Calculating polar alignment error?

[jambouk]

364

Is it possible to calculate the amount of error in ones polar alignment by looking at two images spaced apart in time?

These two images were taken 364 minutes apart with the mount left tracking between.

The kit was: ED80, 0.85x SW reducer, Canon 6D, unguided (obviously)

Target: M81/M82

Thanks.

James

[Thalestris24]

Hi James

If you put a number of unguided consecutive images into DSS it will calculate the dx, dy values which you can plot and use to calculate drift rate. Polar Alignment error shouldn't vary over time (as it's mechanically fixed via mount adjustments) but the degree of field rotation should give an indication of how much PA error you have though I think it will likely only show up with fairly long exposures (unless your PA is way out!). Had you guided with PHD2 then the log gives an estimated PA error.

Well, that's what I believe. Maybe someone else knows more.

Louise

[jambouk]

That is really helpful, thank you. I flicked between the two subs and there doesn't seem to be much field rotation, so maybe most of the drift is cumulative period error? Does periodic error accumulate and result in drift, or is it cyclic in the sense of it adds a bit of drift, then removes that same amount of drift?

I'll load the two subs into dss tomorrow and see what data it gives me.

Thanks.

James

[Thalestris24]

That is really helpful, thank you. I flicked between the two subs and there doesn't seem to be much field rotation, so maybe most of the drift is cumulative period error? Does periodic error accumulate and result in drift, or is it cyclic in the sense of it adds a bit of drift, then removes that same amount of drift?

I'll load the two subs into dss tomorrow and see what data it gives me.

Thanks.

James

Hi

I believe drift accumulates over time. It will show up round the edges if you stack lots of short subs - hence the need to crop the stacked image. I'm not 100% sure about how periodic error manifests when unguided. It doesn't prevent the mount being held steady by guiding corrections. I guess unguided it would result in a small oscillatory effect between star positions across subs over time but, as I say, I'm guessing there! Normally stacking takes care of all these things. I suppose if you have a lot of drift and take lots of subs over a long time, you could have problems stacking else have to crop a lot off... It doesn't look like you have a great deal of drift considering the time interval between your two subs above.

Louise

[jambouk]

Thanks again. I think the actual amount of drift is about one degree.

I am interested to see what dss says; i'd get out of bed now and look, but itms currently stacking hundreds of those images and i dare not try and open it twice as my laptop is on its last legs and crashes at even the thought of having to process more than one thing at a time!

[Thalestris24]

Thanks again. I think the actual amount of drift is about one degree.

I am interested to see what dss says; i'd get out of bed now and look, but itms currently stacking hundreds of those images and i dare not try and open it twice as my laptop is on its last legs and crashes at even the thought of having to process more than one thing at a time!

It will keep! I wouldn't worry about it anyway

[jambouk]

I just loaded up the two images (the ones 6 hours apart) into DSS and let it work out the dx and dy...

Not sure how I then convert this into the amount of drift. Any ideas:

[Thalestris24]

Hi James

There used to be a Wiki for this but it's no longer available though there is discussion in the Yahoo group https://groups.yahoo.com/neo/groups/DeepSkyStacker/conversations/topics/6840. I think what you do is put the dx, dy values into a spreadsheet and plot them. However, you need a number of dx,dy values really - maybe 10. Since you know the time that it's taken, you can calculate the rate of drift. Apparently you can calculate PA error from the drift rate: http://celestialwonders.com/tools/driftRateCalc.html

See also: http://celestialwonders.com/articles/polaralignment/index.html

Hope that helps

Louise

[jambouk]

Oh, that is getting serious... I might just put this thought process on hold for a time, I can't face the maths Plus I've had a glass of wine!

James

[Thalestris24]

Hi James

TBH honest I wouldn't worry about it! You can just re-centre your target periodically

Louise

[Carina Lass]

The periodic error cycles around the same point whilst drift - does exactly that.

It is an angle across the sky as your RA progresses.

If the same stars are visible in your pics after 6 hrs, the drift is quite small as the usual angle of the pic is less than 1 degree - that's one degree in 6x15degrees of RA.

If you can see how much it has moved, could be a good move to adjust your dec bolt by half the amount and try again!

[jambouk]

duplicate post, sorry

[jambouk]

I'm being a bit thick, not enough sleep... How do you know it is the altitude which is slightly off rather than the azimuth [i presume you mean altitude when you said dec].

James

[kens]

The declination error due to polar alignment reaches a maximum after 6 hours - it is a sine function so after 6 hours the error reduces and gets back to zero after 12 hours.

I worked out on the back of an envelope one time that the error after 6 hours was 23 times the error after 10 minutes.

I found this site helpful to understand it http://www.eanet.com/kodama/astro/equipmnt/p-align/driftaz.htm

Polar alignment error has two components: Azimuth and Altitude. The photos you have are not enough to tell how much each is contributing, For one thing the declination setting for the photos will affect the amount of drift - near the celestial equator it is largest. Near the horizon the altitude error has the most effect whilst at the meridian the azimuth error has most effect.

Also, in your photos it is not certain which way is RA and which is Dec so there could be RA errors in there as well.

RA periodic errors are typically on a cycle of about 10 minutes - depending on the rotational speed of the RA worm. If  component of periodic error is included in your "drift" on the photo then it could be contributing anywhere from zero to as much as 20 arc-seconds.

[jambouk]

I need to have a read of that link.

Thanks for that. It's a lot to get my simple head around so I'll come back to your post once I've read that document.

Thanks for the detailed reply, much appreciated.

James

[jambouk]

That is a useful document; and interesting. Thank you. It's good to see some of those things i know a bit about down on paper.

I'm going to look at the images again and see if there is a trend in the drift.

Thanks.

James

[kens]

Oops - I only gave you half the link and left off the bit about altitude

Here is a link to the main page http://www.eanet.com/kodama/astro/equipmnt/p-align/index.htm

And the page on drift alignment which link to both the azimuth and altitude examples http://www.eanet.com/kodama/astro/equipmnt/p-align/drift.htm

When I was trying to work out my drift and periodic errors I found it useful to take a series of shots about 1.5 or 2 minutes apart over a bit more than 10 minutes, lining up the camera so that RA was left/right and dec was up/down. This was done with the scope pointing at the celestial equator(Dec 0 deg)  at either the meridian (for azimuth error) or the horizon (for altitude error). I could then overlay those shots to show both the drift in declination and any periodic error. Total alignment error in either azimuth or altitude is then 23 times the 10 minute drift.

Now I use PHD and its a lot easier and faster. Point at the same stars as above, turn off declination correction and turn on trend lines. Within a few seconds it calculates the total error