When trying to take flat fields with the internal Qz lamps (Cass cage calibration unit), using the 10% flux levels that we used back in October 2015, there was no flux in the ARCoIRIS spectra. Eventually, we realized that the new calibration system is extremely nonlinear in its flux-vs-power level setting, and has very different settings compared to its as-installed settings. After much trial and error, and visual inspection of the calibration unit hardware, we found that appropriate flatfield flux was achieved with power levels of 70-75%.
One additional problem was that remote TCS-GUI operation of the Qz flat-field lamps seemed to yield *much* lower flux levels than manual control of the system from the Cass Cage. We visually verified this effect by removing the housing unit of the Qz lamp, and by-eye observing the lamp with both manual and TCS-GUI operation at 100% flux. For the TCS-GUI remote operation, the bulb was fairly bright, but not opressively so. For the manual operation from inside the Cass cage, the bulb was so bright as to force us to look away from the bulb and shield our eyes. This seems very odd.
Many thanks to Hernan Tirado, Javier Rojas, Humberto Orrega and Claudio Aguilera who spent four hours diagnosing these problems.
As a note to the ETS-folks, we should try to explain why the flat-field flux of the TCS-GUI remote operation is so much lower than the tests we ran in October 2015 with the exact same set-up. Something must have changed.
| Total Times | Time Observed | Time Engineering | Time lost technical | Time lost weather | Time lost other | Total Program Time |
|---|---|---|---|---|---|---|
| 2016-01-23 | 0.0 | 8.0 | 0.0 | 0.0 | 0.0 | 8 |
Links
[1] mailto:djj@ctio.noao.edu
[2] http://www.ctio.noao.edu/noao/field-collection/field-nr-prog/2899/delete?destination=print/content/enr-ctio-blanco-4-m-2016-01-23
[3] http://www.ctio.noao.edu/noao/field-collection/field-nr-prog/2899/edit?destination=print/content/enr-ctio-blanco-4-m-2016-01-23
[4] http://www.ctio.noao.edu/noao/field-collection/field-nr-prog/add/node/9258?destination=print/content/enr-ctio-blanco-4-m-2016-01-23
First night of last ARCoIRIS engineering run.
Early part of the evening, engineering tasks centred on software updates with Marco and flat-field calibration system problems. We remained closed, and very high humidity from 02:30 (UT) onwards has kept us closed the remainder of the night. Several engineering tasks completed:
(a) Marco Bonati has now included a ¨write expo¨ feature into the spectrograph detector window, which writes out the last read of an ARCoIRIS exposure to be read out separately into an image, so one can check for evidence of saturation, that might not be present in the delivered difference image science spectra (note to self: for the ¨expo¨ images, data values are below a floor of 65535 --> that means all pixel values are 65535 - datum) . A very useful addition, thanks Marco.
(b) Katelyn Allers has made vast ameliorations to her data reduction suite of Spextools codes, including improved wavelength solutions (especially for order 7), which now successfully work for early ARCOIRIS commissioning run datasets using either OH- emission sky lines or discharge lamp arc spectra.
(c) We tested the DECam LED system and the COSMOS top-ring flat-field lamps with the DECam flat-field screen to test appropriate levels of a colour-balanced set of flat field data. The DECam LEDs are far too low in flux to provide good ARCoIRIS flat-fields, and in fact, for the COSMOS lamps, we found them to give a far better colour balance than the Qz lamp spectrum emitted from the Cass cage calibration lamp unit. Extensive tests with COSMOS lamp power and flux levels have allowed us to settle on an efficient and statistically sound flat-field data taking procedure with the DECam flat-field screen (thanks to Sean Points for remote assistance with operation of the top-ring COSMOS lamps).
A perfectly good engineering night, especially as the sky conditions were poor.
DJJ, KNA, ES