Plate solving error.

[Davey-T]

Finally got to try out the 10Micron mount last night using Pers Modelmaker software which connects to Maxim and Pinpoint for plate solving via stored catalogue not internet.

The first time I started it using a 10 point model just to get the hang of it it all worked and solved the images apart from one failed solve.

I deleted that model then selected another 3 base points which it solves first but the solves failed with this message, anyone know what it means and is the problem with the mount, Modelmaker or Pinpoint or ?

Turned everything off and on again but got same message.

Dave

 

 

 

[ChrisLX200]

Thats... odd. It seems to be compaining about pixel scale in one axis only. If you open Visual Pinpoint and look at Solve Settings/Plate Parameters does it show the correct values for Horiz and Vert scales? Also, in MaximDL have you set the value for Focal Length under the Site and Optics tab. I take it you are using Per's mount driver and not the 10-Micron one? - you also enter the value for focal length and aperture here. I can't think of anything else off the top of my head...

ChrisH

[Davey-T]

Thanks Chris, I set the scope specs in Pers MM and Maxim, will have a look in PP, shame nice clear night last night and thought I might get my first image after finally getting it to play nicely with Win10 .

Strange it worked the first time and then didn't work again after making no changes.

Also got a problem with MM not fitting on the screen vertically the last couple of lines hang off the bottom of the screen, something else to Google

Dave

[Davey-T]

Does this look right arc/secs pixel scale in PP 2.0, 1.59, Maxim calculates the 1.59, did PP calculate the 2.0 ?

Do I need to download ansvr to plate solve sans internet ?

Do I need to fill in the path to GSC/1.1

Dave

[ChrisLX200]

Yes you need to the path to the catalog - not sure how you got it to work at all without it TBH but the error you have was nothing to do with that anyway. Are you sure the plate scale is exactly 2.0"/pixel? The scale should obviously be the same in both axies with square pixels (and the Atik website says they are square for the 314L). You can have a 25% plate scale error (by default) and PinPoint should still solve but it's as well to use an accurate figure to start with (to two decimal places). Just solve an image off line and use the calculated figure from that. The other parameters I cannot recall what I have them set at - I'll have to look later tonight..

ChrisH

[Davey-T]

Camera is QSI683 5.4 pixels Maxim calculates 1.59 arc/secs pixel don't know where PP got 2.0 from.

Says something in PP  instructions about needing to download ansvr to solve offline but not very clear.

Thanks,

Dave

[ChrisLX200]

OK, wrong camera   However, putting in an accurate figure for plate scale will still help because a calculated figure is never quite right. PinPoint is more relaxed about scale than (say..) Elbrus, which would simply fail every time unless the scale were accurate to 2 decimal places, but PinPoint may be very close to its error limit if the real scale is 1.59 and it is using 2.0. The only way to know is to solve an image first and read off what it says. I still don't understand why the error refers to just the horizontal axis though - I;ve never seen that. It usually just says the plate scale is 'probably incorrect'.

I do have ansvr installed on the Obs PC - mainly because I used Elbrus off-line a couple of years ago. I don't recall that requirement for PinPoint though which will solve anyway if you have a catalog available on disk. Maybe it does... I have several catalogs on my HD apart from the GSC, like USNO A2 and UCAC4, but these days I just use GSC for general solving (the others for asteroid atrometry).

ChrisH

[Davey-T]

Had another go at this and still can't get the plate solving to work.

I start off in the normal weights down aiming north position and can see Polaris in a 30mm eyepiece

Start MM as per pic connects to mount detects GSC star catalogue, connects to Maxim.

Haven't started Stickstation but would that stop MM working ?

So put in three points and run and get error as pic, takes about 30sec's to fail and MM stops responding, see pic'

[ChrisLX200]

I would use your StickStation for the refraction data - it makes a big difference to positions nearer the horizon.

Increase your Solve Expansion to 60% - offers some insurance against larger than expected positional offset.

Increase exposure time to 8sec in case PinPoint is failing to separate stars from background noise (especially when the moon is bright)

Max Stars in Model should be set to 100 in case you feel like doing a full model sometime

Slew settle time I have at 2 seconds - problably irrelevant.

Unless you have a good reason to change the default directory for the catalog it's best to leave it at C:\GSC11 (again, should make any difference but...)

You don't need to go as deep as Mag 20 (in PinPoint's Use Stars setting) - mine is set to mag 18, and that only because I use the same settings with the ODK12

Have ModelMaker set to 'Clear Model in Mount' each time you run a new model - otherwise it will add the alignments to existing model with results which may not be what you expected (i.e., invalid data)

That's all I can think of right now..

 

ChrisH

[Davey-T]

Thanks Chris, will hopefully have another go tonight.

I put the GSC catalouge in C:\GSC11 but Maxim ignores this and wants it in  PP so I had to put it there, I see someone on YT has it in the Maxim folder ?

Do you know any reason the mount update can't be done wirelessly ? (no ethernet socket on this laptop) it says connect by cable in the instructions, I need to update to get the GPS module supplied by 10Micron working as strangely it says it needs latest firmware to work, its only a standard Navilock GPS unit so no idea why it needs the mount updating ?

Looking into the ansvr thing I think it's only required for blind solving as far as I can make out, do you think this is right ?

Dave

 

[ChrisLX200]

Sorry, I've never connected mine wirelessly and I'm uncertain as to whether it's possible or not to update that way. There was some comment on the 10M forum about the new issue of GPS module requiring you to use the latest firmware but again, I took no notice because mine (old version I guess) worked anyway. I think we're up to v.2.13.22 now.

Yes to the ansvr thing, it allowed Elbrus to solve off-line when I used it. Since replaced with PinPoint but ansvr still loads at boot time (I never bothered to change anything).

I just noticed on your screen grabs that MM was hanging whilst trying to solve?

ChrisH

[Davey-T]

18 minutes ago, ChrisLX200 said:

I just noticed on your screen grabs that MM was hanging whilst trying to solve?

ChrisH

Yes, said "not responding" after trying to solve for about 20sec's but then when it finishes / fails to solve it comes back to life ?

Dave

[ChrisLX200]

Hmm. That sounds like PinPoint timing out. You can set the length of time it will spend attempting to solve somewhere in the program... Whatever, PinPoint is working with what it thinks is invalid data - either poorly focused, noisey background, wrong image scale, or way off target. I think I covered everything   Thing is, the errors reported can be misleading - I've had some weird ones too like 'Too many stars' with the G4-16000 camera. The 'No Matching Stars' error you have there is a catch-all which may have many causes (as above).

ChrisH

[Davey-T]

Well making a bit of progress tonight, it manages to solve sort of, had a few goes just using three points with varied results, sometimes getting message "Modelmaker not responding"

Also message (see pic') about not liking one of the points and deleting it, then, failed to delete, then, bailing out.

Should I do more points or does it need to do the first three successfully before adding more ?

Dave

[ChrisLX200]

Whoa. 3079" ? That's almost a degree out! I've never seen figures like that. I should disable model clean-up for now as it isn't helping (might be useful for a full 100-point run but not here).

Your 3 base points are all on similar DEC, would be better to vary that, but stay away from the Pole.

You need to find out why your initial pointing accuracy is so far away from where the mount thinks it is pointing. I assume the time is correct? and that daylight savings time = on?

If you leave it to run a model of at least 25 points - assuming it solves - you should get a sensible readout of PA. That may be substantially out. The first three points will always have the largest error but the mount should rapidly home in on where it's really pointing and the RMS error for each additional point will drop rapidly after that. It should be easy to get PA accuracy to <30" after the first run. Remember to clear the model after making a physical adjustment to PA.

ChrisH

 

 

[Davey-T]

OK, managed to complete a 10 point model and had a twiddle with the PA, cleared the model and did a 27 point model reporting an RMS of 9.56, and made an horizon mask.

I still get a message Modelmaker not responding if the solving takes more than about 10sec's but it comes back to life once it's solved

Just left it doing a 10min sub on M31 as it's opposite the Moon.

Seems we might have a winner, thanks for all your help Chris.

Dave

[ChrisLX200]

Well at least it's getting there   It should really solve inside of 1-2 seconds but I guess that's dependant on the computer to a some extent. Once you have it aligned you can reduce the solve expansion % again - tweak until it starts to fail - that will reduce the the area PinPoint searches for a match and so the process will be quicker. My best RMS recently is 2.5" but it's usually around 3.5" (this with a refractor and good focuser = little or no flexure). With the ODK12 I'm lucky to get 5" and more likely 10"RMS, but remember that's just the all-sky pointing error (which is only as good as the points that were included). Because the mount's tracking and pointing performance is totally dependant on the quality of the mount model it is wise to give it the best/fullest dataset you can, so use the full 100-point model if possible. You may even see the RMS increase as you add more points beyond the 27 you used but that's OK - the mount will have a better idea of the flexure in the system and how it should compensate for it. I know some owners use a very limited number of points and get an apparently low RMS value, but that will not translate to a consistently better performance.

 

ChrisH

[Davey-T]

Thanks Chris, managed a 15min sub with good FWHM and flatness measurements in Maxim.

It was solving in a few seconds generally with the occasional glitch, computer is new and fast, SSD, don't know if wireless slows it down ?

Still on the Meade tripod ATM must get the pier bolted down and run 100 point model, still getting the "not responding" message occasionally in MM, will have a browse around the 10Micron forum to see if anyone has had the same thing, doesn't seem to affect the end result but may be a bit annoying on a 100 point model as sometimes it dies for about 30sec's.

Must get FocusMax working, must get cables routed so much to do as well as solar imaging 

Found a guy on the 10Micron forum who has updated wirelessly so will give it a go later.

Dave

[Davey-T]

Hi Chris, just done a 50 point model and ended up with RMS below 10, PA is still a little out but going to give it a go at it's first image, have to be HA with the Moon.

Still got about four MM not responding messages but it still solved after a longer time than the "normal" time of  few seconds.

I only really have a lowish horizon to the East so North, South and west restricted to high Dec', given the restricted sky do I still need to try and squeeze 100 points in or does it depend on the density of the ones I can get in, will it limit the achievable RMS not having access to a full sky view ?

Hows the solar animation going ?

Dave

[ChrisLX200]

All you can do is cover the sky you can see, I've got some big obstructions too in the NEast and nowhere can I get all the way down to the horizon. Another tip is to sort your list of points in order of azimuth (I used Excel for that), that way the mount only does a few meridian flips as it crosses South, so the model run consists of shorter slews and thus makes for a quicker job. It won't limit the apparent RMS having a limited view but the RMS alone won't tell you the whole story either. You will get good tracking and pointing only in the areas you have modelled - but you won't be imaging outside of those areas anyway so it doesn't matter

Finished the animation but I'm not sure it was worth the effort! Just 30 frames in total, there's got to be an easier way of doing this..

ChrisH

 

[Davey-T]

I was getting along OK last night, deleted the model had a twiddle with he PA just started a new model and it clouded over about 1.30am so had to pack it in.

On the animation front this one I did a couple of days ago is only 13 frames about 5 min's apart, it didn't get much in the way of processing as I was heading outside 

Dave

 

 

[Davey-T]

Had more time last night, did an 85 point model (no room for any more) with RMS of less than 10, should I set the RMS target lower ? will it improve the model ?

Checked the PA and it's spot on.

Single HA subs no calibration just stretched in PS, no flattener / reducer as I didn't want to introduce any variables.

Took a few test images and stars appear to go eggy in E / W direction, not sure if it's tracking too fast or too slow ?

Dave

                                                              600 seconds WO110FLT QSI683

                                                                               900 seconds

                                                                               900 seconds

 

[Davey-T]

Just been checking a few things while waiting for the forecast clear sky to appear.

I haven't got round to updating the firmware yet so still can't use the GPS so still setting the time manually but it wont let me set the correct time,  no matter what I put in it resets itself 10 sec's slow, can't figure out where the hell it's getting the time from, it's 10 sec's behind the computer time which is the same as the time shown on my phone.

Dave

[ChrisLX200]

That's odd. I use Per's driver to continuously update time and barrometric pressure - but that's my computer time (which is maintained using an NTP Daemon). It needs to be accurate to a few millisecond..

ChrisH

[Davey-T]

8 minutes ago, ChrisLX200 said:

That's odd. I use Per's driver to continuously update time and barrometric pressure - but that's my computer time (which is maintained using an NTP Daemon). It needs to be accurate to a few millisecond..

ChrisH

Does that work via satellite or does it need continuous internet connection for the server ?

Dave