Help - how to use platesolving to control CG5

[RobertI]

Thanks so much for doing this Louise, I tried with your settings and it finally solved in...1092 seconds. The search radius was set to 150 (which I know is huge) so I reduced this to 10 and it did it in 111 seconds. Some tweaking to do to speed up, but at least it works know! Next step is connecting, scope via ASCOM and camera using APT. 

Will try some more later when I’m back from work. Thanks so much for your help. 

Also thanks Furtysocks2.

[Thalestris24]

Hi

Yeah, tweaking the search radius is helpful - it relies on knowing where the target is. Normally, you'd be pretty close so the search radius can be small. Also, you can tweak the sigma value to limit the number of stars. Increasing sigma reduces the number of stars, decreasing it does the opposite. You don't want sigma too small as that would decrease the signal/noise. Obviously centering on a 'live' target is a bit different to solving a saved image. I usually expose for just 2-3 secs but you'll have try out to see what works for you. Working at a small fov seems intuitively more challenging - my 115mm apo + dslr gives me around 1.5 - 2 degrees at f5.5. Oh, having a fast computer is good  . You can also tweak the number of index files you need/use. You can tell by enabling the log viewer. You can move unused index files to a separate folder. They are installed in cygwin/usr/share/astrometry/data (I think!).

Anyway, good luck!

Louise

[stash_old]

1 hour ago, Thalestris24 said:

Hi

Yeah, tweaking the search radius is helpful - it relies on knowing where the target is. Normally, you'd be pretty close so the search radius can be small. Also, you can tweak the sigma value to limit the number of stars. Increasing sigma reduces the number of stars, decreasing it does the opposite. You don't want sigma too small as that would decrease the signal/noise. Obviously centering on a 'live' target is a bit different to solving a saved image. I usually expose for just 2-3 secs but you'll have try out to see what works for you. Working at a small fov seems intuitively more challenging - my 115mm apo + dslr gives me around 1.5 - 2 degrees at f5.5. Oh, having a fast computer is good  . You can also tweak the number of index files you need/use. You can tell by enabling the log viewer. You can move unused index files to a separate folder. They are installed in cygwin/usr/share/astrometry/data (I think!).

Anyway, good luck!

Louise

Louise is spot on with where the index files are (so long as you haven't done a custom install) and is also spot on with using the log to see what indexes are used and then trimming(renaming in case you need them later !) what indexes Astrotortilla uses.

You must not start out with the wrong impression -  as Louise states you must start platesolving with RA/Dec positions that are near to the object else  you are solving "blind" (no RA/DEC or too far away) and there is no way you will get fast solving - too much sky to look at.

Same as the FOV - luckily APT will fill in sensor sizes and together with F/length calculates the FOV for most DSLR's - not sure about CCD's

If you need blind solving then use the astronomy.net they have larger/ faster computers.

You will find you extra info here http://aptforum.com/phpbb/viewforum.php?f=24&sid=ef74abacb65721acbf102fee029266a6 if you haven't already

It will take time to get the settings right and being comfortable using plate solving but once its working -  

[Robrj]

Just curious, what camera and scope was that image taken with?  I want to test my spreadsheet to see what settings it generates.

[RobertI]

 

15 minutes ago, Robrj said:

Just curious, what camera and scope was that image taken with?  I want to test my spreadsheet to see what settings it generates.

It's an RC6 with an F3.3 reducer giving around F4 I think and a Lodestar mono camera. Sorry the vagueness of the F ratio probably doesn't help you!

[RobertI]

16 minutes ago, RobertI said:

 

It's an RC6 with an F3.3 reducer giving around F4 I think and a Lodestar mono camera. Sorry the vagueness of the F ratio probably doesn't help you!

Hmm on reflection, this looks like a much longer focal length (I calculated about 1060mm) so really not sure what reducer and f-ratio I was using, possibly the .63 reducer. Not helpful I know. 

[Robrj]

So I assume your scope is f/9.  With a .63 reducer, I get the following:

Scale Minimum is 0.3

Scale Maximum is 0.5

Indexes should include 4200 - 4206

My Custom Options line would look like this:

--sigma 70 --noplots -N none -H 0.5 -L 0.3 -r -objs 100

 

At 0.33x, I get:

Scale Minimum is 0.6

Scale Maximum is 0.8

Indexes would include 4202 to 4208

Custom Options:

--sigma 70 --noplots -N none -H 0.8 -L 0.6 -r --objs 100

 

To calculate what focal ratio you're at, you can get the arcsec/pixel from Astrometry.net and plug your info into the 5th calculator down:

http://www.wilmslowastro.com/software/formulae.htm

[RobertI]

13 hours ago, Thalestris24 said:

You can tell by enabling the log viewer. You can move unused index files to a separate folder. They are installed in cygwin/usr/share/astrometry/data (I think!).

 

3 hours ago, Robrj said:

So I assume your scope is f/9.  With a .63 reducer, I get the following:

Scale Minimum is 0.3

Scale Maximum is 0.5

Indexes should include 4200 - 4206

My Custom Options line would look like this:

--sigma 70 --noplots -N none -H 0.5 -L 0.3 -r -objs 100

 

At 0.33x, I get:

Scale Minimum is 0.6

Scale Maximum is 0.8

Indexes would include 4202 to 4208

Custom Options:

--sigma 70 --noplots -N none -H 0.8 -L 0.6 -r --objs 100

 

To calculate what focal ratio you're at, you can get the arcsec/pixel from Astrometry.net and plug your info into the 5th calculator down:

http://www.wilmslowastro.com/software/formulae.htm

Thanks Louise and Rob, I checked the files and I have 4202 to 4211, I think I also downloaded the files to support my wider FOV DSLR, but sounds like moving 4209 to 4211 might speed things up a tad.

13 hours ago, stash_old said:

t will take time to get the settings right and being comfortable using plate solving but once its working -  

Thanks for the help and encouragement Stash, it's looking good so far....

[RobertI]

Well it's proving tough, it stopped solving that same image, eventually my laptop started to do some wierd things, and virtually locked, so I had to reboot - this solved the problem and it started solving again (not quickly though), but it has just stopped again, not sure I have been methodical enough to be able to tell you what is going on, but when it takes 10+ minutes to come back with a solve/no solve, it's difficult to do any testing. I will persevere and post more information as and when....thanks for your patience!

[Robrj]

13 minutes ago, RobertI said:

 

Thanks Louise and Rob, I checked the files and I have 4202 to 4211, I think I also downloaded the files to support my wider FOV DSLR, but sounds like moving 4209 to 4211 might speed things up a tad.

Thanks for the help and encouragement Stash, it's looking good so far....

Adding files you don't need would actually slow it down, not speed it up because it Astrotortilla may try to search through them (which is a waste of time if you don't need them).   I would move everything above 4208 out of the folder Louise mentioned.  The log file that Louise mentioned will show which index actually solved.  I would keep the indexes that solved plus maybe one level above and below and move all the rest out.  If you find you're not using 4200 or 4201, move those out too.  They're really big.

[RobertI]

I just moved files 2008 to 2011 to another directory but it is taking ages (1296 seconds) and still not solving. The only thing I have changed since the last successful solve is the search radius now 20 instead of 10 (because 10 stopped working) and the scope position very slightly.

I have attached the log file and the screen capture of the parameters below - interestingly the last successful solution shown is not the location of the lemon slice nebula (the file used) so not sure what is going on there.

Need to rest and have another bash tomorrow.

 

 

log.txt

 

 

[RobertI]

I just moved the scope to the vicinity of the last solve (which is not where I expect the lemon slice to be), set the search radius to 10 and it solved in 33.9 seconds! Used file 4204-03.

log2.txt

Either it is solving the wrong thing, or I am looking at a different object to what I thought! Just about to put it into Astrometry.net to confirm.....

[RobertI]

OK, problem solved, my stupidity! It is not the lemon slice nebula it is NGC1501, it was not solving because the scope was too far away. That's what comes of trying to solve things late at night.

At least we are working again and I can move on.

[spillage]

If you are going to use APT then you just do all of this from within it. Just put platesolve into a folder install asps. Click on gear, click connect mount click on platesolve, Click settings and tell apt where both programs are and tweak the other settings.

You can then choose a list of objects or load up a previous image (using the IMG tab) to platesolve. You will need to choose the target from a previous image from the list in APT or if you are blind solving you just click the box for the mount current position.

Using a previous image once solved you click the solved button at the bottom and click goto++ and off it goes. It will take a few images and move the mount to within how ever many pixels you have set up in the setting option. I have mine at 5 pixels and take 5 second exposures.

If you are trying to blind solve and you have issue just stop it and then click solve and it will use the image it just took.

Hope this makes sense.

Spill.

I guess I could have just linked the users guide....

 

[stash_old]

Rob ,

Ignore this if you are already doing this !

To be able to do this during the day !!!!!!!!!!!!!

Something I did to help things move along and gets things working faster was use the Ascom Telescope simulator,CDC or such like and some images of different objects. Then use (as you are using Astrotortilla ) the Open Dialogue (which I think you have used from the above picture) to load an image.

Position the scope bang on the object with CDC and Platesolve. This will show you the indexes used in he log (hoping it solved else your settings are correct)

Then Position the scope further away (not too far) and repeat the above.

Once this is working its not difficult to set up APT to use Astrotortilla and it should be spot on - you can of course you different images from different camera to find the correct settings/indexes.

FYI Cygwin which is what Astrotortilla uses, is not a native Windows but in fact an Linux emulator. 

[RobertI]

Things are not going well. I managed to get AT to communicate with CDC and the ASCOM Telescope simulator, initially worked fine then takes longer and longer to solve. Eventually it seems to slow down other applications like Firefox and IE, to the point were they won't start at all.

I spent hours tweaking and testing over the weekend to no avail, so as a final test I did another reboot, set the parameters to make it a quick solve (see attached) and it solved in 60 seconds. Did exactly the same thing for a second time and it solved in 299 second. Did a third time and it solved in 2510 seconds. It seemed to be getting significantly longer each time to do exactly the same thing. I've attached the log too.

So I really don't know what is going on. I ran a virus check and malware check and removed redundant startup process, didn't make a difference sadly.

I don't want to push my luck with all the people who have provided so much help so far, but does anyone have any idea what is going on? 

 

 

Index sample.txt

[Thalestris24]

Hi

I don't know if there might be a bug in the software but I've seen the same thing when trying to blind solve a target that's already been solved. In normal operation/use it seems fine. Perhaps you could try it for real? Have faith!

Louise

[stash_old]

Dont know what version of AT you are using but I upgraded to 7.?? from 7.??(think v8 now but no experience of that)  and I had loads of problems - the older version seemed to work better.

Make you virus checker ignore the directory where your indexes are  held - just depends on the algorithm but the virus checker might be slowing your down - try switching of the Virus checker for 10 mins while you are not connected to the Internet and preferably after a scan.

As I pointed out AT uses CYGWIN which emulates Linux as creates an overhead.

Also it is doing alot of reading and comparing so hopefully you have plenty of memory (32bit max <4Gb, 64bit > 4GB) and if you get it working reasonably but want faster times get an SSD for the system and/or Index's.

You have moved from no solving to solving but slowly the only way is forward. Clear Skies! ----- and faster Platesolving LOL

[RobertI]

2 hours ago, Thalestris24 said:

Hi

I don't know if there might be a bug in the software but I've seen the same thing when trying to blind solve a target that's already been solved. In normal operation/use it seems fine. Perhaps you could try it for real? Have faith!

Louise

Thanks for the encouragement Louise, perhaps I won't give up quite yet! It really is bizarre though.

 

2 hours ago, stash_old said:

Dont know what version of AT you are using but I upgraded to 7.?? from 7.??(think v8 now but no experience of that)  and I had loads of problems - the older version seemed to work better.

Make you virus checker ignore the directory where your indexes are  held - just depends on the algorithm but the virus checker might be slowing your down - try switching of the Virus checker for 10 mins while you are not connected to the Internet and preferably after a scan.

As I pointed out AT uses CYGWIN which emulates Linux as creates an overhead.

Also it is doing alot of reading and comparing so hopefully you have plenty of memory (32bit max <4Gb, 64bit > 4GB) and if you get it working reasonably but want faster times get an SSD for the system and/or Index's.

You have moved from no solving to solving but slowly the only way is forward. Clear Skies! ----- and faster Platesolving LOL

I was on version 7.0.0 I think, so perhaps I shall install previous version and see what happens. Checked resources and memory and CPU were fine while it was running (its an i3 with 6Gb memory). Its odd that it seems to be causing all kinds of other issues with the laptop (I have just rebooted as it ground to a halt - never done this before), perhaps its something to do withCYGWIN. we shall soldier on for a bit more..... ?

[stash_old]

Or just try ASPS in APT - As I say its is a Windows program (oh sorry APP LOL) , FREE and intergrates nicely with APT

Good Luck Robert.

[Robrj]

17 hours ago, RobertI said:

I was on version 7.0.0 I think, so perhaps I shall install previous version and see what happens. Checked resources and memory and CPU were fine while it was running (its an i3 with 6Gb memory). Its odd that it seems to be causing all kinds of other issues with the laptop (I have just rebooted as it ground to a halt - never done this before), perhaps its something to do withCYGWIN. we shall soldier on for a bit more..... ?

I would just try it in a live setting.  I use 0.7 with my setup.  My computer is an Intel Computestick (2nd gen) which has a quad core  atom processor and only 2gb of ram.  AT 0.8 is an alpha version so it may not be stable.

[RobertI]

Just to continue the story on this, I installed All Sky Plate Solve and really liked the interface, but got a .dll erro when installing the index file, and it never seemed to work properly. 

So last night was clear and I had a session, after 20 minutes of careful aligning of the scope, i knocked the power cable out of the mount! So decided to try the CDC/Astrotortilla route. I managed to connect the mount to CDC, but when I tried a slew, it gave me the message "cannot slew unless telescope has been aligned". I could not see a way of doing this in CDC, so went and did a quick align using the handset (didn't seem the right to do, but couldnt think of anything else). Slewed the scope and it ended up pointing to the neighbours shed, so something wrong. I may have input the date wrong. Gave up at this point and did a manual align and reverted to my usual goto approach, which actually worked perfectly on this occasion. I'll post the results in the Video Astronomy forum tomorrow.

I am not giving up on plate solving and will try again next time, but any ideas on the CDC message?

Thanks for you continuing patience with my saga!

[Robrj]

For mine, I do my usual alignment and then use Astrotortilla, so I would have done what you did at the end first.  For the AVX, I use Starsense so the scope aligns automatically but with my 130SLT, I just do a manual alignment with the hand controller.  After that point, I can control it with Stellarium and plate solve with Astrotortilla after the goto.

[RobertI]

8 hours ago, Robrj said:

For mine, I do my usual alignment and then use Astrotortilla, so I would have done what you did at the end first.  For the AVX, I use Starsense so the scope aligns automatically but with my 130SLT, I just do a manual alignment with the hand controller.  After that point, I can control it with Stellarium and plate solve with Astrotortilla after the goto.

Thanks Rob, I think I was being confused the way NexRemote works - once you use the hand controller to align, you should't use the NexRemote software, and vice versa. But it sounds like with CdC/astrotortilla/Ascom setup you can align with the handcontroller and then use CdC & AT. I will give this another go next time, thanks. ?

[Thalestris24]

Hi Rob

Using Astrotortilla there's no real need to star align. However, having good balance and good PA is important regardless. Then you should naturally be quite close to star alignment and AT will centre your target. I know you have a small fov and that might make things a bit more challenging...

Louise