MGTechDVP
22-03-2015, 03:42 AM
Hi All,
I was wondering about long exposure and deep sky imaging using the Imaging source DMK41. I was googling for info online, not much was out there about it, and the small amount all stated that the noise due to heating up of the sensor makes it too noisy for anything over 30seconds... also the ADC was only 8 bit so that would limit the stretchability of the final image.
So I'm thinking if it's only heat and bit depth of the image that the obstacle, (since the software allows upto an hour of exposure) then surely there is a way to cool the CCD and stacking of multiple subs will increase the bit depth.
So I attached two 12V peltiers, heatsinks and fans to the DMK41, ran a few tests and surely the noise did lessen, all but near dissapeared even in a 30 minute test dark frame... all it took in photoshop was a noise reduction of 1 pixel with threshold of 5 to remove nothing but the overactive pixels.
Results are as follows. The focal length used was F10 on my 8" SCT. for guiding I was using an off axis guider, and had to use the DMK21 and PHD for guiding since, there were only 2 stars within the reach of the prism, and neither was bright enough for the nexguide to pickup.
I also found that the focus change very slightly as the CCD reached it's lowest temperature, atleast I hope that that's why the final image is a bit soft, but on both nights the focus was slightly off after the cooler ran for about 10-15 minutes, only a slight adjustment was needed to to refocus the stars to points.
Attached images are:
image 1. The cooling setup
image 2. Uncooled 15 min dark sub
image 3. Cooled 15 min dark sub
image 4. Uncooled 30 min dark sub
image 5. Cooled 30 min dark sub
image 6. raw 10 minute with 520 gain exposure of the M104 galaxy
image 7. M104 with 6 10 min subs and 4 15 min subs stacked and stretched and 5 minutes of unguided R, G & B 1 min subs 1023 gain for color.
Notice how there is no more noise on the 30 minute cooled dark frame test compared to the 15 minute test.
Next time I'll let the DMK cool for 15 minutes before imaging, refocus, grab 10, 20 & 30 min subs as well as more color data.
I think the results are promising, what do you think?
Mariusz
I was wondering about long exposure and deep sky imaging using the Imaging source DMK41. I was googling for info online, not much was out there about it, and the small amount all stated that the noise due to heating up of the sensor makes it too noisy for anything over 30seconds... also the ADC was only 8 bit so that would limit the stretchability of the final image.
So I'm thinking if it's only heat and bit depth of the image that the obstacle, (since the software allows upto an hour of exposure) then surely there is a way to cool the CCD and stacking of multiple subs will increase the bit depth.
So I attached two 12V peltiers, heatsinks and fans to the DMK41, ran a few tests and surely the noise did lessen, all but near dissapeared even in a 30 minute test dark frame... all it took in photoshop was a noise reduction of 1 pixel with threshold of 5 to remove nothing but the overactive pixels.
Results are as follows. The focal length used was F10 on my 8" SCT. for guiding I was using an off axis guider, and had to use the DMK21 and PHD for guiding since, there were only 2 stars within the reach of the prism, and neither was bright enough for the nexguide to pickup.
I also found that the focus change very slightly as the CCD reached it's lowest temperature, atleast I hope that that's why the final image is a bit soft, but on both nights the focus was slightly off after the cooler ran for about 10-15 minutes, only a slight adjustment was needed to to refocus the stars to points.
Attached images are:
image 1. The cooling setup
image 2. Uncooled 15 min dark sub
image 3. Cooled 15 min dark sub
image 4. Uncooled 30 min dark sub
image 5. Cooled 30 min dark sub
image 6. raw 10 minute with 520 gain exposure of the M104 galaxy
image 7. M104 with 6 10 min subs and 4 15 min subs stacked and stretched and 5 minutes of unguided R, G & B 1 min subs 1023 gain for color.
Notice how there is no more noise on the 30 minute cooled dark frame test compared to the 15 minute test.
Next time I'll let the DMK cool for 15 minutes before imaging, refocus, grab 10, 20 & 30 min subs as well as more color data.
I think the results are promising, what do you think?
Mariusz