I bought a ZWO ASI 1600MC at the end of August last year, which is winter time here in Australia. In my initial imaging I didn’t notice any issues with noise in my images and I was imaging for up to 3 minutes. More recently however, I have been really struggling with amp glow and it now always requires a huge amount of work to reduce the glow in my images to an acceptable level for further processing. In fact the glow limits what I can do with these images. Naturally I have to do a fair bit of cropping, which greatly reduces the working area of an already fairly small sensor. This issue is generally not eliminated by calibration images, in fact as we'll see the darks often make it worse. The glow is right around the periphery of the image but is at it’s worst towards the right top and bottom corners.
This is a boosted stretch of a single dark, 240 sec exposure, Gain 0 and Temperature of -20C which illustrates quite well where it is worst.
https://www.dropbox.com/s/7mjwqr7qxq7k5xs/Single%20Dark_240sec_Gain%200_Temp%20-20C_boosted%20stretch.JPG?dl=0
So why did I not have a problem at the outset, but I do now? My theory is that it has to do with the ambient temperature. In winter the camera did not have to work hard at all to get down to -20C and moreover the camera itself was cool or mildly warm to touch. Now when the ambient temperature is around 20C, the camera is working much harder to pull the sensor temperature down to -20C. In fact I notice that the whole front of the camera is really hot to touch……I’m guessing 50C-60C. So is the cooling working? Well yes, the fan is blowing and the set temperature is being reached. I notice that the camera is capable of pulling the temperature down by about 41C-42C, so the power consumption is close to or at 100%.
To test this out I ran two calibration runs of 40 dark images, with an exposure of 240 secs, Gain 150, Set Temperature -20C. The first one was performed with the camera in the fridge (Temp 6C) and the second in my study (Temp 22C). In the case of the test performed in the fridge, the camera was cool to touch and the power consumption was only ~ 20%. For the test in my study, the camera was again very hot to touch and the power consumption was ~ 100%. The following are the stretched images of the MasterDarks I created from these data after image integration/ stacking. The MasterDark from the fridge test has a mean ADU of 1.2, whereas the one from my study (i.e. 22C ambient) has a mean ADU of 5.1. The amp glow in the right hand image is clear to see in the top and bottom right corners, whereas it is absent in the left hand image.
https://www.dropbox.com/s/485uswx4h79pkvy/MasterDarks_120%20sec_Gain%20150_Set%20Temp%20-20C_Ambient%20Temp%206C%20vs%2020C.JPG?dl=0
Clearly if I run two dark calibration runs, where the conditions are all the same, then the darks should be identical regardless of the ambient temperature. This is not the case as I have shown here.
What makes this worse is that MasterDarks so produced cannot be reliably subtracted from the lights! This is because the amount of amp glow depends not only on the ambient temperature but on how long the camera has been working. This is in turn because there is time dependency on the warming of the internals of the camera and therefore a variable amount of heat that is leaking to the sensor, warming its periphery and of course changing the dark current noise produced there.
Theoretically subtracting a MasterDark to calibrate the individual lights ought to substantially remove the amp glow, since it should be present in both the lights and the darks of the same duration exposure. But that only works if the dark current noise is not varying over time. Let me illustrate this with the latest imaging run I did on Thor’s Helmet. This was on my C14 Edge HD, with the Lepus Telecompressor x 0.62 focal length reducer and the ZWO ASI 1600MC.
I took 38 lights with a Gain of 150, set Temperature of -20C and exposure time of 300 secs. I then took 40 darks with the same settings, 40 Bias and 40 Flats.
After normal calibration, with PixInsight, the following was the MASTERSTACK produced in PixInsight. Ugh!
https://www.dropbox.com/s/nvam4yv12vvms82/MASTERSTACK_300%20sec_Gain%20150_Temp%20-20C.jpg?dl=0
Here is an individual debayered sub. I debayered it so you can see the amp glow. This is the subtle blue-white hue in the corners of the image…again top and bottom right, but also right around the periphery:
https://www.dropbox.com/s/5lsb1qc5t1moujr/Single%20Debayered%20Raw%20Light.jpg?dl=0
I tried subtracting the Bias only….no real difference, however, when I subtracted the darks only (no bias, no flats) from this same sub this was the result. The amp glow looks worse and has a green hue.
https://www.dropbox.com/s/ernulkqhfi0d82x/Single%20Debayered%20Light_Darks%20calibrated%20only.jpg?dl=0
Clearly the darks are part of the problem.
When I looked at the stretched MasterDark, this is what I got:
https://www.dropbox.com/s/82zce3b83n8lbou/MasterDark_300%20sec_Gain%20150_Temp%20-20C.JPG?dl=0
The MasterDark clearly reflects the amp glow, however the fact the dark subtraction didn’t cancel out the amp glow in the light subs, suggests that the darks do not reflect the dark current noise in these lights. I can only conclude that by the time the darks were taken at the end of the imaging run, they no longer represented the dark current in the lights because something had changed over time. My theory is that the camera had continued to get hotter through the imaging run such that more and more heat was getting to the sensor and warming it, particularly around the periphery and the right side. This despite the temperature probe telling me that the sensor was being cooled to -20C.
I was also wondering about the green hue where the amp glow is occurring in my calibrated images. I suspect it has to do with the colour of the amp glow noise which seems blueish. If the darks, having been created when the camera was warmer now have more amp glow than the lights which were taken earlier, then when this extra blue is subtracted what will be left is the dark green and red hue I see in the calibrated images. How do I know it has red? Well when I apply SCNR in PixInsight, to subtract green, the glow turns red.
To test this I integrated the first 10 dark images and then the last 10 dark images of 40 from my run on Thor’s Helmet and then compared the two. The aim was to see if my theory about the camera warming and increasing the amp glow bears out.
https://www.dropbox.com/s/cpsuwftv1r5dgwi/Changing%20Darks%201st%2010%20vesus%20last%2010%20of%20a%20run%20of%2040%20and%20the%20difference.JPG?dl=0
This proves fairly conclusively that the amp glow is getting worse through the run. The respective ADUs were 1.6 and 1.9. I then used Pixel Math in PixInsight to subtract the two…the 1st 10 from the last 10 (see bottom image). This shows that the amp glow has increased as there is a positive amount of glow in the top and bottom right hand corners and interestingly more in the middle as well.....(bottom image)
This proves conclusively that the darks are inconsistent over time and therefore cannot be relied upon to be subtracted from the lights.
In conclusion, there is an unacceptable level of amp glow with my ZWO ASI 1600MC when the temperature is something over 10C, which gets worse as the ambient temperature gets higher. I cannot say whether this is a design fault or a flaw in my particular camera. The level of this amp glow virtual renders the camera unusable in the conditions I often encounter during the night time in summer when imaging in my observatory in the central west of NSW.
It is my belief that when the camera is operating near maximum capacity which occurs when the ambient temperature is around 20C or above, the heat is not being adequately conducted away from the body of the camera to be removed by the cooling air. Instead it is causing the temperature of the whole body of the camera to be elevated. In due course it is noticeably hot to touch.
It would be difficult to imagine that the high temperature of the camera would not cause at least some heat to be conducted to the sensor causing localised heating at the periphery. This in turn increases the dark current noise being generated in those areas. Furthermore, since the camera’s temperature is time dependent, given it starts at ambient temperature and is then getting hotter, darks taken at a different time to lights cannot be relied upon to reflect the dark current noise in the sensor when the lights were taken nor to subtract the effect from the light images.
macnenia
Posts: 1
Joined: Thu Feb 22, 2018 12 pm
