TSX - Plate solves with ZWO 1600MM-c often failing

[g__day]

I am getting back into astrophotography after a break of several years. Rebuilding an astro lab computer and testing all the focusers, ZWO cameras, mount, dome control etc.


I am finding using The SkyX Professional Edition with camera add on that a 30 second shot with a 1600MM-c on a 9.25" SCT is very hit or miss trying to plate solve. In the past when I used a Canon 40D plate solves where correct 9.8 times out of 10. With the ZWO 1600MM-c my plate solves are abysmal.



Can folks give me ideas what parameters I should be setting in the SkyX to get a decent chance of a plate solve?


Link to a solved image - saying image scale is 0.66 arc seconds per pixel (must be 2x2 binned by the resolution - though I didn't ask it to do that???):


https://scontent-syd2-1.xx.fbcdn.net/v/t1.0-9/93323719_3674342262640916_602761513 2509536256_n.jpg?_nc_cat=111&_nc_sid=110474&_nc_ohc=B_OQvYUglHwAX__az8u&_nc_ht=scontent-syd2-1.xx&oh=0b61c034e9351a31cec073f3753b4b60&oe=5EBF8BEB



Link to a noisy (-15C normal shot):


https://scontent-syd2-1.xx.fbcdn.net/v/t1.0-9/94231911_3674342372640905_344880228 9932828672_n.jpg?_nc_cat=102&_nc_sid=110474&_nc_ohc=L7bhbWPmOnIAX9RE9Cx&_nc_ht=scontent-syd2-1.xx&oh=755e0bbba0cbda68d630591646e575fe&oe=5EC116F2

[brisen]

Hi Matthew

I use the Sky X as well but with a different camera, but think I can see the issue with why you are having issues.

The image tolerance is set at 1% - way too low and if there is any margin of error for it here it is minimal. Try setting this at 10% and this will allow some more tolerance. This is what I am using and it isn't giving grief. I am using 2 sec images to link without an issue. I think you can cut the 30 sec images down a lot and this will make linking quicker as well.

It is definitely binning 2 x 2 which is not an issue for linking, the setting of this is the Setup Tab of the Image Link Window - there is an Automated Image Link Settings button in the left hand column, second from the bottom of the column. Click this and the settings will open. Binning is in there and can be set when the camera is connected. It also allows you to set the exposure time and image scale (unbinned) and the binning and filter to use for automated links. The search FOV is 2 by default and needs to be a bit higher. Also increase the retries near each target to 2 - default is 0.

In the Image Link window go to setup and also check the Search Parameters - I have mine set to 4 fields of view, image scale tolerance I have set to 0.89 s/ pixel (which is the resolution of the system I use).

Looking at the FOV from your setup this shouldn't run into issues with linking images.

Hope this helps. Once its setup it generally cooperates well.

Brian

[g__day]

Many thanks Brian - will try all of these things! What are the sign posts of success? What duration shot and how many stars should be found roughly?



By the way were is image tolerance set - or is it a derived parameter?

[brisen]

Hi Matthew

From memory it will plate solve with about 8 stars as a minimum. I am running 1 second exposures and getting it to solve easily. I was imaging last night and was using a 1 second exposure with about 10 stars showing up on the image and was successfully solving.

A quick approach to know if it is working is to do a closed loop slew (CLS) - select a target and tell it to CLS. This should trigger the mount to slew to the target and take an image, this will then attempt to solve and if it does it will trigger another slew to center the target and take an image to confirm. If it fails to solve, it will try an error message and fail. If this happens, I just increase the exposure time a bit and have another go at the process.

I just remembered that it is necessary to use the same binning when doing the CLS as what is in the image link settings. I use 2x2 for T point runs and focusing, but use 1 x 1 for imaging and CLS. It has a fit and will fail if the resolution is not correct for the CLS.

Brian

[g__day]

Made a fair bit of progress with my settings – and did a much better initial alignment of my scope on three stars. So using 120 Gain on 3-4 second shots binned 2×2 at a known image scale of 0.66 arc seconds per pixel I say very interesting behaviour.


All sky solves generally nailed it in 2-3 seconds, general solves failed most often after 2+ minutes on 7-8 FOV searches with 3 retries.
Is this saying that initially the object was too far from the where the telescope thought it was pointing? The first few closed loop slews failed. My last attempt after improving the mount’s handcontroller’s pointing model and the main scopes focus nailed a Closed Loop Slew to M20 dead centre of frame.


So if All of sky Plate Solves generally work whilst the search tab Plate Solves don’t generally work can someone please suggest some ways of fixing it?


Thanks, Matthew

[brisen]

By search tab I assume you mean Find? The one where it brings up all the details of the given object.

Slew will slew to the object but won’t solve it, closed loop slew as you indicate will easily centre an onject even if the alignment is out. It sounds like pointing accuracy is partly to blame for the failures as if the scope is pointing somewhere outside the parameters it will fail. CLS doesn’t have this constraint to the same level.

I purchased T point a couple of years ago and have used it to significant improve pointing accuracy and with supermodel I am routinely getting under 5 arc second pointing accuracy which improved the accuracy of closed loop slews for multiple nights as well as improving guiding.

Brian

[g__day]

Brian,


The Closed loop slews were failing Image find from the plate solve. Raw pointing accuracy is about +/- 2 arc minutes with this set up. An old Tpoint 130 star model might have been impacting things - but all this used to work really well with the less sensitive Canon 40D - so I would ponder something not paramaterised correctly by me.