M109 and relatives--amidst a strange problem

[Rodd]

This is a 45 min (90 30 sec subs) red stack.  Any longer than 30 seconds and the histogram of the unstretched subs separates from the left hand margin--my sky is too bright.  But collecting lots of short subs does have its advantages--it yields very clean data and (for the ASI 1600 FWC) improves star color.  This stack will eventually consist of 400-600 subs.

But all is not roses.  I am experiencing a strange problem that no one, not the gurus at TAK America, not the clever folks at Pixinsight, and no one from SGL (so far).  Here is the 2 minute mystery (Haledjian, or Holmes, or Dupin....pick your hero):

All is well until I do a meridian flip.  The goto spot centers the focus star after the flip, so the scope is aligned on the mount.  Its not I go to align the subs for the evening.  After aligning the subs to a single sub, I place all aligned subs into a tool called blink, which allows a rapid scan of all subs.  When I zoom in, I notice that there is a distinct shift--the stars do not line up between the east subs and the west subs.  There is a distinct, frame wide shift.  This makes no sense.  registration is supposed to work for subs of various scales, various angles, various framing.  I have registered many subs that barely align, and the registration process perfectly aligns the stars that are in all frames--there will be a edge crop--often large--but the stars will be aligned.

When the subs are blinked and the the tool paused at the last eastern sub, then advanced to the first western sub, there is an obvious shift--but it almost seems like there is an upper layer that is shifting, while the bottom layer remains constant.  If I integrate all subs--there is a ghost edge to many stars--representing the shifted stars from the shifted subs.  Very strange.  large stars shift completely--small stars only shift a tiny bit.  The strangest thing is when I registrar all east subs to an east sub, and all western subs to a west sub, then integrate two stacks (an east and a west), then combine the stacks--all stars align.   

This method of aligning all east and west is gruesome, as I will have numerous nights per filter, hence many alignment subs and stacks to keep track of.  This is intolerable--mainly because it is new, and restricted to the FSQ 106.  The scope came back from Japan far from pristine--better, the stars are better-but hard to accept as they still are not really right.  This new issue may suggest something amiss with the scope.  I have viewed distortion maps of all subs and the distortion maps for the east subs are very different from the west subs.  But all east subs are similar, and all west subs are similar.  

I do not have flexure in my system as everything is threaded and buttoned down hard-unless its the focuser--which is, indeed, a problem with the scope.  Anyone have any ideas?

FSQ 106 with .6x reducer and ASI 1600  90 30 sec red subs--lots of fuzzies.  If I had known, I would have pushed teh framing to the left, so the galaxy along the left hand margin would not be so close to the edge.

 

[ollypenrice]

What happens if you stack just the subs from just one side of the meridian? What are the stars like? I would try this for both sides, ending up with two stacks, but maybe you've done this.

On the physical side, the obvious source of variance between meridian sides would be sag in the focuser, which would invert relative to the sky. It there is a slight distortion which varies each side of the flip, then the software's calculation of the stellar centroids may be thrown out of kilter, prducing bad registration.

If all the subs on one side stack nicely, and all the subs on the other side stack nicely (using a sigma routine) then I wouldn't hesitate to align them in Registar and average combine the two stacks.

Olly

 

[Rodd]

17 minutes ago, ollypenrice said:

What happens if you stack just the subs from just one side of the meridian? What are the stars like? I would try this for both sides, ending up with two stacks, but maybe you've done this.

On the physical side, the obvious source of variance between meridian sides would be sag in the focuser, which would invert relative to the sky. It there is a slight distortion which varies each side of the flip, then the software's calculation of the stellar centroids may be thrown out of kilter, prducing bad registration.

If all the subs on one side stack nicely, and all the subs on the other side stack nicely (using a sigma routine) then I wouldn't hesitate to align them in Registar and average combine the two stacks.

Olly

 

Thanks Olly. Yes, I do this. The problem is, the two stacks from opposing sides do register. However, the stars are bigger. There is no ghost fringe, but the stars are that much larger.  I am not 100% sure if this. But in several attempts this has been the case.  However, conditions were different, so guiding and seeing entered the equation.  But even if it works, this means a great deal of extra work.  I never saw this before the return from Japan, and I don’t see it in any other scope either same camera. If it is the focuser, as I said, that means it’s the scope and I am inclined to not accept it—not after what I have been through. I am negotiating a buy back cost with TNR.  I think they should replace the scope.  But they will buy it back and if I want to get another one it will cost me another $2,500. So I am leaning toward a full frame camera and chroma filters and the new TOA flattener (1um spot size on axis, 2um spot size at 30 mm and 3 um spot size at 60 mm.  I will lose my wide field, but full frame back illuminated 16 bit camera on TOA with new flattener is pretty spectacular. I can always get a redcat 51 or something for real widefield for low cost if I want to.   This will be the only way I will ever be able to get a 6200 and 50 mm filters. Otherwise, it’s the 2600. The problem I have is giving up a lifelong and even inheritable scope for a camera that will last maybe 10 years. Maybe 15.  I guess I will still have the filters, and they cost more than the camera.  I am conflicted

Edited November 19, 2022 by Rodd

[simmo39]

Glad somebody could help, I have no idea about these kind of problems i just like to look at the pics. Yep im a philistine! LOL

[Rodd]

20 hours ago, ollypenrice said:

What happens if you stack just the subs from just one side of the meridian? What are the stars like? I would try this for both sides, ending up with two stacks, but maybe you've done this.

On the physical side, the obvious source of variance between meridian sides would be sag in the focuser, which would invert relative to the sky. It there is a slight distortion which varies each side of the flip, then the software's calculation of the stellar centroids may be thrown out of kilter, prducing bad registration.

If all the subs on one side stack nicely, and all the subs on the other side stack nicely (using a sigma routine) then I wouldn't hesitate to align them in Registar and average combine the two stacks.

Olly

 

Also, Olly, if there is a shift, technically the flats won’t work and I will have to have a different set for East and west

[ollypenrice]

14 minutes ago, Rodd said:

Also, Olly, if there is a shift, technically the flats won’t work and I will have to have a different set for East and west

I doubt that it would be significant, though.

Olly

[Rodd]

21 minutes ago, ollypenrice said:

I doubt that it would be significant, though.

Olly

I don't know--in the past the slightest rotation or shift causes flats to fail--dust bunnies partially corrected--leaving arcs.  Do you see this in your FSQs?  If this is normal behavior I fail to see the attraction