Polar Alignment using AstroTortilla

[Guy_P]

Hi,

I've just started using AstroTortilla and find it very useful. I was therefore very interested in its Polar Alignment tool.

I tried it the other night and it was plate solving OK but the alignment error reported varied a lot (+/- 10's of arcminutes) - even when I didn't adjust the mount.

I noticed that other people have reported similar problems in the past.

Is this a current problem or is there anyone using it reliably who can help me resolve this issue?

Thanks,

Guy
 

[Thalestris24]

Hi and Welcome!

I use AT and I use the PA tool . It is a bit iffy... But it's the only way I can do PA as I image from my flat and have a very limited view of the sky. Fortunately I can see due East so usually get reliable PA alt error at about 30 deg DEC. I can only do azimuth at about 150 deg which isn't ideal - I think you really need to be close to the meridian. You might find it useful to incorporate a little utility called Polefocus http://www.scopefocus.info/polar-alignment which lets you adjust on a sync'd star. Or, if you have APT, you can use that instead http://aptforum.com/phpbb/viewtopic.php?f=17&t=169. I think it's quite important to keep exposure times down to no more than about 3sec for PA adjustment. You may have to do several iterations. If you're doing guided imaging, PA errors are fairly tolerable, I think. However, if you get an unexpected, crazy sounding PA error from AT then it's probably a good idea to just repeat the measurement

Hope that helps

Louise

[Guy_P]

Thanks Louise,

Thats good news that it works - even if it is a little tempermental.  As you say it is very useful since it doesn't need a specific set of stars.

I tried it the other night using Polefocus to help with adjusting the mount.  When I went back and checked it it was further out than before.

I checked it again and it gave me varying results.

I did have a 5s exposure time so that may be part of the problem.  I'll try and reduce it to 1 or 2s to see if that helps.

I use APT but I haven't used its offset tool - do you use that?

One question I'm not sure of is whether I should use JNow or J2000 in AT - which do you use?

Next time I'll measure the alignment error using the drift method and see if it correlates.

Thanks for the help

Guy

[Thalestris24]

Hi

I suppose AT is essentially drift aligning programmatically. It might be less accurate with longer focal lengths - I wonder what scope/mount/camera you have? How long does it take you to solve usually? If in doubt, you can always drift align in a conventional way. I use 'Jnow'. I've not used the APT tool. My scope is more or less permanently setup so I don't have to PA from scratch each time - just a quick check. Not that I get to image very much anyway what with the Glasgow weather .

Louise

[Guy_P]

My main scope is an SW 200 pds.  Most of the time it takes 20-30s to solve - although I'm still refining the configuration.

I know what you mean about the weather.  i'm just trying to make sure I can set up as quickly and accurately as possible for the rare clear moonless night.

Getting AT polar alignment working reliably would help.

I'll try the shorter exposure time next time.

Thanks

Guy

[Thalestris24]

My main scope is an SW 200 pds.  Most of the time it takes 20-30s to solve - although I'm still refining the configuration.

I know what you mean about the weather.  i'm just trying to make sure I can set up as quickly and accurately as possible for the rare clear moonless night.

Getting AT polar alignment working reliably would help.

I'll try the shorter exposure time next time.

Thanks

Guy

Hi

What camera? You should be able to get AT to solve in about 4-20secs - if you can adjust yours to get it to solve quicker than currently, that should help with the PA accuracy, I think.

Louise

[Guy_P]

The camera is a Canon1000D and the mount is an HEQ5 Pro -   I calculated the FoV to be 1.27 x 0.85 degrees so I've downloaded indexes between 4204 and 4210.

So you think the errors I'm getting might be associated with my solve time as well?

I thought 20-30s was pretty fast - it is running on an old laptop - I've done some optimisation to get it to that - any tips to reduce it further?

Guy

[Thalestris24]

Hi

Longer solve time might contribute. The stars will have moved a bit from where it first thinks they were. But that's an assumption on my part - I'm no expert. It might be worth posting a question about it on the AT forum. It is quite processor demanding so a slower laptop will take longer, I suppose. As I said before you can monitor which indexes are being used to solve in the view log option. That way you can determine which aren't being used and move them to a separate folder. If you find you suddenly can't solve in a particular area of sky you may need to move one or two back. I think the smallest indexes might be used to solve star dense regions i.e. in and around the milky way. I limit my search radius to 15 deg. Usually, if it can't solve I find it's because it can't see enough stars. It (briefly) tells you how many have been detected at the bottom of the window -  it must say in the log also.

Louise

[Guy_P]

I'll try removing some of the files it doesnt seem to use.  Have you removed some files - if so which?

I've limited the search radius to 15 degrees too

I check the number of stars detected and adjust capture time, downscaling and sigma accordingly to be sure I get a decent number of stars. 

So far I've been able to get it to solve and I've been trying tuning the settings using the old AT images in the AT temp directory - which is how I got down to 20s.

As a test I can probably temporarily remove all the files except the one it needs for that area of sky - that should reduce the time. 

At least that way I can see if solve time is a factor in the PA measurement accuracy

Next time I the weather gives me a chance I'll try all these ideas.

What sort of solve times are you getting?

Guy

[Thalestris24]

Hi

I can't remember! It's been a while since I last changed anything... I can tell you what I currently have for my 150pds/qhy8l combo. I have indexes 4208-4213. That's with the 64 bit version of AT running on an i5 PC. It does tend to fail to solve around the Milky Way so maybe I need to add back some of the smaller indexes though I can usually solve nearby and still slew to target accurately via Stellarium. It also tends to fail if it's already spot-on. Don't know why that is but it doesn't cause a problem.  Imaging with a dslr, it's worth downloading raw + jpg in APT. It's quicker for APT to download jpg directly from the camera rather than having to convert from raw. I think solving from files will always be slower since they have to be read into memory from disk first. This is my 150pds configuration:

It usually returns a solve time of about 4-15s so I guess you should expect similar times with live images. I have to say the lp is bad where I am plus the atmosphere tends to be misty - both of which combine to make imaging and plate-solving challenging...

Louise

[Guy_P]

Hi.

Thanks for the information.

I'll look at any differences in my setup and try them next time I have a chance.

In some cases AT might fail to solve an image and then after that it continues to fail no matter what you do.  Its like it just repeats the previous result.

In these cases I've found the --overwrite parameter helps.

I've been going back over old posts in the Sourceforge AstroTortilla forum and have found a couple of interesting posts about possible things to try to reduce the erratic PA measurement:

i)  I may be taking images too close to the horizon - try taking them at 45 degrees altitude or so

ii) see if there is a difference between taking measurements East of the meridian or West

So when the weather clears there is a lot I need to try out.

I'll post any definite results I get.

Guy