How accurate do you polar align?

[frugal]

The weather has finally cleared enough that I can actually use the scope the I bought in the FLO clearance sale a while ago. I spent a good while setting up the scope in the light. I collimated the polar scope and when it got dark I polar aligned as accurately as I could.

When I performed a two star alignment it reported a polar alignment error of about 1 arc minute in either axis. Seeking across to targets got them on screen first time every time.

My question is: Is a polar error of 1 arc minute considered accurate?

About half of the 150 second subs had nice round stars, however the other half had eggy stars (but I think that this had more to do with the balance being slightly camera end heavy)

Sent from my SM-P900 using Tapatalk

[Leveye]

Always, always,always balance and align to the very very best you possibly can. 

Edited May 18, 2014 by Leveye

[oldpink]

when imaging the better your PA and balance are critical to getting the best out of your images

[ollypenrice]

Always, always,always balance and align to the very very best you possibly can. 

Well, I don't necessarily agree. If running unguided then, yes, the best possible alignment is best. But if running guided, a slight misalignment allows you to run the autoguide Dec corrections in one direction only and this can remove the problem of Dec oscillation. I have to do this on one of our mounts routinely.

As for balance, like many people I like to run just a little heavy on the east to keep the backlash down. In perfect balalnce the payload can float around in the backlash.

Frugal, if your stars are sometimes round then the problem is not PA. That would affect every single sub in the same way. It usually shows as a rotation in the corners.

Olly

http://ollypenrice.smugmug.com/Other/Best-of-Les-Granges/22435624_WLMPTM#!i=2266922474&k=Sc3kgzc

Edited May 19, 2014 by ollypenrice

[Dave]

I tend to agree with Olly.
It's been ages since I checked my polar alignment, and my auto-guided images are still coming out OK.
My mantra: If it's working, don't fiddle.

[frugal]

Thanks for the responses, however I am still not sure how accurate I should be getting the alignment in the first place. I know I should be getting the alignment as good as possible, however I do not know how close I am to the target accuracy.

I managed a 1 arc minute deviation in Alt and Az. Is that considered accurate? Should I be trying to get it down to only a couple of arc seconds? Or is anything under a degree good enough?

After a star alignment the mount will give me a deviation from alignment, but I have no idea if that deviation is acceptable or if I need to go through the process again.

Sent from my SM-N9005 using Tapatalk

[davew]

As no one has specifically answered I'll put neck on the block and say 1 arc min is fine. If I could walk outside each night and be sure to align that accurately I'd be a very happy bunny. In fact, I'll put my neck on the block twice and say you'll be more accurate than some who drift align

Dave.

[wxsatuser]

Don't know how accurate mine is but I get nice stars with a 3 minute DARV.

First image is star near the meridian 3minutes each way, bang on.

Second image star to the East,  3minutes each way, slightly off.

I fixed the East one with the PHD drift method as it's quicker.

[PhotoGav]

An interesting question indeed! I have normally managed to get my PA to just under 1 arcminute accuracy as reported in Alignmaster and have had great guiding results. I set up on Saturday night and ran Alignmaster quite a few times, but just couldn't get a, what I thought was, good PA. The accuracy was around 8 or 9 arcminutes I think. I finally decided to stop the futile fiddling and started imaging. Guiding was spot on, stars are round and all was fine for two nights in a row (wow I hear you say, two consecutive nights imaging... that's a rarity!). So, my conclusion is that, when guiding, you can get away with a pretty large inaccuracy.

When unguided, it's a whole different matter though... The closer to perfection the longer the subs you will manage. My unguided test shots over the last two nights were pretty shocking! I did a 60s sub of M27 and the trails are horrendous!

[jambouk]

It depends how accurately you eant/need it.

For visual, then 30 arc minutes is probably good enough; i've had it much worse.

For unguided imaging as accurate as possible. A friend and i did the polar align routine on the skywatcher handset the other night and managed to get one down to 45 arc seconds and the other down (alledegely) to 2 arc seconds. We were only taking 60 second subs at a focal length of 1200ml, but it seems ok, and we were quite poorly balanced in RA as he needs another counter weight.

I had struggled with the PA routine on the handset before, but now i've reduced my cone error and tightened up the back lash, i actually find it works like a treat with a reticle eye piece. Very happy with it.

Jd

[sharkmelley]

But if running guided, a slight misalignment allows you to run the autoguide Dec corrections in one direction only and this can remove the problem of Dec oscillation. I have to do this on one of our mounts routinely.

100% agree.  I recently began to adopt this technique and it makes a huge difference to my guiding accuarcy in Declination.

Frugal, if your stars are sometimes round then the problem is not PA. That would affect every single sub in the same way. It usually shows as a rotation in the corners.

It may not be quite that simple.  Although bad polar alignment certainly does manifest itself as a field rotation within subs, in some parts of the sky the rotation will be clockwise; in other parts anticlockwise and in some other parts of the sky no discernible rotation at all.

Mark

[cfpendock]

When guiding, as Mark says, in some areas of the sky (close to polaris), I find that polar alignment is much less critical.  There is an equation which might be interesting for you.  It indicates the accuracy required, depending on where you are looking in the sky:

E = (4500 x S x cosD) / (T x F x A)

Where :

E is the maximum allowable polar misalignment in arcseconds

S is the worst case length of star trails (in microns)

D is the declination of the target in degrees

T is the exposure time in minutes

F is the focal length in mm

A is the angle between the guide star and the target in degrees

Chris

[sharkmelley]

When guiding, as Mark says, in some areas of the sky (close to polaris), I find that polar alignment is much less critical.  There is an equation which might be interesting for you.  It indicates the accuracy required, depending on where you are looking in the sky:

The equation doesn't quite agree with your earlier sentence.  Since Polaris is at declination 90 (more or less) then cosD is zero (more or less) so the maximum allowable aligenment error is also zero (more or less).  So the equation is saying that a high degree of accuracy is required for imaging near the poles.

By the way, do you have a source for the equation?  I'm interested in how it is derived.

Mark

[cfpendock]

Thanks for pointing that out, Mark.  You are correct, of course.  I cannot explain why, but the equation comes from a paper by Frank Barret: http://celestialwonders.com/articles/polaralignment/PolarAlignmentAccuracy.pdf.  Can you explain it?

Chris

[sharkmelley]

Thanks for the link to the paper.  It'll take me time to digest its contents.

In the meantime, however, there is something I find troubling about it: the equation seems to suggest that field rotation near the celestial poles is greater than field rotation near the celestial equator.

However, consider a camera with a very wide angle lens on an imperfectly aligned tracking mount (without guiding).  The field of view is wide enough to contain both the celestial pole and equator.  Certainly field rotation will be noticeable in the image but it cannot be case that the rate of rotation in one part of the image is greater than rotation rate in another.  Or can it?

If another image is now taken with guiding switched on then, there will be no trailing on the guidestar and the stars in the rest of the image must all rotate at the same angular rate around the guidestar. Or maybe not?

To be honest, I have great trouble visualising what is going on so I could well be wrong.

Mark

[cfpendock]

I think that although the field rotation will be the same at the pole as on the equator (the angular rate will be the same), the actual length of any star trails will of course increase at the equator.  My problem with understanding the equation comes from the definition of "maximum allowable misalignment".  A sort of combination of negatives.....You are right - it is difficult to visualise!

Chris

[sharkmelley]

There's another thought experiment that gives another insight.  Consider viewing Orion through an alt-az scope.  In a sense this is a very misaligned mount (here in the UK it is approx 40 degrees off the celestial pole). 

If you view Orion rising, Orion is leaning backwards compared to the horizon. At transit, Orion stands upright and as Orion sets it is leaning forwards. So Orion has been rotating clockwise. In addition, during the time Orion is below the horizon, it is rotating anti-clockwise to eventually resume the leaning back posture. So, compared to the local horizon Orion is rotating clockwise when in the South; anticlockwise when North and it is not rotating at all somewhere near the East and West. 

This is what I mean when I say the rotation rate depends on where it appears in the sky.

Mark

[glowingturnip]

just thinking it through, suppose you were taking a picture of the celestial pole, on a slightly misaligned scope, then the sky would be rotating about the true pole whereas your scope is rotating around a close but different point - wouldn't that result in fierce field rotation ?  Indeed, stars between the false and real pole would even streak in a different direction than stars on the outsides since the scope then would be rotating against their direction of rotation (all hypothetical of course, I don't think I could get my aligned scope to point at the pole).

However, if you were pointing at something on the celestial equator, then you'd just have a very slow drift in declination, all one way for 12 hours, then back for the next 12 hours, so field rotation would be much less and guiding would easily fix it.

[Scott]

All these equations confuse me . 

What I do know is that I spend a bit of time and effort getting pa as good as I can and a mate just plonks it down aiming the mount northish.

His Images beat mine every time and I'm starting to hate him for it . Well apart from the hating him bit, the rest is all true and I just don't get it. (both scopes are guided).

[glowingturnip]

i like the theory that a slight mis-align means that your guiding doesn't keep switching on dec - sounds like an excuse for me to be lazy  ;-)

[Kirscovitch]

Equations boggle my mind as well but I agree that you should get it as close to perfect PA as possible. Sunday night I tried best I could to get the error to 0 but at best was a few seconds off. But even though my alignment wasn't perfect (according to the hand control) I was still able to get 190 second subs unguided. Makes me wonder how long I could go if I had really nailed it. Eventually I am going to try drift alignment and see what results I can get with that. I would like to get a guider but for now I just don't have the funds so will have to make due.

[glowingturnip]

i suspect there is no such thing as a perfect polar alignment anyway, since the axis of rotation of your scope is different to the axis of rotation of the earth (by the earth's diameter at your latitude) and the axis of scope rotation is also rotating around the earth's axis !

Mind you, I think I'm over-thinking it now and the error distances are literally astronomically small.

Anyway, polar scope with Polaris in the little circle works for me.

[Scott]

i suspect there is no such thing as a perfect polar alignment anyway, since the axis of rotation of your scope is different to the axis of rotation of the earth (by the earth's diameter at your latitude) and the axis of scope rotation is also rotating around the earth's axis !

Mind you, I think I'm over-thinking it now and the error distances are literally astronomically small.

Anyway, polar scope with Polaris in the little circle works for me.

Providing your little circle is in the right place

[glowingturnip]

haha, yup

[sharkmelley]

I have now read the paper by Barrett and also the paper by Hook that Barrett refers to.  It all makes sense to me now and I can visualise what is going on.

The easiest way for me to describe it is to consider a scope on an Alt-Az mount. If you are following a star in your scope eyepiece that passes close to zenith then fairly fast movements in Azimuth are required to continue to follow the star as it passes more or less overhead and this causes the field of view to rotate fairly quickly in the telescope eyepiece.

If instead you point the main axis of Alt-Az scope near the celestial pole you effectively have a (mis-aligned) German equatorial mount.  But you still have a problem when the star you are following passes close to the axis of the scope.  A fast rotation of the axis is required in order to keep the star in view and this causes field rotation in the eyepiece.

Similarly, when guiding with a German equatorial mount (or an Alt-Az on a wedge), the closer the guide star gets to the scope's main axis, the worse the field rotation problem.   To prevent this, the scope axis must be aligned to celestial pole with a much greater accuracy than the closest approach of the guide star to the scope axis.  The equation quantifies just how accurate this alignment needs to be.

In fact, when imaging near the celestial pole it can actually be better to switch off guiding than risk the possibility of the guide star passing too close to the scope's axis. Unless, of course, your polar alignment is extremely accurate.

Mark