A little bit struggling with my learning auto-guiding using the main OTA+cam. Well, the purpose was actually to learn how to train PEC and to fine tune the system so that autoguiding will become unnecessary for my targeted subexposure lengths at 60s.
The guiding star was an auto-selected one next to Arcturus. Near meridian.
Initially I could not get the PhD2 calibration right. And the reported dec-ra angle error went over 40deg! Playing with various parameters and setting, and the result was not satisfactory. Maybe the bad thin high clouds and unsteady atmosphere and the terrible pollution are all to blame?
When the sky got a little better, recalibration seemed to work -- the reported dec-ra axis error was still 3.3deg. Something I need to look into.
Q1. What would be deemed a good range for the error? What could be the cause of the error? (Later I measured the polar alignment error at 4' -- how much would that affect this calibration error?)
The next autoguiding test saw the guiding lines in a better shape but from time to time deviated more than 2", with RMS at .9" in both dec and ra.
Q2. Could it be due to the poor seeing? I don't know how to measure the seeing. Anyway, the sky did not look very good at the time.
In the autoguiding test, I also noticed that, during the first couple of minutes, the guiding lines were rough (for the said RMS above). But after a while the autoguiding become more steady, and most of the time slowly drifting within 1.5" and -1.5" -- I think the RMS would be much smaller but I did not have time for a new measurement session. I repeated the test and observed the same phenomenon.
Q3. Does PhD2 need to take time to adapt and stabilize the autoguiding?
I was trying PEC calibration on the mount. The initial few tests (when it was still a mess -- before the Q1 point) did not see any signs of improvement -- sometimes the result seemed to point the opposite way.
Learning from Q3, I waited for the autoguiding to become seemingly stabilized (and thanks to be better sky then), and ran the 400-sec calibration again.
With PEC correction turned on, I then spent the next few minutes (194sec) on testing PhD2's guide assistant and this is the screenshot of the result.
Click here for the full org-size image. https://photos.app.goo.gl/2YxEiLyMAfCrJm7C8
The PEC does seem helpful. Without autoguiding (until the last few seconds), the drifting in both DEC and RA was relatively steady.
The first section of the error lines seems to be significantly better in terms of smaller deviations from the trend lines.
Q4a: the main concern here is about the RA drift. It should not have happened, right?
Q4b: Does it mean that this PEC calibration worked better for the then first part of worms engaged, but was not trained well for the 2nd half? How would I better train the PEC so that it could perform consistently as in the first 1/3 of the session here (within 1" deviation from the trend lines)?