I could use some help figuring out my Target ADU for Flats. I read in a post on CN that Flats should be taken in the Linear portion of the camera’s sensor and that to determine the linear vs. non-linear ADU levels that you can construct a light curve. I constructed the attached light curve for my Atik 16200 on an EdgeHD 1100 with an EL Panel sitting on top of the OTA when pointed at Zenith.
The attached graph was created by taking a series of flats incrementing the exposure time by 0.1 seconds for exposure times between 0.4 and 1.5 seconds. Based on the Median ADU levels this exposure time appears to cover the typical ADU levels of 20 – 30K used for Flats, but I am not able to identify a clear break between the linear and non-linear ADU levels for my sensor. It looks like it might be around 0.9 seconds, but then the curve seems to flatten out again. The lines in the graph are the linear regression trendlines for the Median and Max ADU data points. I ran the sequence 3 times and took the average of the ADU levels for each exposure time.
Can someone familiar with camera sensor linearity help me determine what the best target ADU level would be for flats for this ATIK 16200 sensor based on the attached light curves? I have been having issues processing my images where the flats are not removing the vignetting or the dust motes and I think it has to do with the fact that I am not shooting flats at the correct ADU level. Any guidance that could be offered would be greatly appreciated.
Update: I just got some advice on another forum that I need to keep exposing for longer times in order to get to the non-linear ADU levels for my sensor. I was operating under the apparently false understanding that the Median ADU had stay below 30K. I have been advised that the Median ADU can go as high as 50K before I will start to saturate the pixels and see break in the trend line. I will repeat the exercise tomorrow with longer exposure times.
The usual advise it to get the median to about half the max ADU which is your case is about 32k. This is the value that SGP uses for the flats wizard. So based on your graph, 1.3 seconds. However, with my cameras I have found that that short an exposure is problematic. If you can reduce the intensity of your light source, try for an exposure of 15 to 30 seconds with an median ADU of 32k.
Hello again. I re-ran my flats sequence again today to collect the an average Median and Max ADU values of three flat images take at each of the shown exposure times. This time I kept increasing the exposure time by 0.1 second until I saw a clear break in the Median ADU Curve. I was surprise to find that the 16200 sensor remains linear almost to the point of saturation at 65535 ADU.
Given these results, it would appear my Target ADU for Flats for this sensor should be around 45K. This is close to highest Median ADU value before pixels start to get saturated. This ADU level should still give me enough “breathing room” for variations between different flat sequences. Let me know you folks agree with my assessment or do you think I should use a lower ADU level.
Note: I am running my Spike-a-flat EL Panel at around 15% of the max brightness setting which is about as low as I can go before I start to get a flicker. I have already added 3 layers of white Plexiglas to the front of the panel to dim it down. I may need to add a few more layers if you guys still think the 1.5 second exposure time is too short of an exposure to get to the target ADU. This camera does have a mechanical shutter.
I appreciate the help to double check my data and to confirm I am reaching the correct conclusions about my Flat Target ADU levels.
I think you are trying the wrong thing by looking for the non-linear part of the curve. As I said before, use a median ADU of 32K and if you can dim your light source try to get to an exposure of between 15 and 30 seconds. Especially with a mechanical shutter, you need to have a longer exposure time.Using 1/2 the max ADU is the standard way to do flats.
Thanks DesertSky. I will give 32K ADU Flats a try. I know when I was using Flats with an ADU around 25K they did not effectively remove the dust motes and vignetting. Hopefully the higher ADU level will work. Do you recommend using the same ADU level for all the LRGB + Ha, OIII and SII filters?
32K is for all filters. 25K is a little low but it should have been ok for processing. If it didn’t work, I would look elsewhere at what processing steps you used.
