Log in

View Full Version here: : PHD2 Image Shifting


kkara4
11-05-2015, 08:09 PM
OK so I have an interesting problem and that is image shift while guiding. Specifically it is about 2.7 arcseconds every 3 minutes. I am able to repeat this on Eta Carinae, even over time, such that I drift approximately 9 arcseconds every 9-10 minutes. The drift is perfectly opposite to the sky rotation, ie drift opposite to RA. The way I determined this was via stellarium simulation of Eta Carinae with equatorial grid turned on. I took 2 10 minute exposures and the stars had all moved 9 arcseconds (approx). Attached is an image of said star.


This drift is most pronounced near the celestial equator. As expected, the drift is in the opposite direction to RA moving North of the celestial equator as well (but this manifests itself as an image shift the other way on the imaging camera since I don’t change its orientation). This is all while imaging West of the meridian.

Polar alignment was performed in DEC only (and using Az knob accordingly). I got it to 0 drift in 10 minutes. I did not bother with ALT adjustment, since it was fairly close (about 3 arcseconds drift after 5 minutes in RA, according to RA slope in drift alignment tool of PHD2).

Since I was specifically imaging to determine the root of this problem, I also screenshotted my PHD2 graphs (guiding with QHY5L-II + Orion ST80, 1.93 arcsec/pixel). Guidescope FL and pixel size were entered correctly into PHD2, so the graph scale should be meaningful. Here are the two screenshots (second screenshot overlaps because I kept guiding).

The scope was made a touch West heavy (Eta Carinae west of meridian). The mount is a Losmandy G11 with Gemini 2, and PEC done by previous owner. Motors are the new high torque ones, and upgraded worm that people usually do (cant remember if RA or DEC). Guide cam and mount both controlled by ASCOM. PHD2 settings can be seen in above graphs plus settings screenshots attached.


I know the graph is very choppy, but I am really not sure what to change to make it less so. I was afraid to start playing in real-time because I wanted to capture the problem and current state of my setup.

I really don’t think this is flexure (too close to following lines of RA, scalable drift to exposure length, seemingly independent of guidescope/gravity orientation, pretty significant image shift of nearly 1 arcsecond per minute, etc).

It is almost like the scope isn’t tracking RA properly (ie tracking it to fast, so the stars streak in the opposite direction), which shouldn’t happen while guiding???:shrug::question::confuse d2:

I have attached an image showing the imaging camera view with respect to the guidescope too. I then approximately superimposed an equatorial grid (red and orange lines as per stellarium simulation), and blue arrow showing the direction of image shift. Eta Carina star is just off the bottom of the view shown (cant see it in guide scope, but is centered in imaging scope), you can see the other bright stars in the Keyhole region.

Any input is appreciated, except: I am not going to post details of my imaging train/scope. The last thing I want is people telling me my FL is too large, get an OAG, blah blah. I am happy to post up extra information in terms of settings, mount details, etc :).

Cheers!

Amaranthus
11-05-2015, 09:35 PM
You might not want to hear it, but it really does look like differential flexure.

kkara4
12-05-2015, 05:26 AM
Haha that is fine Barry :). If that is what it is then that is what I will attempt to fix. I just want to make sure that it is in fact flex, and not some mount or PHD2 setting that is manifesting itself as appearing like differential flex.

otherwise i will literally be chasing my tail! :lol:

kkara4
12-05-2015, 05:14 PM
Barry, have you had flex in the past? If so how did you fix it (or at least try to fix it) and were you successful?

I have determined that the proper tube rings may be to blame with the felt lining allowing flex (even when clamped tight).

I was able to calculate about 15 microns of tilt every 10 minutes. Given the camera is hard bolted with steel and aluminium to the dovetail this is the only other point on the guidescope for movement.

Thermal expansion and subsequent tilt between camera clamp and guidescope tube rings yields << 1 micron of tilt.

I cam confident it is not the imaging train at this stage.

I still really wonder if it is a setting somehow causing the star to be continuously chased in RA :question:

Amaranthus
12-05-2015, 06:32 PM
Yes Krishan, I've had it, as documented in this thread:
http://www.iceinspace.com.au/forum/showthread.php?t=124484

Look at the image and animation I did of it - look familiar? :)

I fixed it by going to an OAG. Wonderful solution; I'd never go back to a guidescope now. Even short FL really do benefit from OAG, and for long FL, it's THE way to go. I really don't understand why many people are reticent to use OAG, it's easy, logical, and gives superb results. The only real key is to get a good guide camera (I use an ASI120MM).

kkara4
12-05-2015, 07:43 PM
Thankyou Barry. OAG I am reluctant to go to at the moment for two reasons, namely the "apparent problem" of finding a guidestar given the small pickoff prism in comparison to the imaging FOV, and the relatively slow f ratio I am imaging at at present. I have a QHY5L-II, so i dont doubt I have the camera for such an endeavour.

I fully understand why people use an OAG, and also why people get put off by flex. Thankyou for providing the link to your animation/thread, that is EXACTLY what it looks like.

I did some more thinking and initially i thought the flex could not be due to gravity. But to completely understand my problem, I took pictures of the orientation of the guidescope with respect to the ground, and it is certainly in the realm of possibility that it is due to gravity.

5 microns could be due to felt compression. I will remove it and the glue and replace with very thin electrical tape (or a thin plastic strip). will see what that yields :).

I am not ready to give up on my guiding setup yet. THere are many improvements I can make. And this is using two orion tube rings, two munson rings and 3 losmandy DVAs to bolt the thing down!

but as i said, just want to make sure it is flex and not some other gremlin :)

tlgerdes
13-05-2015, 08:04 AM
As Barry said Differential Flexure would be the culprit. Again this is the difference of flexure between the 2 camera. So it might not only be your main camera moving, but could be your guide camera/scope flexing/moving against the main scope as well.