Plate solving and goto with AstroTortilla

[Realt]

Thanks it solved ! 31 secs

What is odd though is AT changes the min and max values after a solve. And then when i run the same image again it does not solve with those values.....

Thanks Russ and Realt !

Result...Well done

Re changing values- This may be to do with scale refinement setting.

0 is off

If you set it to ,for example, 0.1 it will set min max within 10% of the value it finds for the solved image.

This value should be good for all subsequent images with the same scope/camera set up.

[kirkster501]

This is really cool! With those min and max values its now solving all my old pictures with that setup!

So I connected it all up to scope (indoors of course with this naff weather ). I do an align on one star to give CdC/EQMod a frame of reference then slew to say M31. I then do a plate solve and AT then fine tunes the alignment ! Its brilliant ! Is that alignment necessary though? Just forget it?

I have noticed a few slight bugs though. If I do a slew to another object and do another plate solve it will successfully solve the plate OK but then the status of AT in the bottom bar stays on "re-centering" and seems to be stuck in that state and I have to do an "abort" which then seems to lock up AT. Anyone else seen that?

I suppose the question is once you do one plate solve do you then use the object that you are currently on to do a "sync" in EQmod so this value is then available to new slews? Or once you do one plate solve that's it for that session - no need to do further plate solves because EQmod is programmed precisely after the first solve and even if a slew to the other part of the sky it will be spot on? Do you need to do a plate solve after every slew to a new imaging target?

Sorry for so many questions. Its just so difficult to test these thing out with this weather!

Rgds, Steve

[chrisshillito]

Steve,

AT is not doing any goto modeling itself. It is simply providing an accurate sync point for EQMOD to use. If you want an alignment model that accurately maps your mount / scope to the sky then you will need multiple alignment points to cover the sky - so multiple plate solves.

In reality though most folks don't bother with an all sky model and only create alignment points locally to the objects they visit. So if you change target just do the goto, plate solve/ sync and goto again.

Chris.

[Sp@ce_d]

This is really cool! With those min and max values its now solving all my old pictures with that setup!

So I connected it all up to scope (indoors of course with this naff weather ). I do an align on one star to give CdC/EQMod a frame of reference then slew to say M31. I then do a plate solve and AT then fine tunes the alignment ! Its brilliant ! Is that alignment necessary though? Just forget it?

Mine is on a pier in the obsy but I reckon you'll still get away with same as me after doing a PA.. Just slew to your target then do a solve. It will correct and center target. Dont bother doing a star align first.. thats what solving is for.

I have noticed a few slight bugs though. If I do a slew to another object and do another plate solve it will successfully solve the plate OK but then the status of AT in the bottom bar stays on "re-centering" and seems to be stuck in that state and I have to do an "abort" which then seems to lock up AT. Anyone else seen that?

I've seen this using Nebulosity where it stops talking to AT. If you run the log viewer it'll be waiting for camera. I just restart Neb. Normally its when I switch between the DSLR & the Atik... Whats your log output when this happens?

I suppose the question is once you do one plate solve do you then use the object that you are currently on to do a "sync" in EQmod so this value is then available to new slews? Or once you do one plate solve that's it for that session - no need to do further plate solves because EQmod is programmed precisely after the first solve and even if a slew to the other part of the sky it will be spot on? Do you need to do a plate solve after every slew to a new imaging target?

Sorry for so many questions. Its just so difficult to test these thing out with this weather!

Rgds, Steve

It automatically adds each solve sync to EQmod anyway.. obviously the more you do the better it gets. Although for imaging you're only normally on 1 or 2 targets a night .. (if you are really lucky with the weather) so I only solve for the target I'm on. It all depends on if your target is in the centre.. or where you want it after the slew. I normally find my target on the sensor somewhere when slewing away after 1 solve, even if I slew to the other side of the mount

EDIT: Ah.. Chris beat me to it..

[kirkster501]

Steve,

AT is not doing any goto modeling itself. It is simply providing an accurate sync point for EQMOD to use. If you want an alignment model that accurately maps your mount / scope to the sky then you will need multiple alignment points to cover the sky - so multiple plate solves.

In reality though most folks don't bother with an all sky model and only create alignment points locally to the objects they visit. So if you change target just do the goto, plate solve/ sync and goto again.

Chris.

Thanks Chris, that's my conclusion too. I need to get under the sky to test this out properly to see how robust it is and do some further testing with it.

Steve

[kirkster501]

.....Just slew to your target then do a solve. It will correct and center target. Don't bother doing a star align first.. that's what solving is for.

That would be incredibly powerful if so. The least enjoyable part of the whole hobby for me is the alignment on the first star. Its a PITA 'cos the first slew is normally some way off. If the plate-solve eliminates that and provides an initial - and highly accurate - starting location vector to EQmod, well that simplifies the whole setup procedure dramatically! I can't wait to test this under a clear sky - probably in four months at this rate....

[Sp@ce_d]

That would be incredibly powerful if so. The least enjoyable part of the whole hobby for me is the alignment on the first star. Its a PITA 'cos the first slew is normally some way off. If the plate-solve eliminates that and provides an initial - and highly accurate - starting location vector to EQmod, well that simplifies the whole setup procedure dramatically! I can't wait to test this under a clear sky - probably in four months at this rate....

AT is sooo cool... It had me dribbling with excitement the first time I clicked on the "Make it so" button, stood back and collected my jaw off the floor. It gets my vote for the single addition to my kit that has changed my life... my other half will testify to that too, she no longer has to cover her ears when I come back in from the obsy...

[mickut]

Thanks it solved ! 31 secs.

I need to get my head around the values to input I think and read the documentation a little more. At least I am set up and configured correctly!

Since it's been dull and cloudy here for the past few months, I've been playing around a bit with the plate solver parameters and the 4000-series indexes. Adding the extra options -r -c 0.02 you can make the solves significantly faster. The best improvement on one test-frame was from 67 seconds down to 8s. With the 200-series indexes these parameters are not really that necessary. These are also covered in the FAQ now.

Overall the solver works faster if you give it a 2x2 binned image and don't bother downscaling in software. With my KAF-8300 and 2x2 binning using 2-5 seconds exposures I get a fairly consistent 200-500 detected objects pretty much anywhere on the sky. This seems to be the "magic" number for fast plate-solves, although it usually solves with the first ten or twenty objects. Playing with the --sigma and exposure time using a few stored test shots pays off when the skies clear.

Regards,

Antti

AstroTortilla author

[Sp@ce_d]

Since it's been dull and cloudy here for the past few months, I've been playing around a bit with the plate solver parameters and the 4000-series indexes. Adding the extra options -r -c 0.02 you can make the solves significantly faster. The best improvement on one test-frame was from 67 seconds down to 8s. With the 200-series indexes these parameters are not really that necessary. These are also covered in the FAQ now.

Overall the solver works faster if you give it a 2x2 binned image and don't bother downscaling in software. With my KAF-8300 and 2x2 binning using 2-5 seconds exposures I get a fairly consistent 200-500 detected objects pretty much anywhere on the sky. This seems to be the "magic" number for fast plate-solves, although it usually solves with the first ten or twenty objects. Playing with the --sigma and exposure time using a few stored test shots pays off when the skies clear.

Regards,

Antti

AstroTortilla author

Hi Antti

You've done a great job with AT. Thank you

[kirkster501]

This application is amazing. I can see quite clearly it can revolutionise the workflow of an imager. I can't wat to give it a try under the skies. I've had it hooked up to my mount and works a treat.

Well done Antti, your work is massively appreciated please keep the innovations coming. We'll help you with input!!!

[Realt]

Finally got outside under reasonable skies with AT last night.

When using with APT I get this error message;

camera error 'DON'

I can solve the images saved by APT no problem using the file dialog and everything works hunky-dory.

I am using Canon 1100d and 120ed (no ff/fr) on neq6

A search online doesn't shed any light.

Anybody got any ideas.???

[russellhq]

Have you got the latest version of AT? I had this issue with version 0.4.2, but installed the latest version (0.4.3) and it's now fixed. What version of APT are you using?

[Realt]

Have you got the latest version of AT? I had this issue with version 0.4.2, but installed the latest version (0.4.3) and it's now fixed. What version of APT are you using?

I have 0.4.2 version of AT

I'll update and see if it does the trick.

I'm using unregistered APT ver 2.01 which is pretty recent I think.

Cheers russellhq

[russellhq]

There is now a 2.10 version of APT which has a Bahtinov focus mask tool and live view framing mask which looks quite handy! I've not had a chance to use either though as it's been cloudy for the past week.

[Realt]

Sorted I think.

Installed version 0.4.3 and it stopped giving that error message.

Only tested indoors obviously but definately seems to have been a bug in 0.4.2

Thanks again russellhq.

Hoping to use it to track asteroid 2012DA14 this evening if these clouds would naff off.

[Realt]

There is now a 2.10 version of APT which has a Bahtinov focus mask tool and live view framing mask which looks quite handy! I've not had a chance to use either though as it's been cloudy for the past week.

Yeah I was actually aware of that new version.

Wouldn't make any signifigant difference to me right now but the improvements do look interesting.

[russellhq]

Hoping to use it to track asteroid 2012DA14 this evening if these clouds would naff off.

How will you be doing that?? I've only used the scope for tracking the stars and planets.

[kirkster501]

Thanks for the heads up about 4.3. Installing it and APT 2.10 ready for, hopefully, tonight. All seems to solve from photos from my ED80 and RC8 scopes but want to test under the sky. Constant cloud for weeks has prevented me doing so.

[Realt]

How will you be doing that?? I've only used the scope for tracking the stars and planets.

I don't mean tracking in the normal sense.i.e the scope following the object to keep it centred.

I mean positioning the scope accurately at various points along the predicted path of the asteroid, (usually followed by photographing the starfield and then blinking the images.) Astrotorrtilla should make this a doddle.

I do know however that you can use software to get your mount to track sattelites so I don't see any reason why you couldn't track asteroids. Eqmod has a custom tracking rate-I suppose this could be used

[chrisshillito]

EQMOD not only has a custom tracking option but with the latest test release V1.27d this can be defined by ephemeris data obtained from the minor planet center.

Chris.

[Realt]

EQMOD not only has a custom tracking option but with the latest test release V1.27d this can be defined by ephemeris data obtained from the minor planet center.

Chris.

Excellent. You are fast becoming my all time hero Chris.

Thanks for all your excellent work.

[Realt]

And thanks to all the other free software developers also of course.

[kirkster501]

Well I am out imaging and AT and EQMod got The Rossette nebula 100% bang on centre !!! Capturing as we speak , going to try some other targets as well later.

Sent from my iPhone using Tapatalk

[Paul Laws]

I've had nothing but trouble with the 0.4 series but found a new 0.5 version so I'm running this now - without problems

Regards.

[kirkster501]

Well, After two nights of use I have to say this software has transformed my workflow. All tests done on my Ed80 scope first.

No longer need to do three star align or any of that malarkey.

1. Polar Align and set scope to home position.

2. Slew to Dubhe. Because my home position is pretty good now I get Dubhe in FoV. Center Dubhe then swap EP for camera and focus with Bahtinov. Sync on Dubhe.

3. Slew to object and plate solve. It is 100% EXACT in every test I did.

I tested this with M31, Horsehead, M67, Rosette, M81. Every time, AT put the subject absolutely bullseye dead center in the FoV.

Most incredibly impressive!!!!

Next is to test on RC8 scope.