Finding a guide star

[Ian]

I was out last night playing with guiding again.

Has anyone any tips for finding a guide star? Is it just a case of wiggling the guidescope around until an appropriate one turns up the CCD or is there a more rigorous approach?

[JohnP]

Ian - That's the way I do it.. I normally just hope that there will be one there but if not I just start adjusting my guide rings until I get one. On a few ocassions I have had to take the guide cam out & put an eyepiece back in & centre a star that way...

John

[Mike]

Yep that's pretty much it. Luckily with a cooled CCD for a guide cam you can find a star 99 times out of 100, but as long as you pick any star along the same declination it will work, even if it's nowhere near where you are imaging.

[Rocket Pooch]

I do the wiggle scope bit works for me.

[Fay]

I usually use a focal reducer, as It gives me more choice

[Mike]

However with a wider field of view the star moves less and you therefore get less accuracy with the guiding.

[Ian]

thanks for the replies, I wasnt sure that I had the best approach.

I'm now going to ponder Mikes comment about the dec. I wonder if guiding might be better if you try and get the a guidestar with as low a dec as possible...

[RobertM]

Hi Ian,

The dec would have to be relatively close otherwise you may run into diffraction effects depending on how much atmospheric the light from your guide star has to come through.  It'll get worse the further from the meridian you are guiding.

You could always increase the guider exposure time.  I use 2-5 seconds with my cmos guider at approx 500mm fl and haven't failed to find something to guide on, even if it is a hot pixel ...

[MarkS]

There's another good reason for keeping the guide star close by - field rotation.   This is not easy to explain without diagrams but I'll do my best.

Even for an equatorial mount, if your polar axis is not aligned dead accurately then you will suffer drift and also mild field rotation - up to a few tens of arcseconds per minute depending on where in the sky you are pointing (it is worst near the pole).  But when guiding, the drift component is removed leaving you with a residual field rotation.  Your guide star effectively becomes the centre of this residual field rotation since it is the object that is kept stationary.  If your imaging scope is pointed well away from this centre of rotation, then the stars being imaged will form slight trails during a long exposure.

An alt-azimuth scope is simply a very extreme case of the above effect.

[RobertM]

That's very good point well explained.

[Mike]

If your guide star and imaging area are in the same declination then field rotation will be irrelevant as they will both be rotating at the same speed.

[Ian]

hmm. Not sure field rotation isn't a red herring. need to do some sums, and I'm currently wearing socks... Getting my head round it empirically isn't working as you don't get field rotation with an Alt-Az at either the pole or the equator, but does that mean it's worse at 45deg, I wonder?

But, looking at this thread I wonder whether a guidestar at a similar altitude is best.

Of course, I could always just STFU and get on with it, but it's cloudy 

[Mike]

No because the earths axis is tilted to the celestial axis and therefore objects at different altitudes will rotate at different speeds. Whereas objects at the same declination all rotate at the same speed and your mount is tilted to the same axis.

[MarkS]

Mike,

It is not sufficient to choose a guide star with the same declination. 
Here's a diagram that illustrates what is happening.

                X
A    P     X                         G
                X

P is the celestial pole
A is where the axis of our misaligned scope is pointing
G is the guide star
The X-cluster is what we want to image.

The relationship of the guidescope, telescope and axis is always rigidly fixed.  In this example the telescope points halfway between the guidestar and the axis and so it starts off pointing directly at the X-cluster.

6 hours later the stars have rotated around the celestial pole (see diagram below).

A    P     
                 
      X
    X  X



      G

But the telescope is still pointing halfway between A & G.  So, from the telescope's point of view, the X-cluster has moved off to the right and rotated - hence the long exposure ends up trailing/rotating.  The same argument holds whereever you choose your guidestar (just place it in your chosen place in the digrams above)

[Mike]

I disagree. If they are on the same declination, not altitude, then they are rotating around the zenith at exactly the same speed. If your guide star is on the same altitude then it will naturally rotate as we are at latitude 53 degrees and as such different areas along the same altitude cross different declinations.

[Rick]

...but if they differ in RA by 90 degrees then they will also be moving within their respective fields of view in directions 90 degrees apart.

[Mike]

..at exactly the same rotational speed. So as long as your guide star is rotating at the same speed as the area you are imaging and can send correctional pulses to your mount to correct for any deficiencies in the mounts speed of rotation, then, along with corections for RA you will keep your area to be imaged steady.

[Rick]

That's true for the main-axis rotation, which is the largest single effect, but it's not true for the field rotation, which is a much smaller effect. I'd be interested to know just how off-axis your alignment has to be before the field rotation effect becomes significant (for a given exposure length).

[RobertM]

That relies on both target and guide star having the same center of rotation as seen by the OTA i.e. the polar axis is exactly aligned.  Any differences in alignment will be exagerated by focal length (longer being worse) and atmospheric refraction depending on how far they are from the meridian.  If there was no atmospheric refraction and the polar axis was exactly aligned then you could guide on any star, assuming it was visible, as it would have exactly the same angular movement across the sky as the target.

[Ian]

after a bit more thought I see what Mark is getting at. Just picking a guide star on the same Dec isn't good enough if it's RA is massively different from that of the target. The question to answer is, where is the centre of rotation about which the field is rotating? Without any guiding correction, it's based solely on the alignment of the scope. However, if you're guiding, then the mount will be adjusted, as expected and this will shift the centre of rotation to the guide star. However, is this better or worse or indifferent to where it is without any guiding. I'm going to try and understand that bit.

This website has a quite in depth analysis. I'm in the process of plugging some of our numbers in to see if I can work out how far away a guide star can be before images trail...

[Rick]

...so I guess, in reality, it comes down to "The nearer your guide star is to your target, the better."

[Ian]

or does it come down to, how far away can my guide star be and get away with it?

[Rick]

There's a spreadsheet linked on http://www.fvastro.org/presentations/FieldRotation.htm which is claimed to calculate something relevant...

[Rocket Pooch]

Wow

[MarkS]

I'm going to come up with my own estimate for the magnitude of this effect - i.e. the trailing caused by having a distant guidestar when the scope is slightly misaligned.   It interests me greatly because of the difficulty I have in finding a guidestar bright enough for my SPC900.

[Mike]

I understand what Mark is saying here but I think it would only imply if you had a seriusly nasty bit of polar alignment and were trying to guide on a star a long way away from the imaging area.

In the real world your Polar alignment is going to be at least half decent and the Kochab's Clock method I use is good enough. Also, you wouldn't have to move too far away from your imaging area to find a decent guide star unless you were doing something like using a non-long exposure camera on a small aperture guide scope.

Image scale is obviously important as well as focal length but rotation is going to be a negligable effect as far as I can see unles someone can prove otherwise.

[JohnP]

Mark - I'd be interested in your findings as well. I always seem to suffer from small movement between subs even though I am fairly sure my polar alignment & guiding is good. I thinking now maybe it is rotational shift due to mis-alignment between guidescope & imaging scope??? They do point fairly close to one another but I do tend to 'wiggle' guide scope around until I find a suitable star. It is my understanding that if polar alignment is excellent & guiding is working well then subsequent subs in a sequence of shots should be on top of each other - I always tend to get drift of one or two pixels in same direction which distorts star images (elongates)..

John

[Mike]

Would you not get that anyway inbetween subs when it is not guiding? Do you have it set to have jitter on? Having the subs NOT line up is actually a good thing as it avoids perfectly square stars and 'jitters' the image slightly, which is actually as setting you can turn on in MaximDL. I always have jitter turned on to smooth out images and give an antialiasing effect.

[JohnP]

Yes - but I am guiding so I don't understand why stars don't stay in exact same position - also stars are slightly elongated....

[MarkS]

I agree with Mike - a small amount of drift is definitely a good thing.  It prevents square stars and it alleviates problems caused by hot-pixels because sigma stacking can then remove all the hotties  that remain after dark subtraction.  But you don't want drift to be large enough to become apparent on a single sub.

Last night, I did work out an estimate of the the size of the guide star problem. Any individual case is very dependent on the geometry of the situation i.e. in which direction the misalignment of the polar axis occurs and in which direction the guide star is away form the imagaing area.  But I have derived a pretty accurate upper bound.

For the Canon at prime focus of the C11 (focal length = 2800mm i.e. 0.6arcsec/pixel) and with a polar misalignment of 0.2deg then for a 5 min exposure, a maximum 1 pixel star trail will result if the guide star is 5% of the distance of the the area of being imaged from the pole.

Example:

Cluster being imaged is at declination 70deg.  This is 20deg from the pole, so my guide star must be within 1deg (=5% of 20deg) of the cluster to prevent trails longer than 1 pixel on a 5min exposure (Canon on C11).

If you halve  (or double) the focal length you halve  (or double) the trail length.
If you halve (or double) the exposure time you halve (or double) the trail length.
If you halve (or double) the polar misalignment you halve (or double) the trail length.

This approximation works very well acros the whole sky up to a declination of 85deg .  I'll show you the maths at some point.

So for a typical scope (1000mm) the whole issue is a red herring - use Kochab's clock to align your axis and keep your guide star within a few degrees of the imaging area and you'll be fine.  As long as you aren't imaging near the pole or using extreme exposure times.

I think flexure in the mount, the side by side plate or the guide rings is generally more likely to be a  problem.

[JohnP]

Phew! - Good work Mark - I can cross rotation of possible reasons for the problems I am seeing. I'd be interested in seeing the maths at some time.

John

[Mike]

John have you checked your software is not purposely moving the image inbetween subs or that there is a delay in the guide pulses inbetween subs that would allow some small drift?