There were a couple of days good weather over the past weekend after month-long raining here in Sydney. So I took a trip to a dark site near Orange to test a 12 inch Newtonian I rebuilt using Skywatcher original. I rebuilt two 12 inch Skywatcher scopes, a 300F4 and 1 300F5, replacing original metal tubes with aluminium bars, saving some 4 kg each, so than they could be mounted on my CEM60. There had been collimation and auto-focusing issues for the 300F4 for some time, resulting in spiky stars. These turned out to be caused by plays in the focuser itself. Star shapes had never been a problem when I first rebuilt it using ASI294mm. But after switching to ASI2600mm problems persisted. I had to refine every structural aspects as well as ways mounting focuser, but the final fix was in fact tightening screws on the focuser itself, indicating main cause of the problems. However, I was able to improve structural sturdiness of the 300F5 using ways I came about fixing the 300F4. Stars on the 300F5 were always good except guiding errors used to be twice as much as with 300F4. After reinforcing focuser mounting parts I was able to get similar guiding accuracies with the two scopes, indicating flexures on parts holding the focuser on 300F5 were fixed.
While refining the 300F4 I took quite a bit of data for M83, albeit with overblown stars. Focusing changes due to temporature drops were far more rapid than the F5 scope so I thought I might just use data from 300F4 as RGB, and take some L using 300F5. Here are some pictures of these. LRGB data for 300F4 were 76, 14, 18 and 19 x 5min, respectively. 36x5min of Ha using 3nm filter were added, which was taken in my Sydney home with B8 skies. I've done over 140x 5min L using 300F5, but due to seeing variations and focusing changes in-between auto-focusing 64 best subs were giving me best clarities and noise levels. Seeing condition changes were also noticed by a friend of mine using AM3 who came with me for the trip. Guiding RMS errors were as good as 0.3 to 0.4 arcsec for each axis but can go to as high as 1 for some periods. RGB data were a bit low for this image as L data with 300F4 were pretty much overridden by 300F5 data, so it's hard to tell a total image time for the final image. I also had to remove colour fringes due to large and irregular star sized due to collimation errors with 300F4 while testing and refining it. Best might be to reddo RGB as well but I might have to it for now.
