Back from Oregon Star Party, with some more image data to process. This year saw a big improvement in conditions, thanks to an earlier date intended to avoid smoke from wildfires that proved to be such a problem last year. This strategy was successful, and we had only a touch of smoke to contend with, but had the additional consequence of reduced hours of darkness. Actual available imaging time started around 10 pm, and ended shortly after 4 am. A few of my sub-exposures into the twilight hours after 4 am are useable, but show a significantly washed out brightness from the advancing sun, which I was able to deal with using normalization options in PixInsight’s integration tool.

This is also my second outing with the camera I acquired last year: an SBIG STL-11000M including the optional 8-position FW8-STL filter wheel and installed with a set of 50mm diameter Astrodon Gen II Series I LRGB filters. As seen in the image of the setup, the camera and filter wheel are physically very big and have to be carefully positioned to avoid hitting the mount during slews.

With last year’s star party bringing such challenging conditions and not yielding very much data to work with, I’m still learning how to deal with this new setup. After a week of collecting and processing new data, I’m confident now that I have a good handle on the issues, and there were a few things to deal with.

Firstly, the cooling of the camera: not sure if this is something that has deteriorated due to the age of the camera (manufactured around 2006) or just a limited capability, but the thermo-electric cooler for the image sensor seems to really struggle to achieve low temperatures. The white bucket to the left of the mount is the cooling water system; a submerged pump in the bucket circulates cold water through the heat exchanger integrated in the camera. Even with this installed, I can only hit -10 ° C reliably, and at night with ambient at 55 ° F, the cooler is still operating at 60 – 70% power just to hit that temperature. Ideally I’d like to be able to run at -20 ° C, but that doesn’t appear to be practical.

The second issue to deal with is the general crankiness of the camera; the driver appears to be quite buggy and when provoked in the wrong way has the effect of immediately crashing the PC on which it is running and triggering a reboot. For the first few days of the star party, I had multiple such crashes each night, but by gradually learning how to not push the camera in the wrong way, was able to survive the last two nights crash-free (for some reason, adjusting anything in the autoguider settings dialog in TheSkyX seems almost guaranteed to provoke a crash). Along the way, I was able to learn how to mitigate the effects of a crash and be up and running again with the minimal delay. A key learning: following a PC crash, the mount just keeps running, and knows where it is, just seeing the PC as having disconnected temporarily. Once the PC is rebooted, there is no need to reset the mount to a park position and unpark. Just reconnect, and the mount will resync the driver (and TheSkyX) to the correct position.

Thirdly, the image scale is quite a challenge. With the FSQ-106 focal length of 530 mm and the STL-11000M pixel size of 9 um, the image scale is at around 3.5 arcseconds per pixel, resulting in a substantial under-sampling of the sky. The resulting images show fainter stars in particular to be extremely “blocky” and look almost like uncorrected hot pixels. Following a suggestion from Wade Hilmo, I was able to successfully mitigate this using the drizzle integration tool in PixInsight. This took a little while to figure out (it’s not at all intuitive – I’ll write a post about that later) but once I had that included in the process workflow, the problem seems to be well resolved.

Lastly, the raw data from the camera is just remarkably messy, and requires careful calibration to clean up before further processing. Some of this is due to the nature of the image sensor chip; the interline sensors such as the KAI-11002 seem to have a lot of junk in the image (showing up primarily in bias frames) compared to the full-frame KAF series sensors from the same manufacturer. In addition, though, this is a really old sensor. CCD sensors are well-known to “age”, acquiring new defects such as bad columns and hot pixels over time, believed to be the result of damage caused by high energy particles impacting the chip, and this particular example has some interesting and unusual patches of defects. Proper calibration removes most of this, and PixInsight’s CosmeticCorrection tool can clean up whatever the calibration doesn’t catch.

As for the image here: NGC7822 is a very large emission nebula in the constellation Cepheus, about 3,000 light years away and about 100 light years across. The image shows an area of sky roughly the size of a golf ball held at arm’s length.

The image is a combination of a total of just over 6 hours of exposure time, captured between July 30th and August 1st 2019. All processing is in PixInsight. The processing workflow uses drizzle integration (the image shown here is reduced size) and deconvolution to improve the sharpness.

Details: