I ran some tests and found that the dark current on my 1600MM Pro Starts high and drops off sharply over time. The 'knee' is at around 2 minutes with only a small improvement at 5 and 10 minutes.
Dark current is supposed to be linear and this plainly isn't. Has anyone else experienced this?
I'm doing some other tests to see if I can isolate the issue.
Perhaps there is thermal noise lingering from previous sensor readouts.
I compared two series; one where frames were captured immediately after the last had finished (red line), and another series where there was a 10-minute delay between frames. I also reversed the order too - I started with the 600s exposures and ended up with 0.000032 second exposures taken 600 seconds apart (blue).
The 'spaced' series has worse dark current for about 1 min 10 secs, and thereafter has less than half the dark current of the non-spaced series.
Not sure what is going on here. This was shot at gain 0 to capture full wells. I'll try at different gain settings to see if anything changes.
It seems that dark current is not linear across all gains (see below).
The amount of dark current e-/sec drops off over time. That is to say, the dark current still accumulates more the longer you expose for, but that the rate of accumulation decreases over time.
I didn't remove the 1600 from my OAG setup, so I took the following precautions; I had a black 'filter' (black aluminium foil) in place in the filter wheel, and the guide camera in place (no light coming in through the stalk) and aluminium foil in place over the rubber lens cap of the Field Flattener.
I wanted to work with exact ADU levels as near as I could determine experimentally using Stark's method, so the data would be in multiples of each other eg;
76 = 2
100 = 1.5
135 = 1
195 - 0.5
253 = 0.25
313 = 0.125
For each gain setting I shot 250 bias frames at 0.000032s each, integrated them with the settings below in Pixinsight. Then I used statistics to read out the mean value from the integrated master bias.
I then took 2 or 4 dark frames (depending on which run of tests it was - I did a second run to verify some data points) with exposures at 30s, 60s, 120s, 300s and 600s in SGP. Temperature was -15C for all subs (and bias). The setup for the dark frames was the same as for the bias.
In statistics, I took the mean values and averaged them in a spreadsheet.
The result of shooting those gain settings at various exposure lengths is what you see in the post above.
We tested the 1600 MM Pro with your method. It has the similar result with yours.
So thanks for your support first.
For the result, we think it may be the following reasons.
When we have an operation, for the circuit, it's like dropping a pebble on a calm water. There will be some disturbance in the circuit. When the gain is large, the disturbance is greatly amplified. For long exposure, there are some different operations with short exposure.
Beside, consider that the short exposure is different with long exposure in the camera, so I think it is better to measure at long exposure. It means you should use 2 seconds exposure time to replace the image of 32uS exposure.
As you suggested, it changed the data greatly. Now my graph that compares dark noise in e/sec for various gain settings is looking much better.
I'm retesting the data at gain 313 because that series still has a small anomaly that will probably disappear with more data. I'll repost when done.
Thanks very much for your suggestions.
Can I ask - does this mean bias frames should be shot at 2 seconds for this camera to get a truly accurate noise profile? Unless you're using it as a planetary camera, you're unlikely to have subs that are <2 secs.