Electronic Polar Alignment Scope

[kens]

No, not the PoleMaster although it does apparently now work in the Southern Hemisphere.
But inspired by the idea I rummaged around in my spares box where I had two polar scopes. One for the Super Polaris and the other for the SW EQ6.
Both are pretty much identical with the exception of the screw thread for the eyepiece, slightly different focal lengths and therefore slightly different reticles. In fact the SP reticle is epoch 1980 as well I think.
And of course both are pretty much useless especially under light polluted skies.
But tinkering about I discovered that the thread for the screw on ring for the month dial is exactly the same as that on my DBK21AU04.
But to make it work I'd have to cut off the tube where the eyepiece screws in to expose those threads. So which scope to sacrifice? The SW scope drew the short straw. I'm keen to keep the SP in as close to original condition as possible and its polar scope is decidedly better quality than the SW.
I hacksawed off the tube then filed it down smooth. I had a couple of goes at it as at first it was still too long to focus.
I tested the focus first on the Dandenong TV towers and then on the moon. When the stars finally came out I refined the focus further. To do this I was using a camera tripod so I had to point close to the pole to reduce the star movement.
Next step was to take a pic and plate solve. This told me FOV (2x1.5 degrees) and pixel size (11 arc-seconds) from which I could calculate a 105mm focal length.
Checking against Stellarium I could see that the dimmest stars visible in the image were around mag 8. Then looking at the FOV around the pole in Stellarium I could see that the stars there are mag 7 to 8 so I should be able to get a usable image.
The theory is that I can take an image and plate solve. Then rotate the RA axis 180 degrees and do the same again. Taking the midpoint should account for any collimation error and tell me where the mount is pointing. Then I can adjust accordingly and iterate as required.

[Stefan Buda]

Yes, but wouldn't it be easier to use the afocal method and not having to cut the polar scope? You only need a cheap Jaycar lens and a 3Dprinted adaptor sleeve.

http://www.thingiverse.com/thing:1144293
http://www.thingiverse.com/thing:1102513

[multiweb]

Sounds good. Regarding LP I actually find it easier to polar align when there are less stars visible in the polar scope. I have the Losmandy one on my G11. The reticule is probably outdated but it still gets me very close and I know how and where to offset the two stars I'm looking for.

[el_draco]

Great idea, especially if you have a buggered up neck and live way South. Be interested in whether anyone has adapted a G11 polar scope like this.

Rom

[kens]

This way is free and collimation should be more reliable.
It's not like the polar scope is useful. It is unnecessarily accurate for visual with a lot of effort to use it, or it is not accurate enough for imaging.
I usually jump straight to drift alignment but the plate solve method can potentially get the PA down to 30". This experiment should tell me one way or the other.
Besides, I've got another one still in the spares box.

[RobF]

Be interested to hear updates Ken. Was cogitating over something similar recently. Key thing is to make it worthwhile time-wise for the benefit in an initial set up (I'm thinking astrophotography).

[Stefan Buda]

Recently I made an adaptor for a Losmandy polar scope and just a few minutes ago a published the design on Thingiverse:

http://www.thingiverse.com/thing:1363374

[el_draco]

So this adapter slips over the polar scope in situ?

[Stefan Buda]

Yes it does and it relies on the elasticity of the plastic to get a grip.
The Thingiverse site lets you rotate the model and look inside it without having to download it. If you do that then you'll understand how it works.

[kens]

So tonight I started to use the scope to polar align experimentally. Plate solving was good to get close to the pole. But then I hit a snag as I was looking at Stellarium with JNOW coordinates whilst astrometry was giving the position in J2000. Until I realised what was going on I was puzzled as to why I appeared to be at the pole visually in Stellarium while astrometry wsa telling me I wasn't. The difference is about 4 arc-minute which is signficant. It seems it would be easiest to just use the nearest asterism and visually centre on the pole in relation to that. BQ Octanis is almost exactly 10 arc-minutes from the pole and was easily visible in the scope. I was about to try that when the clouds rolled in.

[kens]

Finally got to do a serious test last night even though it was a bit rough and ready. I just plonked the tripod down and took a pic and plate solved it. I needed about 8 seconds exposure to get a solvable image. The astrometry program gives results in J2000 so I plugged the results in Stellarium and it gave me back the Azimuth and Altitude. So all I had to do was adjust Azimuth to be 180 and Altitude to be my Latitude. I ended up getting close to the SCP. However, I knew that was illusory as the next test was to see how much collimation error there was. To do this I took an image with the mount rotated 6 hours west and 6 hours east, Plate-solving each one. Rather than calculate the results outside with the mozzies I then retired indoors for some analysis.
Using the plate-solved images with significant objects highlighted I could plot the centre of each image (all 3) onto a single picture. Then joining the dots gives a circle whose centre is where the mount is really pointing. This coincides with the centre of the line joining the East and West images but is useful for future reference.
Next step is to get an image with the mount actually pointing at the SCP as a reference for alignment. Basically the idea is to be able to say if BQ Oct and HIP42708 are horizontal on the image then the image should be centred at...
Alternatively, any ideas for collimation are most welcome

[OICURMT]

For reference.

https://www.youtube.com/watch?v=5L6IvfievRE

[bojan]

It seems that PoleMaster software works only with PoleMaster hardware.. and that hardware is just USB camera with some specific ID that enables PoleMaster software.

Is there equivalent application, that will worlk with any camera?

[Eden]

Writing an ASCOM-based application to work with any camera would be a simple enough task, but the QHY PoleMaster is covered by Chinese patent. Aside from that, PoleMaster and the supplied software costs around the same as a low-end guiding camera.

The software needs a couple of minor tweaks for usage at the SCP, but why hack something together and disregard patent when you could be supporting a company which is delivering affordable imaging products to the Australian market?

[kens]

That's more or less what I'm working/playing on. Thus far I have put together a spreadsheet to calculate Alt and Az from a plate solved image taken along the polar axis. Much easier than using Stellarium to do the calcs. I'm taking two measurements 180deg apart in RA so as to calculate the midpoint to account for collimation error.
From alt and az it's trivial to work out the adjustments needed for PA.
In due course it should not be hard to write up a program to do the calcs and read the FITS file header for the correct time when the image was taken, for more accurate calculation of alt and az.

[bojan]

Brett, I don't think US$268 (https://www.optcorp.com/qhy-polemaster-eq-mount-polar-alignment-camera.html)is affordable at all. Not for me, and I am sure I am not alone here.
And, we are not professionals... IMO, DIY and free sharing of the knowledge is always preffered approach for us amateurs, and I wish it to be adopted much wider in our community, and actually encouraged - it is beneficial for our AU economy in the long run.

Of course - if someone is able and willing to pay for particular off-the shelf commercial product for whatever reason, I have no problem with that.. I bought my EQ6 because I didn't have time/tools to build my own.

[bojan]

I think what PoleMaster does is much simpler, no plate solving is involved - actually, IMO the app first follows the alignment procedure for polar finder (by using the changing position of one star near the pole and rotating the mount for 180° it is easy to determine where the mount axis is pointing at - this is exactly between the two positions of the star on the screen, before and after rotation).
The next step is leaving the star to drift for a while without tracking.. Which will determine the position of the pole on the screen.
All that is left then for app to do is to tell you where to place the image of the star on the screen to align the mount.
The math involved in this process should be very simple... Definitely worth the effort.

Of course, the sensitivity of the camera and the size of the polar finder lens will be the issue.
The modified webcams (with long exposure times) will be a preferred option.

[kens]

Speaking of patents: http://www.google.com/patents/CN204269938U
As I read it the principal claim is the mechanical configuration integrated with the electronic alt-az sensors. But I am no patent attorney.
The software could be based on this work by Rick Kellogg. http://www.syracuse-astro.org/pdf/2012_August_R_Kellogg_Electronic_Po lar_Alignment_Scope.pdf
I need to study Kellogg's paper as I might be able to do something similar using BQ Oct and HIP 42708

[bojan]

I am looking again at the method I described earlier (believing it was what PoleMaster is doing), and the patent..
In patent they are mentioning all sorts of things, and this reading is additionally confused by bad translation into English.. I am sure even patent attorney would have been puzzled (or maybe not..).

Anyway, I am not sure the existence of patent is relevant for amateur activity at all – firstly, because the aligning method of two rotating axles is nothing new and basically a common sense issue, and in public domain for a long time. Secondly, no one here is trying to copy, manufacture and/or sell the product as competitor.

Finding the point where the polar axis is pointed at is accomplished with rotation by 180° all right.. mechanical setting circles can be used for better accuracy, or perhaps the bubble level on the DEC axis .

To find the celestial pole, however, it is better to use couple of stars, located around the pole (it is easy to recognize the asterism around sigma Octantis, and all those stars should be used, perhaps others at the other side of the pole as well).
They should be marked on the screen (and their positions stored) at their starting positions and then the tracking should be stopped. After some time the image on the screen will rotate… and the same stars should be marked again as end positions.
Software needs to place the joining lines between stars start and stop positions, calculate the mid points and then place the lines through those points perpendicular to joining lines. The intersection between all perpendicular lines is true celestial pole.

The vector defined by the point where polar axis is aimed (start) and the abovementioned intersection point (end) determines the amount and direction (negative) to move the star image (by adjusting the mount elevation and azimuth screws) from current marked position to the new one (indicated by software).

[Eden]

Bojan,

You're right, we're not professionals and I apologize if I came across as though I'm against the DIY aspect. I am all for the concept of home-brew and respect that it is and always has been an integral part of the amateur movement.

[glend]

I have to agree with Bojan, I am concerned about the steps taken to 'productise' virtually all aspects of life that can possibly be covered by a broad stroke Patent. Much like 'open system' software projects like ASCOM and others, there is no reason that an open system alignment routine that can work with any sort of webcam could not be developed. Many imagers already have guide cameras that woud be suitable, and while app features like DARV are great for drift alignment there is perhaps a need for a quick dark site field tool that allows tripod setup with minimal time and effort. So Open Source this please.

[kens]

With my camera with its small sensor and the 105mm FL tube I find the FOV is only 2x1.5 degrees so Sigma Oct is out of view. But BQ Oct and HIP42708 (HD 90105) form a nice little asterism along with some other fainter stars as you can see in the photo I posted earlier

[kens]

At last a clear enough night for testing. ! was getting good images that plate solved quickly. My spreadsheet showed I was a few degrees out initially so I stared by manhandling the tripod into position based on the coordinates from a single image- that is uncorrected for any collimation error.
Once within 1 degree of the pole I started to rotate the scope 180 degrees to get two images. Once I started to adjust in the right direction (a small spreadsheet error reversed the azimuth so 179 was showing as 181) I started to get quite close. Ultimately I got to within a couple of minutes on each axis. To put that in perspective, on the SP mount that was about an 80th of a turn on the alt bolt and a 32nd of a turn on the azimuth bolts.
Not bad for just plonking the tripod down on the ground unlevelled.
I wasn't set up to do a drift alignment to confirm o that's for another day.
Attached are the pics taken 180 degrees apart plus the composite of the two showing the circle traced by the scope as it rotates. Sigma Octans is the bright star to the left of the composite.

[skysurfer]

Well, I have a Vixen SP with an excellent polar scope.
Polar alignment is very simple as long I photograph subs of max. 4min at 600mm FL.

I just center BQ Oct (NOT Sigma called 'Polaris Australis' (Australian Pole star ?) !) in the polar scope image. It is only 10' apart from the true pole.
When I need more accuracy then the drift method is the only option as the template is expired.

In the North I put Alpha UMi (they should call it 'Polaris Europeanis') 40' (not 50' as stated on the outdated template) at RA 3:0 apart from the true pole using the template as there is no closer star visible in the polar scope.

[bojan]

This looks very promising!
Both my transportable mounts (EQ3 and LDX75) have sensitive bubble levels and provisions for attaching the compass (mounted on a 1m aluminium 12x12mm "L" profile), so the initial setup always brings the pole close enough for fine adjustment with polar finder… which is a real pain (literary!).
As for the mechanical side of the procedure, if the mount is moved very often, to prevent the inevitable damage to screw threads it is a good idea to go further and replace the screws with ones with larger diameter and finer pitch (I used M10 with hex head instead of original M8, see here (http://www.iceinspace.com.au/forum/showpost.php?p=399967&postcount=14)).
Some people even modified the elevation screw like this (with double pitch, coarse and fine) (http://www.zvjezdarnica.com/forum/index.php?action=dlattach;topic=652 1.0;attach=13950;image), to smooth-up the process.
http://www.3ashop.it/articolo.asp?ID=4030

In the case of EQ6, it is a good idea to go even further and do something like this guy did:
https://stargazerslounge.com/topic/187034-eq6-latitude-lug-modification-or-how-to-avoid-bent-bolts/
or this one:
https://alancatovic.wordpress.com/2015/11/29/modifikacija-postolja/

[kens]

It is promising indeed. I like your idea of a template showing the position of the asterism when you are properly aligned. I started to work on it but the geometry is doing my head in.
I put together an animated GIF showing the position of the asterism at HA of 6, 12 and 18h when close to polar alignment, with the collimation error included. It was certainly not intuitive to me. The red line shows the BQ Oct to HD90105 orientation and the yellow line the BQ Oct to SCP orientation.
What I'll need to do next is to take a series of images at intervals of 30deg around the RA axis after polar aligning just to see what they look like. That might help me work out the maths needed to calculate the orientation.
I'm still somewhat surprised at the magnitude of the collimation error. The centre of the camera FOV is 172 pixels from the RA axis at a scale of 11 arc-sec per pixel: that's half a degree!
Also attaching the spreadsheet I use to do the calculations - uncommented of course :) it converts coords coming from plate solving into Alt-Az

[kens]

Me too. I kept the Vixen polar scope as it is much better made than the Skywatcher one. Only problem is that the reticule is for 1985 or thereabouts. You can get closer to the pole by centering halfway between BQ Oct and the faint pair of stars between it and HD90105.
The reason I'm doing this project is that I find using the polar scope a pain in the back and other parts of my anatomy, it is hard to see the Sigma Oct asterism in light polluted skies, it can be useful if the East and West horizons are not easily visible (made worse by LP) and most of all its a bit of fun and seems to be getting some folks quite animated.

[kens]

Just had an AHA! moment when I realised (after drawing up some sketches) that when the mount is rotated in RA then the image will rotate about some point offset from the image centre when it is not collimated. So the trick is to find that point first with the polar scope and camera assembled. It could be done in daylight or with a bright star. e.g. Take a long exposure while rotating to scribe an arc with the star. Use the arc to determine the centre of rotation. I'm hoping that with PHD I can then set a bookmark on that point.
All you need to do then is to position the SCP on that point. Or somehow fashion an overlay that shows the circles that BQ Oct and HD90105 would trace when properly polar aligned. Those circles would be centred on the cente of rotation already found. Place the stars on their respective circles and bingo!

[bojan]

Yep, exactly right!

[kens]

I made up an image that illustrates how it works. It overlays the images taken 180 degrees apart. One with white stars, the other with yellow stars.
The red dot is the centre of rotation of the camera which is also the RA Axis. It just happens to be close to the SCP in this case.
So to establish the centre of rotation one could start by placing BQ Oct such that when the mount is rotated, its image does not move. It could be any star but best to choose one near the pole or at least near the right altitude so you don't need to put the mount far out of alignment.
The green circles show where BQ Oct and HD90105 would be located if the mount is polar aligned. They are centred on the red dot. The radius of each circle is simply the distance of each from the SCP.
Hopefully, it is apparent from the graphic that no matter the orientation of the mount, if BQ Oct and HD90105 are placed on their green circle then the centre of rotation of both image and RA axis are aligned to the SCP.
Now the trick is to find some software that allows a custom overlay on the image. PHD2 allows bookmarks and SharpCap has a movable, rotatable reticule. Any other possibilities?

[bojan]

Maybe Al's (Sheeney) special reticle could help here ?
Provided the camera wibndow doesnt move during the alignment session.

[kens]

My camera is WDM compatible so I can view it in VLC and add and position an overlay. However the overlay must be wholly within the video boundaries so I would need a cropped one to fit when it needs to be near the edge

[kens]

So last I tried with an overlay in VLC. It would have worked had I sized the overlay to have circles with diameter TWICE the distance of each star from the pole.
Also, finding the centre of rotation was a little tricky. It's important to pick a star near the pole so it doesn't drift away.
Attached snapshot from VLC shows the rotation of BQ oct.You can see a faint semicircle
Another snapshot with the (correct) alignment overlay. This is a mockup as VLC snapshots exclude the overlay

[bojan]

I found couple of apps, could be very useful for markings on screen:

http://epic-pen.com/

reticles, rulers:
http://www.iceinspace.com.au/forum/showthread.php?t=21798&highlight=Al%27s+reticule
https://technet.microsoft.com/en-sg/sysinternals/bb897434
http://davesastrotools.weebly.com/download.html
http://www.arulerforwindows.com/index.html

This one below can even overlay images:
http://www.theimagingsource.com/en_US/support/downloads/details/icmeasure/ (also measures, and stacks)

http://imageoverlayutility.bitbucket.org/ (this is demo, and doesnt save .. otherwise fully functional)

[kens]

Thanks Bojan I'll look into those.
Another thought occurs to me that using the method of finding the centre of rotation, one could position the camera anywhere including the normal guide camera position. The camera would need to be pointing close to the pole and the dec axis could be used to collimate, so the centre of rotation is in the centre of the camera FOV. Cone error adjustment may also be required.
The the overlay can be centred on the FOV and the BQ Oct asterism used to align as before.

[bojan]

Yes, those were my thoughts as well.
BTW, there is an old thread about similar method, someone posted a question after more than a year...
http://www.iceinspace.com.au/forum/showthread.php?t=128890

[AndrewJ]

Gday Ken


Not required, ( but it can be tweaked by the same method ;)).
Ie even if there is a little cone error, "If you dont move the DEC axle" the circle described on the sensor is still purely a factor of where the RA axle is pointing .
I had originally used this method for setting the orthogonality of my LX200 OTA ( by only using 2 spots 180deg apart ), ref
http://www.cloudynights.com/topic/518399-meade-lx-200-10-gps-goto/?p=6894272
In this case, i used a terrestrial target to find the centre of rotation,
then tweaked the OTA to get the "circle" centred on the sensor.
Then i thought it could also be used in reverse as a polar aligning method by simply using the mount controls to centre circles on the sensor ( no matter where they were ) with the SCP asterism.

Andrew

[kens]

Thanks Andrew. I was mainly thinking of how to align the centre of rotation with the centre of the image, specifically along the "y-axis".
Funny how I started out looking to do something similar to a Polemaster using a polar scope and am now ending up not needing any special device at all, other than something that can draw and place circles.
Interestingly, the icmeasure tool from TheImagingSource looks very suitable for that purpose and obviates any need to align the centre of rotation, other than to bring it into the FOV.

[AndrewJ]

Gday Ken

Yep, but then you realise that for purely polar aligning, that's not reqd.
As long as you can get stars to form circles "somewhere" on your sensor ( ie you are close to DEC = +90, and your cone error is small )
all you need to do is centre the circle over the pole astrism by using the mount controls.

Andrew

[kens]

Indeed it is unnecessary - it's just that I just like symmetry :) I had a go just now getting the centre of rotation well within the FOV and it was remarkably easy. I was focusing on an insulator on the pole outside and adjusting declination till the circle was as small as I could make it.
I used icmeasure to first mark the start point of the feature I wanted to measure and the end point after rotating 180 degrees. With icmeasure it's easy to draw a circle to fit the two points and find its centre.

[AndrewJ]

Gday Ken

Then you should buy a fork mount instead of a GEM:lol::lol::lol:

Andrew

[kens]

So tonight I tried aligning using the camera mounted just like a guidescope. At first I had some trouble getting it aligned with the RA axis but eventually got the hang of it. The trick was to start with small rotations to make a small arc while the collimation point was off screen then as it adjusts in you can use larger arcs. I used ICMeasure to record three points on the arc zooming right in to get pixel level accuracy. The I drew a circle connecting the three points. Over that I overlaid three concentric circles of 114, 92 and 75 pixels diameter corresponding to the distances of BQ Oct, HD99685 and TYC-9518-405-1 from the pole.
I ahd to cheat a bit and platesolve to get near the pole. I had just plonked the tripod down on the ground and was over 5 deg out as it turns out and too far to home in without some help.
Once I could see the polar asterism I adjusted alt and az to place the stars on their correct circles.
Attached images are the zoomed in screen shot of the final result and the actual image with overlays. At this focal length each pixel is 11.1 arcseconds so in theory I am within 1 arminute of the pole.. Unfortunately I'm not set up right now to test that with a drift alignment

[bojan]

Most likely drift alignment (and the method you described is actually drift alignment procedure, only applied to the stars near pole) will show the same result..
On both my transportable mounts I have precision bubble level and magnetic compass attached. They are both "null-ed" at the stage your mount is at the moment. by just using them I am always near the pole, which is OK for short focal lengths (less than 300mm) and exposures up to a minute.

[kens]

Tonight I used my regular guide scope following the same technique as before. Then once I finished the polar alignment I did a PHD2 drift to see how good it was. Net result was around 4" dec drift over 15 minutes. This corresponds to less than 2 arc minutes PA error and I suspect some of the drift was in fact flexure plus I was hampered by some light cloud cover. A good result seeing as I didn't take any special care with the alignment. And it was nice to have a confirmation of the accuracy of the method.

[AndrewJ]

Gday Ken

Just for fun, i saw Altronics are doing a special on vari focus CS lenses.
Just picked up a 2.8-12 unit for $32
http://www.altronics.com.au/p/s8955-cop-security-2.8-12mm-vari-focal-manual-iris-lens/

Best thing is being a CS lens, it screws directly onto my DBK41 with no adapters reqd.
It has a built in focusser and also can go from wide view to telephoto mode.
Haven't tried it under the stars yet, but it will be interesting to see what it can pick up in wide mode vs narrow mode, as you could start in wide mode to ensure stars stay on the sensor, then zoom in as you get closer??
Now, when will it be clear in Melb again???

Andrew