Changing background values on STi images. Causing guiding problems and star fades.

Discussion in 'STF Series' started by JoshuaHufford, Apr 22, 2017.

  1. Doug

    Doug Staff Member

    Joined:
    Sep 25, 2014
    Messages:
    6,241
    That would be very helpful. We've never been able to reproduce it here. Hard to fix something that never happens.

    MaxIm DL is supposed to ignore bad frames while guiding. They can happen for any number of reasons, including momentary clouds drifting by. Multistar guiding wasn't coasting through bad frames, but that is working properly now.
     
  2. jerryyyyy

    jerryyyyy Cyanogen Customer

    Joined:
    Oct 10, 2014
    Messages:
    502
    Location:
    Stanford, California
    Sorry for the confusion on the versions Colin.... I am running 6.17... is that the latest? I was watching this again last night and these come up every few minutes. Sky was completely clear. They are one frame. They are not rare. Camera was just in the shop and was checked out.

    The guiding with multi-star is good/better than single star for me given this. for some reason the single star interprets these frames as big offsets and corrects then only to find the next frame OK.

    I have the Mach1 correcting very small changes and will be backing off as I seem to be over-correcting now. My minimal changes is 0.04 sec n DEC/Y and 0.02 sec on RA/X... which at 9.7 arc"/sec are 0.38" and 0.19" respectively. I use aggressivity of 7 and 4 respectively.... About as good as I can do.... the stars still tend to be eccentric a bit in the vertical (Y DEC) axis, but I don't seem to be able to correct that out. The corrected PE is < 2" over 6m. In the plots above you will see where Maxim tries to squash down or up the PE in green. The guider pixel scale is 1.54 arc"/pixel so those graphs in pixels represent 1.54 arc". Probably would be an interesting exercise to turn the aggressivity to zero (uncorrected) and see what it looks like. I have never done that.

    Tracking varies with the seeing, which is not good where I live. So hard to fine tune.
     
  3. Owen Lawrence

    Owen Lawrence Retired Staff

    Joined:
    Oct 1, 2014
    Messages:
    1,397
    Please use Help | Check For Updates... It will show you that the latest version is not 6.17.

    - Owen -
     
  4. jerryyyyy

    jerryyyyy Cyanogen Customer

    Joined:
    Oct 10, 2014
    Messages:
    502
    Location:
    Stanford, California
    OK, I have it. It is not going to screw up the multi-star that is working now if I install it? I have a new license.
     
  5. Doug

    Doug Staff Member

    Joined:
    Sep 25, 2014
    Messages:
    6,241
    Won't affect it. We didn't do any work on guiding for this release.
     
  6. jerryyyyy

    jerryyyyy Cyanogen Customer

    Joined:
    Oct 10, 2014
    Messages:
    502
    Location:
    Stanford, California
    Great. So I get this error now, twice. I cannot guide. Thanks for the update!

    (exposure complete)
    (exposure complete and image downloaded)
    Image finished
    Plate-solve pointing image.
    449 image stars found
    490 catalog stars found
    Solved! 213 stars matched.
    Average residual is 0.34 arcsec.
    Pointing error is 0.635 arcmin @ angle 251.09
    True focal length is 1421.6 mm.
    True binned plate scales (arcsec/pix): H = 3.13 V = 3.13
    True image center (J2000): 18h 18m 47.5s -13° 47' 38.34"
    Imager sky position angle is 0.0 deg.
    [sync] telescope synced
    Within max error tolerance, no re-slew needed.
    Target is now centered.
    (aggregated exposure times, trying to autoguide)
    Switching from Luminance to O-III filter for imaging
    (take guider image at initial interval 5 sec)
    Multi-star guiding, MaxIm chose star at X=313.3249 Y=230.7056
    (guide star SNR=14.0 X=313 Y=230)
    (first exposure estimate 5 sec)
    (guide star SNR=14.0 X=314 Y=230)
    (final exposure for guiding 5 sec)
    (guide star SNR=14.8 X=313 Y=230)
    **Script Error (Tracking has been stopped)**
    Source: MaxIm DL 6
    Message: Bad input parameter
    Traceback:
    Failed in HandleAutoguiding() at line 1767 column 5
    called from Main() at line 1246 column 13
    Script Context:
    1764: End If
    1765: End If
    1766:
    => 1767: If Not SUP.AutoGuide(True) Then
    ^
    1768: If IMG.ExposureInterval > Prefs.AutoGuiding.MaxUnguidedExposureInterval Then
    1769: If STRICT_AUTOGUIDING And Not PLN.BestEfforts Then
    1770: IMG.Status = IMG_FAILED
     
  7. Owen Lawrence

    Owen Lawrence Retired Staff

    Joined:
    Oct 1, 2014
    Messages:
    1,397
    Jerry, can you please supply your MaxIm DL log file? The "Bad input parameter" is only produced by document processing functions such as SaveFile, Crop, various image processing filters, etc., but I can't tell what's going on. The info you posted is going to require some feedback from Bob Denny.

    - Owen -
     
  8. jerryyyyy

    jerryyyyy Cyanogen Customer

    Joined:
    Oct 10, 2014
    Messages:
    502
    Location:
    Stanford, California
    Bob kinda gave up on this. I got the same message with the previous version, then it started working for whatever reason.... let me just do a couple reboots and see.

    It ran find in single star the rest of the night. I was too tired to look for the bad guider frames but will look in the tracking log and see if they show up. The PEC on this mount is tuned to < 2 arc" so it should not be hard to guide.
     
  9. jerryyyyy

    jerryyyyy Cyanogen Customer

    Joined:
    Oct 10, 2014
    Messages:
    502
    Location:
    Stanford, California
    Bump, I did not see any bad guiding on a bad frame in single-start so maybe this is fixed too. Will report back if multi-star works tonight. BTW, we have confirmed that the small eccentricity is in DEC of all places and not RA.
     
  10. jerryyyyy

    jerryyyyy Cyanogen Customer

    Joined:
    Oct 10, 2014
    Messages:
    502
    Location:
    Stanford, California
    Just to follow-up on this with more data, I am not getting the weird fluctuations in guiding with the bad frames in either single or multi-star in the latest iterations of Maxim. I pray this is solved.
     
  11. JoshuaHufford

    JoshuaHufford Cyanogen Customer

    Joined:
    Oct 10, 2014
    Messages:
    330
    That is good to hear, I'm looking forward to trying 6.18. Clouds and bad weather keep preventing me from trying it, has been quite a while since we have had a clear night.
     
  12. jerryyyyy

    jerryyyyy Cyanogen Customer

    Joined:
    Oct 10, 2014
    Messages:
    502
    Location:
    Stanford, California
    It is running fine tonight. Sky a bit clearer. I cranked up the DEC correction a bit as the stars were a touch longer in that direction. Will see what happens. They are OK but I think I can fine tune this and it is an interesting exercise.

    Please see the PE in the image. Green is RA and Blue is DEC. The little Xs are corrections. You can see the up and down of the PE.... I think I can crank that up also but want to fix the DEC first as that is the barely noticeable error.

    Periodic Error.jpg

    Cool.
     
  13. jolo

    jolo Cyanogen Customer

    Joined:
    May 26, 2015
    Messages:
    11
    I'm using ST2000XM camera but I expect that your DriverChecker looks similiar. Check if you have an option to uncheck "Auto Bias Level Correction".


    scrn1.jpg
     
  14. brew

    brew Standard User

    Joined:
    Aug 1, 2015
    Messages:
    8
    Joshua, is this fixed in 6.18 now for single star guiding? I am at 6.08, and of course don't want to buy the update unless I know it has the fix.

    Our weather has turned cool again, but I definitely am still seeing the fluctuations in image background coming from the ST-i. So, maybe the ambient temperature is not the issue.
     
  15. JoshuaHufford

    JoshuaHufford Cyanogen Customer

    Joined:
    Oct 10, 2014
    Messages:
    330
    Sadly I have not been able to test it. I had a mouse get in my conduit that runs all of my power wires (not the USB) and chew through the insulation and short some things out. It is going to require running a new conduit and power wires from the control room to my pier. I also discovered my RA worm gear is loose on my mount, so for right now I'm dead in the water. I'm hoping to spend a few days at the observatory to get everything back up and running soon but life keeps getting in the way including a family member in the hospital for 4 days.
     
  16. Colin Haig

    Colin Haig Staff Member

    Joined:
    Oct 27, 2014
    Messages:
    2,473
    Sorry to hear all this, hopefully things get sorted ok.
     
  17. jerryyyyy

    jerryyyyy Cyanogen Customer

    Joined:
    Oct 10, 2014
    Messages:
    502
    Location:
    Stanford, California
    Hey guys, bet you are happy not to hear from me :)... My system has had a wonderful USB adventure too. The roof stopped closing... turned out the USB/serial port cable developed a short that clobbered by cloud sensor that used the same connection... try to debug that sometime... took a week and I hope it is fixed. Has worked 2 nights now... fingers crossed... of course we have smoke here...

    On the guiding, we use the 8300M and at the end of the day have stuck with the multistar as implemented in 6.18. I have gone back and forth between single and multi and in our long FL scope (1400mm) we get better results with the multi. But both work automatically with ACP Expert Scheduler.

    I am using full calibration with dark and bias of the guider image.

    In the process of debugging the USB problem I fully brought on line my backup Windows 7 mini-desktop (and got the same error). While doing this I installed all the SBIG and Diffraction Unlimited software and it worked without a problem.

    I have to say that if I had more input into the process I think there could be improvements in the guiding algorithm (maybe)... what I see when I actually look at the single star is the scintillations of the atmosphere which really drive the corrections. We have learned with the help of Astrophysics to back off trying to correct everything and get a good estimate of the variance from the DEC (this should not change much if you have good PA). Then do not try to change the RA more than the variance seen in the DEC. This boils down to the settings of the minimal move, where we are pretty much using 0.02 or 0.03 in our system. I would love to see how the variance is calculated for the multistar since you would think it would be a lot better than the single star... but it is only minimally so... (I do a lot of stats in my day job).

    BTW, I also like the plotting in 6.18 of the guide star especially the X-Y scatter plot and the separate X-Y axis plots (separate the guiding problem into two problems). As you can see from the X-Y plots above that PE comes out. Ours is 1 arc" so you do not want to overcorrect. Astrophysics says to calm down and realize the we need to stack images as we cannot correct better than our seeing, which is 3 arc".... it would be nice to fully understand the plots so that I could put up a nice set of X-Y labels that relate directly to my guide chip... I think this is what is being used in the graphs... If you knew what your seeing was you could put some red lines up at that level and not correct errors that fall between the red lines...

    Lastly, we are going to do some work on the long FL scope to try to reduce mirror tilt. Thus I will be back on my old Tak 180ED (to icon photo)... I will be re-colimating this and then guiding again so all the settings will get redone. This I am sure will be an adventure. But, it should be easier to guide with a 500mm FL scope than a 1400mm FL scope. We will see.

    Best wishes for Thanksgiving and the Holidays,

    JY
     
  18. Owen Lawrence

    Owen Lawrence Retired Staff

    Joined:
    Oct 1, 2014
    Messages:
    1,397
    Thanks for the report. Just a word of advice before you switch to your Tak: Use File | Configurations to save your current configuration before making any changes to any setting. Proceed with changes, and when you're happy with them, use File | Configurations again. Think of these as "restore points".

    Both times, choose a name that reflects your current hardware setup. That way when you switch back, it's a simple button press to restore all of your settings, even the ones you forgot you needed.

    If you use different computers or different accounts, do it anyway. It might help down the road if you end up needing to make comparisons.

    - Owen -
     
  19. jerryyyyy

    jerryyyyy Cyanogen Customer

    Joined:
    Oct 10, 2014
    Messages:
    502
    Location:
    Stanford, California
    Thanks for the reminder Owen, I just did this and saw some older configurations including back to when I started the system with Expert Scheduler (there is one with Bob Denny's name on it). These older ones are for the Tak and I should be able to go back to them. Only had one with the long FL, so now I have a good one locked away.

    Seems if I go to the new computer I should be able to move these configurations?

    BTW, the guide star images are rather interesting if you look at them while guiding (I have a laptop next to the TV at night and can watch the system out of the corner of my eye). If I make the track box small on single star (16x16) and use 2X binning even with the 1400mm FL the star looks very solid with good sky conditions, but then you will see a scintillation where you can see the actual pixels change (there are only 16x16 or 256) so there is a jump, but in the next camera cycle 5s later, it may have come back to where it was in the previous image. I should match this on the multi-star and see if it does the same thing... I think it is attenuated with multi-star. I used to do a long of graphics programming in OpenGL and italmost looks like the type of phenomena that could benefit from anti-aliasing. There is also probably some interaction between your algorithm and the amount of binning in the guide image. George at Astrophysics said to use 2X given our long FL. Anti-aliasing would interacting with the binning.... of course what you graph in the image may not relate directly to the computations you are making.

    Cheers,

    JY
     
  20. Owen Lawrence

    Owen Lawrence Retired Staff

    Joined:
    Oct 1, 2014
    Messages:
    1,397
    Yes, you can copy your entire Documents\MaxIm DL 6\Settings folder to another machine. Just accept that there are many paths saved in the configuration files, so if you change account names, disk drive letter, or the like, some settings may no longer be correct on the new machine.

    Anti-aliasing might make pretty pictures, but it will do no good for autoguiding. You're sort of anti-aliasing anyway, in a virtual way, by guiding on the centroid of multiple pixel values. We're just not drawing the "modeled" star back into the image the way an anti-aliasing algorithm might draw calculated pixel values.

    - Owen -
     

Share This Page