Field rotation on PME

[Paul Haese]

This is a bit of a curiosity but I just wondered what the guys with PME's have to say about this issue.

Doing 20 minute subs has led to the discovery that I am getting some field rotation in the deepest parts of the corners on my images. 15 minutes subs are fine but most 20 minute subs show some slight degree of field rotation.

I am according to the Tpoint manual set for around the refracted pole and that is around +72" for my location for the ME.

So is it better to be at the pole? I would have thought so given I am getting some field rotation. I have been doing 20 minute subs for a while; long before I sorted the issues with the pointing, but the PA must have been better.

Thoughts? Recommendations if any.

[Marke]

Just a thought Paul but if the OTA wasnt totally orthoganal to the axis
could it cause this ? I have heard of people having to shim the OTA
a tiny amount to correct for this.

Mark

[Paul Haese]

Hmmm actually I don't know. I suppose that could be a factor, but I had not considered this before. I might have to contact QSI. Given that I get good stars at 15 minutes though I just naturally assumed it was PA. Hmmm now that is something to think about.

[pluck]

Paul,

It would be useful to post an image so that we can ascertain whether it is in fact field rotation, or another issue.

Field rotation over that length of time is most definitely a result of less than perfect polar alignment. Less likely, but theoretically possible, is flexure over a long enough time, and in a direction that mimics the effects of poor polar alignment induced field rotation.

A non-orthogonal optic won't cause field rotation. It's the polar axis - pure and simple - that needs to be aligned with the pole.

Just out of curiosity, are you using Protrack?

I suspect it is you polar alignment.

Post an image or two.

Paul

[Paul Haese]

Hi Paul, No I am not using protrack.

Its not flexure as I am use and OAG, so I think that can be ruled out.

This image (http://paulhaese.net/B92.html)shows some rotation in the corners. Look closely.

This one (http://paulhaese.net/NGC6334QSI.html)too

Yet this (http://paulhaese.net/CoronaAustralisTSAQSI.html)one does not show anything.

The stars in the center are near perfect circles. Leading me to the field rotation idea.

[Paul Haese]

I have attached full size crops of the left corners from both images.

The cats paw one is most evident and was taken on the other side of the meridian to the B92 image.

[bartman]

Hi Paul, Paul and Mark,
Just out of curiosity.... what would one be looking for in the pictures ( that Paul H has posted) in regards to field rotation?

I am sorry that I have nothing to contribute to fix the problem.

I just want to learn what field rotation looks like.....I cant see any distortion on the main pics, but on the zoomed in pics, I see some just slightly egg shape stars or out of focus stars.
To be honest the aberrations are not worth mentioning.....they are beautiful pics!
I understand what field rotation is and does ( google is thy friend), but cant see it in your pics Paul....to that extent........
I guess when I get (:D) to your level I will look back at this post and cringe.....
Bartman

[Hagar]

I really wonder if your problem is field rotation or just registration problems due to field curvature. I know the TSA is well colour corrected but it dose have issues with curvature. I have also found there can be small changes in focus between subs which can make itself evident in the registration of the 3 colours and this would also show up closer in the corners during longer exposures as the stars start to bloat a little with the extended exposure time.
Overall the issues you show here aren't to obvious unless you do a Fred style zoom in.

Just a thought.

[multiweb]

It is very hard to differentiate between field curvature and field rotation in your shots. To check for minute field rotation the best way is to shoot wide field in Ha so you get tiny stars and pick a guide star right in the corner of your frame and look at the opposing corner. On a 20min subs you're bound to see any rotation for sure. Mike BJ (sculptor) has done the maths on field rotation and he has developed a technique of misaligning the scope by a couple of arc seconds from the pole in a line that is perpendicular to the DEC axis of your scope. He routinely does 3h subs with pinpoint stars in narrowband. I'll have to dig the old email he sent me at the time and post it when I find it.

[Paul Haese]

What you look for with field rotation is that each corner has stars that are curve in a radius eminating from the center. In this case I have provide just two corners on the left side but each corner has the problem. This small amount is not much but it is there and the problem is that I know it is there.





I don't think it is Doug, well not entirely. There might well be some registration problems, but I don't think these are caused by field curvature. Curvature has that look where stars are pointed toward the center of the image. These are perpendicular and curved from that line. There might be some registration issue worth looking at though.

I noticed that when I put the reducer on this scope this problem became more evident. The reducer may be causing some distortion.



I will have to give this a shot to know for certain. Like I said, I don't think it is field curvature as the stars don't really match that pattern. However, shooting longs subs in Ha should give me a definitive answer.

Food for thought guys.

[pluck]

Just to add to this one last time, you need to rule out the unknowns. With your RGB composites there is too much going on to be certain. Take 10, 20 ,30, etc minute exposures through clear (or H-a as has been suggested) in order to ascertain whether this is in fact field rotation. Longer trails should give you a better picture of what's going on. If you don't get longer trails, or if they don't appear as you would expect field rotation trails to appear , then the problem lays elsewhere.

At the very least, rule out any effects from field curvature or RGB compositing.

On another note, if you have a large and well modelled TPoint sample, use Protrack. It works extremely well for long exposure tracking by counteracting the effects of flexure over time. Tests we've performed over the years at D21 and D22 at long focal lengths have shown that (provided you have a good model) Protrack works very well indeed.

Good luck,

Paul

[Paul Haese]

Good suggestions Paul. My model is only 180 odd points but I need to do a larger model in the moon lit part of the cycle. I will try for 300 then and apply protrack.

[rat156]

Hi Paul,

One othe thing to note that unless you're guiding around a star in the centre of the field (which would be difficult with my experience of OAG) then the field rotation shouldn't be symmetrical around the centre of your field.

From my experience the field rotation is centred on the guide star.

Cheers
Stuart

[Paul Haese]

Yeah that is true. This rotation is nearly symmetrical.:question: Lums taken on the Eastern side of the meridian have this problem more than on the Western side.

[frolinmod]

In my experience 180 points all over the sky is actually more than sufficient for ProTrack to work well so long as you're using TheSkyX's supermodel feature to create your model.

[Paul Haese]

Hi Ernie. I only have the sky6 at present and don't want to pay for the upgrade to skyx.:)

[rat156]

OK, so that points to a balance problem more than anything.

As the camera/OAG will not be balanced around the axis of rotation, do you run an instrument rotator?

Could be minor flexing of the imaging train, rotationally. Try with the OAG vertical, so it's not placing any torque on the rotator/nosepiece etc.

Additionally, you should notice field rotation between subs, no matter the length, I routinely see this as my Polar Alignment is rarely exceptional (imaging stuff just a bit too heavy for the G11). This time you would have to single star align your subs, then if there's any field rotation then it will be around the alignment star. Alternatively, once the subs are registered, flick through them, the edges of the frame will rotate.

Finding the causes of this stuff can be painstaking, but you have to be barking up the right tree to find the problem.

Cheers
Stuart

[Paul Haese]

I don't think this is really the problem. I have ensured that the PME is well balanced. I am yet to find a part of the sky where the mount moves. Having said that though I will check just to be on the safe side.


The OAG is part of the camera and the camera is not that heavy. I don't have a electronic rotator but I do have the Tak one on board. Worth a try at least.
It could be I suppose, especially since I only get this infrequently, then flexure is the most likely cause, however small that might be.


Just checked that and no rotation seen. The stars stay in exactly the same position over 4 hours. It has raised another idea though.


Yes that is the truth. This might well be a focus related issue. The image that is worse has some subs near the end of the run that are less than sharp and that could be part of the problem too. It is not the spacing of the reducer but if the focus just goes out a little this type of aberation can occur from what I know.

[rat156]

I meant that the camera is not balanced. I use a MOAG and Sbig guide camera, there is considerable torque placed on the rotator by this when it's not vertical.

No rotation between subs means that your PA is fine (congrats, I find this one of the hardest things in imaging).

Could still be the Tak rotator flexing, but it's sounding less likely, focus issues are now the leading contender.

Cheers
Stuart

[Paul Haese]

Yep looking more and more like it Stuart.

Just with the camera, the QSI is pretty small. It is the size of CD is diameter and about 100mm thick I suppose. Probably one of the lightest CCDs on the market with a filter wheel. That's why I thought it was not a problem.

[gregbradley]

Interesting thread. I would have initially thought it was polar alignment.
I don't think the T-point tutorial mentions adjusting for the refracted pole?? Is that a SkyX super model feature? I adjusted mine for refracted pole but got better guiding when simply doing what the model suggested to do without adding that to it. So perhaps that little adjustment was enough to cause some rotation.

Apart from that I see there is some misregistration between coloured subs. Red is not aligned with the others in one image. What are you using to do registration?

I found that using the CCDIS plugin for CCDstack solved any registration issues I had and have never had it miss since I have been using it.

Sometimes one colour sub has larger stars than the others and deconvolution on the master RGB with the largest stars helps get rid of red or blue halos. But your image shows rotation as you point out.

If you got those stars with short exposures I would say it was a spacing issue with your flattener or the flattener did not have a large enough corrected circle (the 8300 chip is small so that should not be an issue with a Tak scope).

If you got it only in 2 corners and not the others then it may be a tilt problem.

Perhaps its a moot point anyway as I doubt 20 minute subs suit your camera unless you are shooting narrowband or at slow F ratios. I get best results from the 8300 chip with 5 minute subs otherwise the small well depth starts to cause problems and you lose star colours and bright stars can start to look ugly.

25,000 electron well depth is very low compared to most chips. However if you get good results with 20 minute subs then thats what works for your setup and try the PA without the adjustment for the refracted pole.

Nonorthoganality of the scope etc are already terms taken up in the T-point model. Right?

With regards to focus perhaps you need to check focus for each filter if not done already to see if your filters are parfocal (they often aren't). Also Roland Christen recommends refocusing with every 1C change of temperature with a refractor. I have found that varies with different refractors. The FSQ is worst and most susceptible to temp shift and focus change. 2C difference and you would find a different focus point. 3C and your images are starting to look a bit soft. The TEC180 not so affected nor the AP140 so check it for your scope.
Were these images taken near the end of an imaging run where temps could have fallen more than 1 or 2C? Odd shaped stars do clean up a bit with focus. Although I doubt out of focus slightly would cause rotationally elongated stars. Dragging cables is another possibility.

Flattener spacing errors would cause elongated stars but they would be more elongated radiating from the centre of the image not rotationally.

Is your PEC curve right? Its easy to have it upside down where your PEC is worsening the mount not improving it. That took me a few months to discover. The check east/west button in Precision PEC has to be used otherwise your resulting PE curve is simply upside down and it is fighting the mount not improving it.


Greg.

[rat156]

Hi Paul,

Forget PEC, dragging cables etc.

These will cause RA errors, resulting in elongated stars, but across the whole field and all in the same (RA) direction.

You have elongation in two directions, if it's elongation at all.

I suppose the easy way to tell is to take a shot deliberately out of focus and see the shape of the stars in the corners.

That's if you can get some clear sky...

Cheers
Stuart

[gregbradley]

Yes that's true Stuart. Unless the cable drag is pulling on the autoguider causing a rotational flex but perhaps
not as likely.

A focus test would tell you if its focus related or PA.

Greg.

[Paul Haese]

thanks guys for the responses.

Greg, the Tpoint manual suggests you shoot for the refracted pole. For my area it is +72 seconds. Since my ME is +70 I left it at that. It could well be that this difference might cause some rotation but why on some images and not on others? Perhaps it is a declination thing. Although the B92/93 images show next to no problems with the stars and it is one of the furthest dec objects I have shot since sorting pointing and guiding.

It could be a little mis registration but on the subs there is some elongation, so perhaps a contributing factor. I am using CCDstack for registration, but I have not used CCDIS. Perhaps I should?

I don't agree about the idea that doing shorter subs is better. Longer subs give you a stronger signal to overwhelm the background noise. Most technical tuts suggest getting a background ADU of 1000. Shooting for 10 minutes only gets my data to 800 with this size scope. I have not seen any significant difference in the star sizes since trying this out. The wells fill but as yet I have not seen oversaturation. The star colours always seem fine, would you not agree? This is shooting at f5.6, perhaps this whole business of shooting shorter is a myth. Someone with the maths can explain here if they like.

I think you are right about the ME. I'll shift that back down and redo a pointing run in the coming weeks.

I am not using the non orthagonality term in my model. It did not seem to make any real difference to the model.

The focus is a good point and that is what I have been doing. Every hour I check focus to ensure that everything is nice and sharp. I have not as yet done a test to see if all the filters are truly parfocal. I will put this on my to do list.

I think PE is ok, but I wanted to check the run again. There was a blip I saw when I did it and wondered what was going on there.

As for cables, I have my cable management well and truly sorted. Nothing stray, all neatly wrapped and going through the mount with a large loop like the manual suggest.

Stuart I will take a gander at that too. A slightly out of focus star image will certainly help decide what is going on.

Once again guys thanks for your suggestions. It is really nothing on the image but I want to sort this out. I am always aiming for perfection.

[gregbradley]

The CCDIS plug in is an awesome accessory and I highly recommend it. It seems as good as Registar.

The well depth issue is probably setup specific and probably aperture related. Smaller aperture may not be an issue although I think you'll see it on objects with really bright stars like Alnitak etc. But then that occurs with any chip. To take the argument to the extreme - would well depth of 5 electrons make no difference compared to a chip with well depth of 100,000? So you can see it is relevant. Once the wells are full there is no differentiation between one pixel and the next to it also full. Like an overexposure. You lose differentiation.

I may have communicated the effect poorly. It is an overexposed look on brighter stars. I have seen it on one or two of your images but as you say it is not really a problem with your setup.

I definitely have seen it many times with the TEC and CDK that I don't see with the 100,000 electron well depth 16803 chip. Usually its a couple of really bright stars that don't look so good. Deep wells is one feature of a CCD that you evaluate a chip by. Small wells means it will have trouble showing both dim and bright areas in the same image and reduce dynamic range. That is what I have noticed, not so much on nebulas etc but on brighter stars.

I am posting some questions about this on another site to find out more about the well depth and what implications it has and I'll let you know.

Its obviously not an issue for your setup so ignore that point. It may be an issue with your 12 inch RC with a reducer, so its worth being aware of it in case it shows up later.

I will check the background ADU on my subs though to make sure I get that exposure time ideal for the setup. SBIG had an exposure calculator on their site at one point. I wonder if it is still there.

I'll be interested to find out what your rotation problem ends up being caused by.

I often find not all guide stars are equal. If I see worse than normal guide errors when I start, the first thing I do is select another guide star. Often a smaller less bright sharp looking star is best. I often see guide errors plummet when I do this. Perhaps it was a simple thing like that? Another thing would be a cloud interrupting guiding for part of an exposure which is more likely with 20 min subs.

Perhaps a combination of subtle factors? Focus a bit off on some colours, worse guiding errors than normal, poor seeing so guiding is chasing the seeing more, a bit of flexure in some angles only, something loose? I remember Mike found filters were moving in his filter wheel a few years back. Filters loose and leaning can cause problems.

Greg.

[Paul Haese]

Yes it could be a combination of lots of tiny things. Most likely something that happens randomly.

For the 12" I have the STL11 though. I had already thought about the size of the stars and I wanted my stars to appear smaller. The sensor size and well depth will take care of that issue.

Yes let me know the results on your question of well depth. Always interested to hear.

[gregbradley]

Here is one response from Richard Crisp who is a professional in the CCD industry:

"Stars saturate first

Nebula signals are rarely more than 5000 electrons in my experience

Ditto for galaxies

There are multiple strategies for managing the star color issue

One is to take shorter exposures and more of them

Low read noise is key in this situation because even though 100 x 1 seconds is the same amount of signal as 1 x 100 seconds; it isn’t the same amount of noise.

Another strategy is to let the stars saturate and then replace them with a “star shot”….

I personally am not a fan of such hocus pocus but it will solve your problem."




So I guess the main thing to keep an eye on is the ADU of the brighter stars and if they start getting overexposed then back off the exposure time.

11002 chip has a well depth of around 60,000 electrons which is quite substantial.

Greg.