One hour Veil with tilt?

[tooth_dr]

Firstly I'm very happy with how much data I can gather with the 268M, the new 3.5nm Ha filter and the Tak Epsilon 180ed.

I've been adjusting the spacing and the stars look smaller, but this isnt consistent across the frame.  Does this look like tilt?  I dont have CCDI but I do have PI, is it possible to measure tilt etc in PI?

 

Cheers

Adam

[tooth_dr]

[tooth_dr]

Single sub of different target below - the above stack of the Veil was rotated 180deg in processing.

[tooth_dr]

Another single sub and corners/centre.

 

[almcl]

Don't have PI or CCDI, so ran the three through ASTAP's CCD Inspector function and got these.  First and last definitely look like tilt, but the middle one doesn't seem nearly as far out.  Not sure if this helps? 

 

[tooth_dr]

1 minute ago, almcl said:

Don't have PI or CCDI, so ran the three through ASTAP's CCD Inspector function and got these.  First and last definitely look like tilt, but the middle one doesn't seem nearly as far out.  Not sure if this helps? 

 

Thank you.  Visually the middle one is bit out too, good to see statistics though. 

[almcl]

Not quite sure what the statistics mean in terms of adjustment.

Having spent an interesting hour or so adjusting the tilt plate on my asi 2600 while it was on the scope, I managed to completely reverse a corner  tilt of similar magnitude with about a half turn of the adjuster. May have to resort to feeler gauges to assess the amount of movement necessary.

Edited July 24, 2021 by almcl

[symmetal]

Hi Adam,

Sorry, only just noticed this thread. I've run them through CCDI and all three indicate significant tilt top to bottom. Either the top or bottom edge stars are out of focus depending on the image orientation. The star shapes themselves don't look too different top to bottom so I would say the spacing distance is consistant top to bottom, it's primarily a focus issue which would imply the main tilt is before the flattener, as if the focuser is drooping. Here are the CCI images. I had to rotate some of the 3D images about the x axis to get them to fit in the window. The absolute values of FWHM shown are incorrect as CCDI made a guess on image scale, as it couldn't get it from a fits header. The relative values give a good indication. I have no idea either what the percentage (compared to what?) tilt values mean either as the third image has over 100% tilt  Again just treat it as an indication. Tilt below 10% is generally a reasonable value.

As I've also found when using autofocus, if there's tilt it tends to determine the best focus is near an edge which makes the difference in focus across the image more pronounced.

Veil

NGC281

IC5146

Alan

[tooth_dr]

52 minutes ago, symmetal said:

Hi Adam,

Sorry, only just noticed this thread. I've run them through CCDI and all three indicate significant tilt top to bottom. Either the top or bottom edge stars are out of focus depending on the image orientation. The star shapes themselves don't look too different top to bottom so I would say the spacing distance is consistant top to bottom, it's primarily a focus issue which would imply the main tilt is before the flattener, as if the focuser is drooping. Here are the CCI images. I had to rotate some of the 3D images about the x axis to get them to fit in the window. The absolute values of FWHM shown are incorrect as CCDI made a guess on image scale, as it couldn't get it from a fits header. The relative values give a good indication. I have no idea either what the percentage (compared to what?) tilt values mean either as the third image has over 100% tilt  Again just treat it as an indication. Tilt below 10% is generally a reasonable value.

As I've also found when using autofocus, if there's tilt it tends to determine the best focus is near an edge which makes the difference in focus across the image more pronounced.

Alan

Hi Alan

Thanks for doing this.  I've no idea exactly what the figures mean, but it looks pretty bad!

I had a very short imaging session last night, and spent some more time tweaking the autofocus routine - I added more points and this seems to have helped.  I dont want to upload any more images as I dont want to take up any more of your time, but the tilting looked less obvious (but still there).  I'll look into getting a trial version of CCDI, as it seems quite costly 😫

The camera is screwed onto the flattener, which in turn is screwed into the focuser tube.  Is it possible the camera chip could be tilted?

 

[tooth_dr]

Is it possible to adjust the front plate of the QHY268M, I thought it was but cant seem to find any info.

[symmetal]

2 hours ago, tooth_dr said:

The camera is screwed onto the flattener, which in turn is screwed into the focuser tube.  Is it possible the camera chip could be tilted?

That's often the case, but I've found in that situation with a flattener you end up with the star shapes being different from one corner to another and altering the spacing distance you can make one corner or side look good, but the other corner/side has elongated or other misshapen stars. You aren't able to get all four corners looking similar.

It's hard to tell exactly from your images but your star shapes are fairly round in all corners and not too misshapen, just well out of focus on one side compared to the other which is what made me think the tilt is not between the flattener and the sensor.

Oops, I've just noticed your using the Tak Epsilon 180ED which doesn't have a field flattener as such as it's corrected internally for a flat field, so camera spacing is not critical I would have thought as long as you can focus. This is why your stars have the same shape across the image which is good. The focus difference does indicate that the camera sensor is not square on to the optical axis so tilt correction is required.

2 hours ago, tooth_dr said:

Is it possible to adjust the front plate of the QHY268M, I thought it was but cant seem to find any info.

From pictures it looks like the front plate of the camera is adjustable for tilt by the three sets of push/pull screws around the front plate. These are difficult or impossible to adjust without removing the camera each time. Check that the tilt plate is parallel to the camera body and tight and hasn't got some tilt set by mistake.

There seems to be two different front plates on the images for the camera on QHYs website. The first one has what look to be tilt correction screws as I've indicated, while the second has three knurled screws which may just do camera rotation. Which one do you have or are both included?

I should have mentioned on the CCDI 3D plot the curvature shown in the plot doesn't mean the sensor appears to have that curvature compared to a flat plane. The best focus/shape stars are the lowest Z value so at the bottom of the plot, coloured black, while poorer stars are plotted with an increasing Z value, and brighter colour, depending on how bad they are.  Stars in front or behind the focus plane due to tilt will give an increasing Z value hence the bowl shape plots you tend to get.

Also the line just away from the centre cross on the CCDI curvature plot is an indication of collimation error, the further from the centre the worse it is. For my refractors this line coincides with the centre cross or is a pixel away. Yours is showing a larger error which may be significant. Here's the CCDI manual page on collimation.

Quote

Precise Collimation

Unlike collimation procedures of the past, CCD Inspector provides a revolutionary new way to collimate a compound-optics telescope.

Collimation makes a huge difference in the quality of image and resolution that can be achieved. With CCDInspector, a collimation error of 10 arcseconds can produce as much as 1 arcseconds increase in FWHM of a star. This means that a good 3.0 arcsecond FWHM image can become a 2.0 arcsecond image with proper collimation! 

By measuring the exact displacement of the optical center from the physical center of the imaging configuration, CCDInspector is capable of detecting the smallest collimation errors with your CCD still attached to the telescope, and with telescope well focused! This is the best possible way to collimate, since:

•  The optical train is not disturbed by removing an eyepiece and replacing the camera after collimation
   
•  Focus position will need only minor adjustments to get to best focus after collimation is completed
   
•  What's more, the collimation can occur right on, or very near-by to the field you will be imaging. This may be the best way to collimate a telescopes with significant mirror flop
   
•  Since collimation is done on hundreds of stars, there's no need to re-center anything after adjusting collimation: just take the next image, and keep adjusting.

With CCDInspector version 2.1.0, we introduce another breakthrough innovation: Single Defocused Star Collimation method. This method will help you achieve the same spectacular results as with the original collimation routine, but using only one bright star centered in the field of view.

 

Differences between Multi-Star and Single Defocused Star collimation methods

• Multi-star method can be used in perfect focus. This is beneficial when the focuser mechanism can introduce changes in the collimation of the system, or if you want to do a quick collimation check without losing focus
• Single-star method can be used with a single bright star. There's always a bright star somewhere in the sky, but sometimes, there is not a rich star cluster, such as is required by the multi-star method
• Multi-star method doesn't require re-centering after making collimation adjustments, Single-star method does
• Single-star method is not as sensitive to tracking errors because a bright defocused star can be used with 1 second or less exposure. Multi-star method can require much longer exposures that are subject to tracking errors

Which method should I use?

That depends on what's most convenient at the time. Decide based the differences listed above. Either method will help you achieve excellent collimation.

[NOTE: CCDInspector also provides a simplified Collimation Viewer display that is even easier to use for real-time collimation]

To facilitate achieving perfect collimation, CCD Inspector provides two sets of crosshair on the screen: the larger one marking the physical center of the chip, and the smaller one marking the current optical center of collimation. By making collimation adjustments to move the small crosshair to the physical center of the chip, the best collimation is achieved. To help, CCD Inspector also provides a numeric reading at the top left of the Curvature Map Viewer window that shows the distance between the two crosshair in arcseconds or pixels (based on the choice made in the main CCD Inspector window).

The procedure to adjust collimation is fairly simple:

1. Find a reasonably crowded star field of reasonably evenly-spread stars, with no extremely bright stars in the field of view. Just point somewhere near the Milky Way and you'll likely see many hundreds of stars in one shot.

2. Start CCD Inspector, and open the Real-Time Curvature Map window from Real-Time/Curvature Map... menu.

3. Start taking exposures using the camera's main chip. The following are some guidelines, actual settings will be different for each individual setup:

•  30 to 60 seconds exposure should be sufficient for best S/N and for collecting enough stars to measure field curvature.  A longer exposure may be necessary with really long focal length telescopes, or if a well-populated star field is not available.  The goal is to see at least 100 or more reasonably bright, but not saturated or bloomed, stars in the shot. More stars are better. A large concentrated star collection anywhere in the image will distort the measurement (such as a globular cluster, for example).
   
•  Bin the chip 1x1 for best results with shorter focal length systems. FWHM of an average star in the field should be around 2 pixels or more. If it's less, the FWHM measurement will not be as precise, resulting in a less sensitive curvature computation. If the seeing is exceptionally good or the image is really undersampled, it would be desirable to defocus the image a bit to achieve the minimum of 2 pixels FWHM.
   
• If your image scale is 0.6 arcsecs/pix
   
•  CCD Inspector will detect conditions when there aren't enough stars in the field of view, or they are not evenly spread out, and will display a message indicating this may not result in an accurate measurement. If you receive this message, it's usually best to stop, and adjust the parameters to get more stars in the field of view (increase exposure, or move the telescope).

4. After each exposure, check the curvature map to see how much, and in what direction to adjust collimation.  Just like in standard collimation techniques, the closer you get to perfect, the smaller the adjustments needed.

5. Make an adjustment, take the next image, and repeat with step 4. Keep doing this until the crosshair at the center overlap, and the error is indicated as only a few arcseconds. At this point, you're in excellent collimation!

Some additional hints to help with collimation:

•    To help see the direction and distance of the collimation error, right click on the field curvature map. From the pop-up menu , select Magnify Crosshair

You can also zoom-in further from the pop-up menu to see the error magnified. You may need to scroll the image to the center to see the crosshair, if the zoomed-in image does not fit completely on the screen.

•  Remove as much  tilt from the camera as possible to achieve best collimation. Use screw-in connectors exclusively, and if your focuser has collimation adjustments, use them to square the camera to the optical train. You can use the CCD Inspector Tilt measurements in the real-time curvature map view to help with doing this.
   
•  Make sure the telescope is cooled down and equalized with ambient air: if not, you will see strange artifacts in the curvature maps that are due to air currents as the OTA cools.

Alan

Edited August 2, 2021 by symmetal

[tooth_dr]

7 hours ago, symmetal said:

Oops, I've just noticed your using the Tak Epsilon 180ED which doesn't have a field flattener as such as it's corrected internally for a flat field, so camera spacing is not critical I would have thought as long as you can focus

Thanks again Alan. Going to take time to read this properly later. The Tak doesn’t have a built-in internal flattener, it’s screwed into the focus tube but is removed for collimation.  It’s spacing is in fact is very critical, with a quoted spacing of 55.2mm required! 

[symmetal]

7 hours ago, tooth_dr said:

Thanks again Alan. Going to take time to read this properly later. The Tak doesn’t have a built-in internal flattener, it’s screwed into the focus tube but is removed for collimation.  It’s spacing is in fact is very critical, with a quoted spacing of 55.2mm required! 

I'll have to retract the earlier Oops. 🤭 I thought it was like the Newtonian equivilant of their refractors. All the other info should still apply though. 🙂

Alan

[Xiga]

@tooth_dr as you know i have the same camera and i'm also in the process of trying to get my stars right as well. Hoping to get my differential flexure problems sorted for good on the next clear night, then i'll be ditching the Baader Clicklock and moving to an all threaded connection to hopefully eliminate any tilt issues. FWIW,I also put some of my test session Ha subs through ASTAP (as a cheap alternative to CCDI) and it told me i had a bit of tilt, usually less than 10% but i found the variability between subs to be huge, even with subsequent subs, which makes no sense to me. One would be ok, and then suddenly the next would indicate quite a lot of tilt, so i don't plan on using ASTAP to determine my tilt issues now. I'm just going to go by eye and draw a line under it when i find i hit a point where it doesn't overly bother me anymore lol. 

On your setup though, i had an idea. For the 3 sets of data you captured above, do any of them include subs from pre and post meridian flip? If any do, then you could take the subs from immediately before and after the flip, register them with each other and then compare. If it is tilt due to the focuser in any way, then the tilt should move to the opposite side and you should clearly see a big difference in star shapes in each corner. Conversely, if there is no major change, then it is probably more like a backfocus, focus, or collimation issue. Something to try out at least. 

[bottletopburly]

@tooth_dr this was posted on YouTube thought may be of use to you looking to check tilt  the chap is also a member on here can’t remember his forum name though , the op also has another video of making device 

 

[blinky]

I watched that vid and it looks good, how does he build it though? I saw the video where he makes it but what's the 2" tube that he uses to hold the camera? I wonder if any Astro manufacturers would consider selling one of these jigs?

[markse68]

you could use a v- block instead of the tube- 2 thin vees cut from ply joined together with a spacer would do

Mark

ps it was @Corpze who posted this in here

https://stargazerslounge.com/topic/380736-diy-optical-rig-for-aligning-cameras-and-collimators/

 

Edited August 7, 2021 by markse68

[Corpze]

1 hour ago, markse68 said:

you could use a v- block instead of the tube- 2 thin vees cut from ply joined together with a spacer would do

Mark

ps it was @Corpze who posted this in here

https://stargazerslounge.com/topic/380736-diy-optical-rig-for-aligning-cameras-and-collimators/

 

And the second video of how it works is just uploaded on my YT channel 😃

 

ps, I just noticed that it was that video you had linked to 😛 

Edited August 7, 2021 by Corpze
Cross minded

[tooth_dr]

Thanks for this info guys.  If I can indeed check for camera tilt this would be fantastic and eliminate/correct that aspect.  I wont be getting time for a couple of weeks to do this but will update once I get a look at it.

[tooth_dr]

Took a few weeks to get around to it, but have now adjusted the tilt, and have had a few clear nights.  I've attached the tif and jpegs.  @symmetal Alan would you mind running them through CCDI?

Image 1: Veil with Epslion 180ED and ZWO2600MC

 

Image 2: Veil with APM 105/650 and QHY268M - I think the spacing is a little too far out with the scope, does anyone agree?

 

 

 

 

TAK_ZWO2600MC.tif APM_QHY268M.tif

[symmetal]

Here they are Adam and I think you'll be happy. 😊

Compared to your original postings they're a big improvement. Both showing a similar symmetrical minimum fall off towards the corners. Note the difference in the min/max FWHM figures (2.5" to 3.1") compared to the originals (8" to 18"). CCDI didn't know the image scale from the tifs and assumed 1.26"/pixel so the stated FWHM are using that figure. The 3D views are symmetrical and pretty flat too.

The APM is showing a tiny amount of residual tilt in the Y axis but is really not worth worrying about. Your image shows a bit of tracking error as all the stars are a little oval all over. I would say the spacing is correct. The top left and bottom right corner stars are showing a very slight coma like spread which is not really a spacing issue.

The TAK indicates that your focuser chose to focus near the top left corner rather than nearer the centre so the centre stars are showing as being more out of focus compared to the edges. The difference is only from 3.88" to 4.57" FWHM, so slight but just noticible on the image.

As I've mentioned before I don't know what the percentage curvature figures refer to and the manual doesn't really say either. 🤔

The tilt jigs are certainly doing their stuff. 😀

Alan

[tooth_dr]

1 hour ago, symmetal said:

Here they are Adam and I think you'll be happy. 😊

Compared to your original postings they're a big improvement. Both showing a similar symmetrical minimum fall off towards the corners. Note the difference in the min/max FWHM figures (2.5" to 3.1") compared to the originals (8" to 18"). CCDI didn't know the image scale from the tifs and assumed 1.26"/pixel so the stated FWHM are using that figure. The 3D views are symmetrical and pretty flat too.

The APM is showing a tiny amount of residual tilt in the Y axis but is really not worth worrying about. Your image shows a bit of tracking error as all the stars are a little oval all over. I would say the spacing is correct. The top left and bottom right corner stars are showing a very slight coma like spread which is not really a spacing issue.

The TAK indicates that your focuser chose to focus near the top left corner rather than nearer the centre so the centre stars are showing as being more out of focus compared to the edges. The difference is only from 3.88" to 4.57" FWHM, so slight but just noticible on the image.

As I've mentioned before I don't know what the percentage curvature figures refer to and the manual doesn't really say either. 🤔

The tilt jigs are certainly doing their stuff. 😀

 

 

 

Alan

Thats amazing Alan, thanks for this detailed analysis.

Do you think it is possible that there is some differential flexure between the two scopes?  My guider is attached to the TAK, and I have the APM on an adjustable JTD alignment saddle.  The blurb says it's good for more than my requirements at 9kg, but it could be a weak link causing the over stars as I tracked at around 0.4-0.5" through the night and my imaging scale is around 1.55"/pxiel.

[symmetal]

26 minutes ago, tooth_dr said:

Thats amazing Alan, thanks for this detailed analysis.

Do you think it is possible that there is some differential flexure between the two scopes?  My guider is attached to the TAK, and I have the APM on an adjustable JTD alignment saddle.  The blurb says it's good for more than my requirements at 9kg, but it could be a weak link causing the over stars as I tracked at around 0.4-0.5" through the night and my imaging scale is around 1.55"/pxiel.

No problem Adam. 😀 I didn't realize both images were taken at the same time on one mount. So it does look like there is some flexure between the scopes as the TAK stars are nicely round while the APM's oval stars are similar and all in the same direction. Things are never easy are they. 😦

Alan

[Astro_mark_c]

I've noticed I have this problem too, what is tilt and how do you fix it? I have a redcat and 183 mc pro.

Edited October 2, 2021 by Sidecontrol

[symmetal]

20 hours ago, Sidecontrol said:

I've noticed I have this problem too, what is tilt and how do you fix it? I have a redcat and 183 mc pro.

Tilt means the camera sensor is not square on to optical axis of the scope, and so all the edges of the image can't be in focus at the same time with one focuser position. The larger the sensor the more critical eliminating tilt becomes. For smaller sensors it may not be an issue. This topic shows a simple method of indicating any tilt, much easier than the trial and error method analysing star images. If the camera doesn't have  tilt adjustment screws on the front plate of the camera, to alter the angle of the camera to the optical axis, then an external tilt adjuster can be inserted in the imaging train to do the same thing. If buying one make sure you get one with the right side threads to match your setup.

Your 183 doesn't have built in tilt adjustment so either an M48 to M48 tilt aduster to fit on the rear of the Redcat, along with the M48 to M42 adapter supplied with the ZWO camera to fit it to the camera, or a M48 to M42 tilt adjuster where it can fit directly to the camera. With the Redcat a specified distance is required between the scope and camera to enable infinity focus to be reached so I would use an M48 to M48 tilt adjuster and M48 spacers as required reducing to M42 at the camera. If you need one and can afford it, the Gerd Neumann M48 tilt adjuster is worth getting as it's much easier to adjust when fitted in the imaging train, than the cheaper push/pull types.

Alan