Tpoint calibration

[Somnium]

I have been trying to use the skyx for the first time with the new paramount MX and am really having trouble with calibration. the mount is connected properly and so is the camera, it homes fine and the pointing seems to work pretty well but clearly my polar alignment is a little off or the scope shifts slightly in its rings when slewing because i don't actually get the target in the FOV. having said that, the scope looks like it is pointing right at the object in the sky. with that done, i try to do a model through automatic calibration and it keeps on returning error 733 the index is out of range. i have set it to look at the max area (10 FOV's) but it just doesnt work. occasionally it will find its location and add the pointing data. the strangest thing is that i did have a run that worked perfectly, i used the information to make PA corrections and now i can't pick up any locations again. i didnt make any changes between the runs ... the QHY9 does have a flipped image, could this be impacting it? i can't find a setting to address this.
autoguider calibration has been an issue too. ... :confused2:

[Atmos]

I am pretty sure that that error doesn't always mean that it hasn't been able to plate solve the image, more so that it is too far outside of the accepted pointing location. I am using an EQ6 Pro, I find I either get this when it is too far off of polar alignment or sometimes after having done a successful model and "Accurate Polar Alignment" routine.

I have personally found that it works best if after doing a model and proposed corrections:
• I tell T-Point that I have finished my calibration data.
• Take an image, run image link and then resync to a portable mount (keeps Super Model) and begins a recalibration model.

If I end up having an issue after that I park the mount, double check that the mount is in correct park position (syncing sometimes moves it), wipe all calibration data and start again. Takes an extra 90 seconds.

Your MX should be far more TSX friendly than it is to EQMOD :)

[PRejto]

Hi Aiden,

I think you need to use the "All Sky" page for Image Linking. To use it though you need to download the All Sky data base which is quite large (a gig). I think that will get you going.

I can really relate to what you are going through with this type of mount and getting one's head around the initial polar alignment. I actually resorted to taking my camera off. The All Sky solution was designed just for this type of problem.

Peter

[gregbradley]

Make sure your image scale is correct.

Unclick use all sky image link for automated pointing runs in the Automated Pointing Run window. For some reason that caused my Tpoint to not plate solve and it took me a while to discover that. As Peter points out you must need to download a database for that to work. I should do that too. It could be handy. I am downloading it now. Its 1.6gb and is at the SB Support site.

I find its best to use the Take Image and Link Photo to get the actual image scale and enter that into the automated pointing run box. If you can't run that image link TPoint wont work either.

Make sure you have the binning of the camera set correctly and the image scale for that binning. I was using 5 second 2x2 binning and then doubled the image scale that was reported for 1x1 binning. Make sure you are using the luminance filter. Make sure its a long enough exposure to give enough stars.

Try that and see how it goes. I find no trouble with my Proline 16803 as it gives lots of stars but found it hit and miss with the Trius 694 (mostly miss!).

Greg.

[Atmos]

I tend to use 1 second 3x3 bin with a QHY9 and don't have any problems at all with Image Link. I have actually tested to see whether it is a problem with the plate solving but I discovered what was happening is that T-Point is saying that the point is too far from where it should be.
What I did try doing once was every time it came up with that error I would pause the automated run, take an image and manually run Image Link. A lot of the time it would find the location within two seconds, if I went to manually add it to the pointing model it would come up with a warning message asking me if I wanted to add a point that was so far out from correct so as to not cause problems with the calibration data (it assuming that it is just a statistical anomaly and shouldn't be added).

[Somnium]

thanks for the help guys, i was running EQMOD and astrotortilla before and it was much more user friendly (that is saying something) i did run take an image and link it, it was picking up the rotation and scale easily enough. also with the plate solving, i think worked but it was too far out of the location it thought it was pointing. i was going to ask if there was all sky plate solving in the software, i am downloading it now. hopefully this sorts it out, then i need to deal with the autoguiding issue. one day i will be able to take a picture :)

[Somnium]

BTW, don't be surprised if there are a few more questions about this over the next few days ... :sadeyes:

[Somnium]

just adding to this, it doesnt appear to be a PA issue, i am taking 4 min subs with no star trails (i could never do that with my EQ6), the date and time are also correct. it must be flexure in the scope and possibly the mount is not 100% level. i did have to adjust for the cement ground which was not perfect so it might still be a little out

[frolinmod]

Please go into TheSkyX->Tools->Preferences->Advanced and set the Bad pointing sample criterion(degrees) to a large value such as 10. That will eliminate the 733 index is out of range error. Also, a new TheSkyX daily build was released last night. If you haven't already fetched it, please do so now.

[gregbradley]

Also once you've installed the All Sky database and check the box in the automated tpoint run to use it then that may also solve your problem.

I was unaware you had to download a special all sky database for the all-sky option to work and found it was preventing me from getting plate solves with tpoint runs and it took me a few days to discover this a few months ago!

Its a complex program but the more I use it the more I like it. Its a world and its worth exploring.

Greg.

[Somnium]

thanks, i will give it a go and report back tonight

[Somnium]

Last night i had the camera working for a couple of hours before it decided to pack up and during that time i was able to capture a decent set of calibration points, around 50. using all sky and expanding the bad plate solving error sorted that out. But i am getting some really strange results. I have attached an image of the output. in some instances the star appears to be on the other side of the sky to the mount location. the date time and location are correct, it is definitely pointing south, and when the telescope points to a target it looks like it is pointing at the right spot so i cant understand the strange data that i am getting. i attempted a super model but got an error as well (seen in the attached screen grab).

i think i am losing my mind with this

[frolinmod]

Those are certainly terrible pointing results. Did you click on supermodel afterwards? Also, did you do a quick polar alignment before all this? That should get you much better pointing right off the bat.

The sample data appearing to be "on the other side of the sky" is just Tpoint's way of noting the side of the mount, beyond the pole and the like.

I know it's a stupid question, but you did go into "Telescope ->Tools ->Bisque TCS ->Utilities ->Hemisphere Setup" and set the mount for Southern Hemisphere operations, right? You probably couldn't even get this far if you hadn't.

[Atmos]

I am wondering if this is a few false hits with the All Sky attempt? It could potentially take one really false hit to push the RMS way out and have it look further afield for all subsequent plate solves. I still haven't gotten around to trying the All Sky thing yet though.

[Somnium]

no question is stupid enough, but yes i did set it to the southern hemisphere. i had my EQ6 set up in the same location before and polar aligned well. i lined up the pier to that and set up the mount. the first polar alignment run i did was moderately successful and i made the minor tweaks to correct the PA. the results now give me multiple degrees out of PA which is strange because i was running 4 min subs as a test and they were coming out perfectly so i dont think the PA is out much at all. i did click on super model after the calibration run.

[Somnium]

possibly. it looks like the first half of the images were giving me really strange results, like 120 degrees out and 12 hours out, the others were more reasonable, a few degrees or a couple of hours out (still more than i think it actually is) even when i eliminate these from the data set i cant egt the super model to run ...

is it possible that it is picking up false stars, should i reduce the sensitivity and give it another go?

[gregbradley]

Did you sync the mount on a star first?

It looks like the mount is not sync'd or is set to the wrong hemisphere. Those errors are massive. 5533 arc sec circle for the errors is way huge. It should be more like 200 or less depending on your mount.

I had this happen with the AP1600 when it was not starting from a known position ie. in AP's case a known park position. When I parked it properly first then manually got it centre a bright star and then sync the mount with that it was all good.

With PMX I had it happen that even though the Sky X said southern hemisphere it wasn't. I had to click on set to southern hemisphere again in TCS and it took a little while and then a lot of the numbers changed and it was fine after that. It was trying to point to northern coordinates.

PMX etc mounts still need to be synced. So check that out.

Greg.

[Somnium]

Hi Greg, apart from homing the mount, i did not perform a sync. the guys from bisque say that ideally the Tpoint model will make this unnecessary. i will sync it next time i am down there (unfortunately i only get to troubleshoot these issues once a week), still doesn't really account for the issues though... the hemisphere is definitely set to south in TCS (attached)

[Bassnut]

Make sure you have daylight saving ticked in sky. I didn't so when the obs PC time changed, pointing was screwed.

[frolinmod]

When Patrick says that, he means it in an ideal world in theory sort of way. That's the way he thinks it should work. And he's right, it really should work that way. But in practice I don't think TheSkyX and/or the mount is able to handle that at the moment. I think it still needs the sync. I could be wrong of course!

The only time you need to do a sync is immediately before doing a full calibration. An image link and sync (take photo, plate solve it, and sync on the plate solved image) is most accurate. You normally don't need to and don't want to do a sync at any other time. Just the one sync before a full calibration.

[gregbradley]

I am pretty sure the manual says to sync and a Tpoint model definitely does not get rid of the need to do that.

Although Chris on this site mentioned he homes and then does a plate solve and used that to sync the mount. I am not sure how you use the plate solve to sync as I tried that the other night and not sure how you get the Sky X to recognise the plate solve as the sync.

Greg.

[Somnium]

thanks guys, i will clear the data, perform the sync and then rerun the Tpoint model and see how that goes. to me there seems to be a bigger issue at hand, but we will see.

[Atmos]

You go to the "Telescope" tab above the "Camera" tab. On the top left there is a drop down menu "Start Up", about half way down there is an option "Synchronise". When you click that it asks if you want to Sync to the current model, start a new calibration run or start a recalibration run.

[Somnium]

i don't get that option ... i get the below

[gregbradley]

Yeah I did that but it was showing a star to sync to which I don't think I did in fact sync to.

Perhaps it was right but but it seemed incorrect.

Greg.

[Atmos]

After you have done an Image Solve there should be a grey box indicating your FOV and what not. You need to select that box first so that you're syncing to the Image Solve and not the star/location you slew to.

[frolinmod]

Ahem, if all else fails you read the User Guide. Open the PDF and search for "Image Link and Sync". It's much easier to do than it is to describe. It goes basically like this. Point the scope anywhere you like. The home position after homing is as good as any. Take a photo. Go into Tools->Image Link, make sure the "Search->Show Photo" checkbox is checked. Image link the photo. Go on over to the virtual sky display. Click anywhere within the displayed image linked photo overlapped on the virtual sky until you see "Linked Photo" as the target (not any other star or object). Then click on "Telescope ->Startup ->Synchronize". Make sure the sync confirmation dialog says you're syncing on "Linked Image" and not something else.

[PRejto]

I'm permanently set up so that is probably a factor. However, I have not synched my mount for a very long time. I just start a new automated T-Point run and then finish it. This probably wouldn't work too well with a portable setup though it might with all sky image linking. If I understand Patrick correctly the finished model essentially is a "synch."

Aidan,

I saw your post at SB but recently can't find it. Did SB ever respond to you?

Peter

[gregbradley]

Cheers for that.



Thanks for that explanation. That could be handy especially if I move my mount to my dark site temporarily and have to do a sync. Its tricky without a finder scope attached to the mount and a narrow FOV!

Greg.

[gregbradley]

I think the problem with that is that you will not get successful plate solves without the all sky link database when you are too far off. If not synced then the mount could be sending the scope to a location too far off for plate solve to work easily. Perhaps the all sky database and having that checked solves that problem. I haven't used that yet in that way but I have used it for a plate solve of an existing image.

Greg.

[Somnium]

I have not had a response since posting the tsx file

[PRejto]

Aiden,

I'd suggest that you bump your post over at SB. Things do go missing when they fall too far behind. The danger seems especially large if you post close to the weekend.

Peter

[Somnium]

will do, but i think i will wait until the weekend when i get an opportunity to run another model after a sync. i will see if that fixes the issue.

[frolinmod]

Since the advent of all-sky image link, the only thing I still find a finder scope to be useful for is the quick polar alignment. That's the one where you home the scope, tell TheSkyX to slew to a any known bright star, then physically move the tripod and/or ALT/AZ adjusters to center the star.

[gregbradley]

Yes I can see it will be very handy to use the all sky image link. It would have saved me a few hours a few weeks ago! Its very difficult to manually line up the telescope with no finder scope or image link with a small sensor even only at 1159mm focal length. It took a while!

Greg.

[Somnium]

so i am currently running another model and it is becoming clear what the issue is but i have no idea what is causing it. the model is running really well, i homed the mount and then synced to a plate solve. i tested my syncing by pointing to a target, NGC 1365 and it landed smack bang in the middle of my image, great ... now i am running my tpoint model and all the RA for the scope are almost exactly 12 hours out from the star RA (attached). to be clear, this is set to southern hemisphere in TCS and the location is correct. also pointing seems fine so i have no idea what this means.

[gregbradley]

I have had it happen where I was getting the mount trying to point as if it were set to Northern Hemisphere yet the TCS said it was set to the Southern Hemisphere.

I clicked the set to Southern Hemisphere again anyway and after a short lag a lot of the numbers it displayed changed.

It was fine after that.

Sky X occasionally tends to be Northern Hemisphere biased.

Greg.

[Somnium]

talking on the bisque forums it seems that the figures are mechanical not stellar ... which means that if you are meridian flipped then they will be 12 hours out so all good. i am finally seeing the power of this mount. i was testing the pointing last night, swapping between many targets and each one landed dead in the middle of the FOV, amazing, i have never had anything like that before. just need to sort out my guiding issues (shouldn't be a major issue) and test my PE :D