DECam read noise when on sky is sufficiently low and stable (<= 15 electrons/pixel) [DECam TD-12]
Measurements of the read noise (and gain) are presumed to already exist from lab tests, probably also from commissioning, and these values are presumed to be within specification prior to the start of SV. The main purpose of this test is to verify that the read noise also meets the specification during regular operations, and in particular when the readout is performed during a telescope slew to a new position.
Nighttime test, non-photometric conditions. The gain is assumed to be known.
- Take 2 bias frames (B1,B2) and slew to a new position during each readout
- Measure the read noise = gain*std(B1-B2)/sqrt(2) for each chip
- The measurement could be made with the
noiseallnewscript available from https://sites.google.com/site/decamtroubleshooting/analysis-software
- Repeat measurement several times during SV to insure the read noise is stable
- Repeat for exposures taken during slew and all motor motions.
- While not a requirement, these data should also be inspected for time-variable pattern noise (e.g. variations in the 2D bias pattern)
The read noise must be <= 15 electrons/pixel.
If the read noise substantially exceeds the specification, we will need to either mitigate sources of noise in the dome or further electrically isolate the camera. This will require involvement by the camera team and the telescope staff.
Numerous bias frames were obtained throughout the SV period. These demonstrate that the read noise is within the specification. The read noise was also calculated from the overscan region of frames obtained during MiniSurvey observations under OBSTAC control with Jiangang Hao's noiseall script and indicate that the read noise meets the specification. Finally, the Eyeball_Examination squad reported a number of images that exhibited substantial noise in the form of a 'diagonal stripe pattern' early in the SV period. This noise was identified as due to the ion pump and the pump was turned off. Eyeball inspection of images obtained after the ion pump was turned off no longer exhibit this noise.
- A description of the Fermilab measurements is here: https://sites.google.com/site/decamtroubleshooting/8-measurement
- Fermilab codes are available here: https://sites.google.com/site/decamtroubleshooting/analysis-software
- Code details are here: https://code.google.com/p/decam-fermi/source/browse/#svn%2Ftrunk