Measuring characteristics of ZWO ASI120MM-S camera

Using technics from The AAVSO DSLR Observing Manual (https://aavso.org/dslr-observing-manual), I have investigated my ASI120MM-S camera built on a CMOS sensor.
Measurements were performed in 1x1 bin mode. SharpCap was used to capture images, in native 12-bit mode (without 16-bit stretching!). Measurements were done using a central region of images.
A notebook screen covered by several sheets of paper (to reduce light) was used as a constant light source. A room was darkened.

*** UPDATE 2019-11-03 **** Measurements via SharpCap Sensor Analysis tool were made too.

Gain in e^-/ADU

The gain was measured through Average ADU vs ADU Variance dependencies. Here is a plot for Gain29:

A slope of the line gives a gain in e^-/ADU.
A dependency on the gain in e^-/ADU on ZWO gain is shown in the next plot. Note that measured values are in excellent agreement with those reported by the manufacturer (https://astronomy-imaging-camera.com/product/asi120mm-s)

New results obtained via SharpCap Sensor Analysis tool are almost the same.

*** UPDATE 2019-11-03 **** SharpCap Sensor Analysis tool results

Full well

Using measured system gain, it is easy to get the full well:

Full well for ZWO Gain = 0 (3.5 e^-/ADU) was measured as 14.3 ke^-. ZWO reports 13000 in "Product description" or 14.5 ke^- in the corresponding plot. New results obtained via SharpCap Sensor Analysis tool are almost the same.

*** UPDATE 2019-11-03 **** SharpCap Sensor Analysis tool results

Read-out noise (RON)

RON, measured from zero-exposure frames, was found smaller than reported by ZWO and almost equal for different gains: ~2e^-. This is strange. See also new results obtained via SharpCap Sensor Analysis tool.

*** UPDATE 2019-11-03 **** SharpCap Sensor Analysis tool results

These results are in the excellent agreement with those reported by the manufacturer (https://astronomy-imaging-camera.com/product/asi120mm-s)

Dark current

The AAVSO DSLR Observing Manual shows troubles while measuring the dark current of DSLR camera sensors. I have to admit that my results for ASI120MM-S are also quite uncertain. I've tried to estimate dark current from values of mean pixel intensity of dark images (1st way) and through standard deviation of mean pixel intensity. The camera has no cooler/temperature regulator which introduces additional uncertainty. The values of the dark current were estimated as ~1 .. 2 e^-/s/pixel at room temperature. Additional measurements are needed!

Linearity

The sensor was found to be very linear:

Strange behavior at low ADU for another CMOS sensors was mentioned by Mark Blackford for Canon EOS 600D (https://www.aavso.org/comment/47248#comment-47248) and by Christian Buil for cooled ZWO ASI1600MM camera (http://www.astrosurf.com/buil/CMOSvsCCD/index.html). Buil thinks that that behavior is caused by imperfect timing for very short exposures he used. I think it is not an exposure timing problem rather some features of the CMOS sensor itself (autotuning of dark level?). The next two plots were obtained using dimmer light and longer exposures, it is seen than "stranges" arose at the approximately the same ADU level:

*** UPDATE 2019-11-03 **** Table of results obtained via SharpCap Sensor Analysis tool

Using ZWO ASI120MM-S planetary camera as a photometric device

Almost all professional photometric studies of variable stars are performed with CCD cameras. Many experienced amateur astronomers use CCD too, however alternative photometric devices, i.e. digital SLR cameras (DSLR), are quite popular too. Modern DSLR cameras use CMOS sensors. There are also plenty of CMOS-based cameras in the market made specifically for amateur astrophotography.
Uncooled CMOS cameras with small sensors are dedicated to planetary imaging, however, they also can be used for entry-level deep-sky imaging.
In the current study, the possibility of using such low-level uncooled CMOS cameras for differential photometry was investigated.
The author tested ZWO ASI120MM-S planetary camera as a photometric device. The camera has a monochromatic 1/3" CMOS sensor (4.8 x 3.6mm) AR0130CS. The pixel size is 3.75 x 3.75 μm, the sensor has 1280 x 960 pixels.  This camera provides 12 bit ADC.
The camera was attached to SkyWatcher 150 f/5 Newtonian on EQ5 motorized mount.
Sharp Capture software was used to gather images. Exposures = 10s, camera gain = 29 (it corresponds 1 e-/ADU accordingly to camera's specification).
There were two test run, a rapidly changing variable XX Cygni of SXPHE type was selected as a target. No filter was used. Preprocessing of images (calibration with flat, dark and bias frames) was performed using IRIS software (http://www.astrosurf.com/buil/iris-software.html). Photometry was done in AstroImageJ (a measurement of fluxes) with subsequent processing data in Excel.
Resulting standardized CV magnitudes were binned by 5 points. The light curves for the variable and a check star are presented in Fig. 1

AAVSO Chart X24817BFV
Comp Star   000-BJV-171
Comp V Mag  10.606
Check Star  000-BJV-173
Check V Mag 11.757

Observing conditions: almost full Moon in the 1st night, however, the sky was very clear; in the 2nd night the transparency was worse, there were haze and sporadic cirrus. This caused bigger random errors in the second data set.

Phase curve built using period and initial epoch from AAVSO VSX database is shown in Fig. 2.

We can see that the observed position of the maximum is shifted slightly from the predicted position. To prove the result comparison with ASAS-SN data was made, it is shown in Fig. 3 (my data marked as "Unfiltered with V zeropoint").

It is seen that positions of maxima are in excellent agreement. Probably the shift is caused by tiny period change.

We can conclude that even such a simple device gives good reproducibility of results and satisfactory precision. The next step will be testing with Johnson V photometric filter.