNR mode?

If the IQ is better than the one used to calculate the exposure times, then the SNR will be reached in a time shorter than entered in PH2. We will then save time on the sky, which allows the QSO team to perform more observations for the same program, or for other highly ranked programs.

9. What happens to the time that is saved?

The time saved by using the SNR mode is not returned to any program in particular. If an A-ranked program is performed in SNR mode, it is very likely that at least part of the time saved on the sky will be used for that program to hasten its completion. This is also true for B- and C-ranked programs, but to a lesser degree.

10. If all my observations were performed using fewer hours than my allocation, can I enter more observations to fill up my allocation?

Under some circumstances, and with prior approval of the QSO Team, PIs may enter a few more observations than their allocation; this is used by the QSO Team to add flexibility, and not necessarily to gather more observations than initially planned.

11. Can my program be considered complete even if I haven't obtained all my allocated hours?

Yes. If all the observations for a program have been obtained with fewer hours than the allocated time, while still reaching the requested SNR, then the program is considered complete.


Back to the QSO main page

For additional information, please contact the QSO team qsoteam -=at=- cfht.hawaii.edu