gnomus

Yes, smaller chip and, yes, I was happy until I started reading your thread and started pixel peeping 

 ......

John

Oh dear. I really am sorry.

No, I use a QSI690 (which was returned to QSI to get the Sony Glow fixed). Hoping I don't have to return my Tak!

Regards

John

I believe that that chip is physically smaller than the Kodak, correct?  Whether or not you return your Tak is, of course, a matter for you.  Are you happy with what it is doing in your corners?  Have you tried the 5 second pointing at the zenith test?

Just an idea. Maybe rotate your camera 90 degrees and repeat the test. If it's always the same bottom right hand corner that shows the most distortion then would that indicate sensor tilt?

I've been taking a closer look at some of my earlier images taken with the FSQ85. Coincidentally, I have elongated stars in the bottom right hand corner, more so that the other 3 corners. The other corners show some elongation but in different directions - radial I believe. This does not appear to be a PA issue. E-mail sent to Mr King for advice!

Regards

John

I had perfectly acceptably (I suppose nothing is perfect) round stars with the same camera on my ED80.  I've spent enough time on this particular scope and it is going back.  I hope that I had a bad copy, but you and I are not the only people who have had radial distortion is our corners.  Are you using a KAF 8300 sensor?  

Yup, that excludes PA conclusively.

I'm really sorry you're having this hassle.

Olly

PS I'd send that crop to the supplier.

Already done. The supplier is getting a new FSQ85 in for me.

No need to be sorry ..... although ..... on reflection .... this is all your fault!!!!

I hope the new one works....

You shouldn't have to focus on the one third lines with this scope.

Having been absolutely delighted by mine (which is now Sara's) I have often extolled its virtures but both Steves have had issues and this is not good. It's a great shame. My subsequent Tak experiences (using my own very second hand FSQ106N and Tom's similar instrument) have also been entirely positive (not that this proves anything. I just mention it.)

At the resolution available on here I find it hard to tell whether or not the corners are worse in the 5 min subs. What do you think, Gnomus?

Olly

PS I had my first Mesu nightmare last night but it turned out to be my fault. I had somehow managed to reverse the sign in the encoder steps setup menu. (Or it might have been the cat. I bet she did it!) The Mesu continues to 'just work' in a way that seems to be beyond the powers of much of the kit we use, as Steve Steppenwolf says.

Thanks Olly.  I have taken a crop from the same region of the 5 second and 300 second exposure and upscaled it a little.  I hope this helps.  There are clearly less stars on the 5 second image.  On the 300 second image the stars are brighter and a bit more 'swollen'.  Nevertheless, the ovoids look to be identical in proportion.  I've circled a few just to give an idea.  What do you think?

Late to this thread, but I was worried about my fsq106ed (new Q) and oval stars!

I decided to make a proper test using CCD inspector!

The advice is to take 10 or so 20 sec sub's..taken from a fairly rich field devoid of globular's and nebula...unguided .....stack them together in CCD inspector and then look at the curvature n tilt.

The logic being that guiding etc will distort the stars.

This should eradicate the sensor and the focuser tilt as being the cause of oval stars which it did for me, though I did also use a DTI to measure all sorts of movement, including the Captains wheel effect, which turned out to be negligible.

Aside from this, there are a set of grub screws on the side of the focuser that you can adjust to just stiffen up the focuser (on the 106 at least). Though friend of mine has loads of wobble on his FSQ 106 and has not been able to remove it by this means.

Frustrating issue with Taks !

Ray

Takahashi state:

As on the FSQ-106ED, the built-in field flattener, perfectly [my emphasis] designed and manufactured, gives an almost homogeneous flat field of 44mm diameter. At prime focus, at the edge of field (22mm from the optical axis), stars remain pinpoints [my emphasis again] ...

My view is that if the scope does not function as promised then it should be returned.  

Here's an off-the-wall thought. Were your latest batch of test images taken with the Ha filter? 

Have you taken a look at the spot diagrams published by Takahashi for the FSQ-85ED scope?   http://www.takahashi-europe.com/en/FSQ-85ED.optics.spots.htm

The spot diagrams show that stars at the Ha wavelength will become distinctly ovoid in shape 14mm away from the centre of the sensor. The corners of a KAF-8300 sized sensor are 11mm from the sensor so maybe it will be similarly affected.  The Ha wavelength is particularly affected by this ovoidness.  Maybe that is why your luminance image of Bode's Nebula looked almost perfect but the Ha images show the issue.

Perhaps the scope is performing more or less within the tolerances of its design and is unable to illuminate the 8300 sensor without distortion at the Ha wavelength?

Mark

Thank you for bringng this to my attention. These images were taken with the Lum filter though - I thought 5 secs through an Ha filter might be a bit optimistic!

When I had the same issue with my Tak 85, it was suggested to me that I should try offset focusing. Now, I was prepared to consider this on a camera lens or cheaper telescope but not on an instrument at this price point - that was not what I bought in to!

I expected a flat field on an 8300 sensor but that's not what I got so it went back. Perhaps my expectations were too high (?) but then, so too was the promise!

Offset focusing might be worth a try for you and if you are happy with the results - and many owners must be - then you have the solution. At least you can now rule out focus slop leaving you with CAA adjustment, sensor tilt or the plain fact that the field is not perfectly flat for your sensor size.

I hope this turns out to be an adjustment for you because the potential FOV of the Tak 85 makes it a very desirable instrument.

Sent from my iPhone from somewhere dark .....

Thanks Steve. Like you I take the view that I was promised a flat field and I should not have to resort to such strategies. In any event I run an autofocus routine through SG Pro and I am not at all sure that I would know how to offset focus with this set up.

I have seen good reviews of the WO 71 and the TS Optics 65. These also give a wide FOV I believe. Does anyone know if these scopes are capable of giving round stars?

Steve is using a Tak 85, which is a Petzval design. There is no reducer or flattener in use 9outside the Petsval) and the design means that, AFAIK, spacing isn't that critical.

Having said that, it'd be worth playing about with spacing.

Is it worth playing with focus as well? If the focus is set bang-on in the centre of the imaging circle then there is a chance that it will be out at the edge, especially with a large sensor. After all, no focal plane is perfectly flat. What about setting the focus bang-on about 20-55% out from the centre? That might give the best balance of focus and minimize distortions? :icon_scratch:

And I am not using the Tak reducer.

I finally managed to get a brief window when the rain stopped and the skies cleared.  I did the test suggested by Sara - that is I took a series of pictures with the telescope pointed at the zenith.  Since the camera and focuser are hanging straight down, these images should not be affected by any play in the focuser or CAA (Camera Angle Adjuster).  In an attempt to rule out a significant polar alignment error, I took exposures of different lengths, from 5 seconds to 300 seconds.  If my problem was field rotation due to polar misalignment then one would expect ovoids not to be visible at all in the really short exposures and for any ovoids to get longer with increased exposure time.  Clearly there will be less stars in the shorter exposures.  The result is that I have ovoids present in at least 3 of the four corners of all images.  To my mind the worst affected area in all images is the bottom right corner.  I do not think that the extent of the ovoid distortion changes significantly from one image to the next (there may be some differences in 'brightness' due to the exposure times and my processing (which is just some stretching)).  

The distortions seem to form in a circular type pattern around the centre (all centre stars are round in my images). 

The final image is an exposure of M31 taken with the same camera and filter wheel but through my ED80.  The stars in the ED80 corners all look acceptable round to me.   

Incidentally, each of the corner panels is cropped in such a way that it shows just under 1000 pixels square - so that is a reasonable chunk of the sensor.

Any ideas as to what might be causing these issues with the Tak?  Is this coma?

(I suggest clicking on the image to get the full-size view):

Hi John

Thank you for those links. I do have the CAA in my train. I am going to try Sara's imaging at the zenith idea. That will tell me whether or not this is 'slop' of one form or another or field rotation. As I said, I thought I saw a lot of elongation in the corners of images of other users on Astrobin who had my scope/camera combo. My problem is that I don't really know what I should be prepared to accept when it comes to corner elongation. I must say that at this price point I was not expecting the corners to be worse than I had before.

I have contacted the supplier. I don't think that I should be having to dissolve glue and be making these fine adjustments on a brand new scope - assuming of course that that is what the problem turns out to be.

If you decide to take your own issues further, then it would be interesting to hear what Mr King has to say. I use a different supplier.

Steve

I wouldn't try to collimate it! It was the focus adjustment that I think you might be able to tweak painlessly if you find evidence of sag. If it's out of collimation in needs to go back.

Olly

Apologies - I meant adjusting the focuser (I thought that was called collimation too). I had early warning of my ineptitude in engineering matters when I tried adjusting the reticle in my polar scope using the three little grub screws. The advice on the internet (just like the advice in the old Hayes manuals) usually begins, "Simply adjust...." It always tickles me that phrase.

Steve

Good ideas from Mark and Sara. A meridian flip on a southern target would turn the scope over but the target would be the same, so you'd see if the distortions reversed themselves as they ought to do if it were the focuser. (I'd give it ten minutes to re-sag, so to speak...) This is certainly becoming a candidate. If the sensor were not tilted in its previous scope then why would it be tilted in this one? If this doesn't happen then we'd be looking at slight mis-collimation which is not unknown.

I know that Per has discussed the Tak focuser and one place to ask for advice on adjusting is Texas Nautical Repair, who are very expert on all things Tak.

Olly

Thanks.  I will give that a go when next we get clear(-ish) skies.  I'm not at all sure that I am up for adjusting the collimation on a scope of this cost.  I would probably make the slight mis-collimation even worse!  In any event, I have only had it for a few weeks - it should have come to me properly set up.  

I spent a little time this morning looking through images on Astrobin that had been taken using the FSQ 85 and the Atik 383L.  It was a little difficult to tell because most images are resized to less than 2000 pixels on the long end.  Nevertheless, I do have to say that I found quite a few images (maybe the majority that I looked at) that appeared to show corner elongation similar to mine.  I began to wonder if I would find these issues if I were using a camera with a smaller sensor.  Is the design of the Tak such that it is particularly prone to orthogonality issues?  On the ED80 the camera was attached in the crudest possible way - a 2" nose piece and the brass compression ring in a Moonlite focuser.  I find it difficult to believe that this beats the out-of-the-box Tak set up.  

It is all a little frustrating since I bought this thing so that I could do some imaging, not spend what few decent nights I get 'testing'.   

If there's any thought that it may be a focuser tilt issue, then how about a test run of focusing at zenith? That way all of the weight will be pulling the focuser directly downwards and not to one side or another.

I can certainly give it a go.

I had a look back at some of my earlier images.  I can see the egginess in my Elephant Nebula shots from early October when I had pretty good PA (these are Tak 85, Atik 383 and Mesu) 20 minutes:

Link to FIT: https://www.dropbox.com/s/bbj874jco0ci58d/Elephant_1200sec_1x1_HA_frame5_c.xisf?dl=0

My Bodes Nebula image from a month earlier is significantly better to my eyes.  There is a slight issue top right, but I would not have noticed I think unless I had been looking for it.  The corners are certainly not as bad as on the Tak.  This is ED80, Atik 383 and CGEM (probably not as well polar aligned as the Mesu is now) - 10 minutes: 

Link to FIT: https://www.dropbox.com/s/wofjg4u8klv44va/Bodes%20Nebula_600sec_1x1_L_frame4.fit?dl=0

It seems that the Mesu is performing better than the CGEM so this must be down to some combination of Tak 85 and Atik 383L.  If this is a sensor tilt issue, then I might have expected to see it on the ED80 but I didn't.  If this is a focuser tilt issue on the Tak then I am not sure that I know how to fix this.  

I used the camera with my ED80 and got round stars in the corners - and this was using unscrewed connections. I believe steppenwolf eventually returned his Tak.

Steve

Nah, it's knackered Steve. I'll take it off your hands for £50 and that's me cutting me own froat.....

Was that your attempt at a Cockney accent Stephen?

Sounds like this could descend into one of those imaging train ? tilted sensor ? ??????????? threads, best sell it all and go back to EQ6 ED80, much less stressful 

Dave

No need to sell it Dave - if it ain't right it's going back.  (You weren't hoping to pick up a bargain were you?)  

Here's the result of running those two frames through my PixInsight script.  Note that an eccentricity figure of 0.4 or below is usually considered round, so the 1200sec image shows good guiding (round stars at the centre of the image)

StarShape1.JPG

StarShape2.JPG

The analysis seems to confirm what is obvious to the eye.  But I can't really explain what we are seeing.  I don't think field rotation is the explanation.

Mark

Thanks again for looking at these Mark.  I can imagine a circle drawn around the yellow lines in your analysis of my 1200 second exposure (except for perhaps the middle left panel).  The 600 second analysis appears to show random errors.  I'm not sure that your figures for the 600 second exposure marry up with what I am seeing on screen.  If I understand your data correctly, your script is suggesting that my roundest stars are in the middle left panel.  This is not what I see in the processed image.  

I should add that last night was quite horrible.  Visibility was only 4 miles according to 'Cloudy Outside'.  Not only that but temperature and dew point were pretty close.  The equipment and observatory walls were dripping wet again after only an hour or so.  When I looked into the objectives at the end of the session, I wondered if I had my dew heaters turned up high enough.  Certainly the PHD2 image seemed to be out of focus by the end of the exposures.   The neighbours decided to set off a number of fireworks and from time to time firework smoke drifted across the field of view.   I do recall also that PHD2 reported one of those "guide camera unresponsive" messages during the 5 min exposure (although I reconnected everything pretty quickly and the guide trace seemed unaffected).  Finally, one disadvantage of having an observatory is that you can get set up and going too quickly.  I knew I only had limited time last night so whipped the roof off and got going immediately without leaving any time for the gear to cool down.  The 10 min exposure was taken first and started before astronomical darkness. 

Thanks again

Steve

I am not using the reducer.  I believe that the Tak has an inbuilt flattener.  I am using the correct Tak rings as specified in the manual. 

....... The simple test is to take a short exposure through a less restrictive filter (luminance) to see if the distortions are still there. In a short sub rotation won't show. On occasion my PA needs a tweak and a rotation visible in 30 minutes may not be visible in 15 minutes......

.....

BTW, below is a typical Mesu guide trace in AstroArt 5, my eccentric choice of guiding software. The guidescope is the same as yours, so has a 400mm FL and the Lodestar is working in bin 2 so a hefty 8.46 arcsecs per (virtual) pixel. The graph is in pixels. I run 4 second guide subs. I can refine this level of accuracy by working on it on a night by night basis but at the scales at which I'm currently imaging on the mount it would be a waste of useful time to do so.

Thank you. As I said in my probably too lengthy post I also shot a 10 min Luminance sub and I couldn't make out any significant ovoid-osity in the corners. Field rotation it is then.

I will never post any smug 'look-at-how-flat-my-guide-trace-is' images again after your AstroArt screenshot! How are you doing that? BTW, I am now using a 600mm guide scope (the ED80) and it does not seem that it is possible to set my ZWO guide camera to bin - there is an option under the PHD2 brain, but it is greyed out. So I think that gives me 1.29 arcseconds per pixel. I intend improving my PA by looking at the video coming from my guide scope - that will work won't it? I don't need to be looking through the imaging scope do I? (I really don't fancy unscrewing all those Tak adapter rings again.)

Congrats Steve

Last night was pretty poor seeing-wise, but I was able to make some progress (I think).  Firstly, the ED80 did find a guide star and did seem able to produce reasonable guiding - here is a screenshot from PHDLog Viewer:

I think this is reasonably good guiding - certainly miles better than I was getting with my CGEM.  There were long periods when the mount seemed to be tracking well with no need for 'correcting impulses' from PHD2.  Is this level of accuracy about what I should be expecting from the MESU, or should it be better?

I took two exposures - one a 10 minute Lum and the other a 20 minute Ha.  I seemed to have cured the issue of star elongation in the centre of the image.  This is a crop from the centre of the 20 min Ha exposure.  I have re-sized this to 400% (how's that for pixel peeping?) and the stars here look bang on to my eye:

In the corners I think things are improved.  I do not see any significant issues in the 10 minute exposure.  In the 20 minute Ha, there is still some elongation, but this is not all in the same direction like before (as evidenced by Mark's PixInsight script in an earlier posting).  Here is a 100% crop (all of these are from the extreme corners):

You might need the 200% sized crop to see what elongation there is - this is from top right:

To me the elongation seen here appears to be radial - that is, the TR corner goes NW to SE whereas the BR corner goes NE to SW, and so on all the way round.  I wondered if this could be field rotation.  The fact that I don't see elongation in the 10 minute exposure suggests that this is not an optical issue.  I did run the PHD Guiding Assistant and it told me that my Polar Alignment error was 1.9 arcminutes.  Previously it had reported an error of 0.1 arcminutes, so I think I must have shunted something a little when taking the scopes on and off during yesterday morning's reconfiguration.  (Of course I have no idea about the reliability or accuracy of the PHD2 Guiding Assistant.)  PHD2 still seemed happy enough with the 1.9 arcminute error, that is it did not ask me to try to improve this.  Would such a relatively small PA error give rise to field rotation?  Am I correct in thinking that the field rotation would look worse when using a wider field of view (my ED80 was 600 mm f/l compared with the 450 mm for the Tak)?  If that is correct would the reducer, by giving me a wider FOV, make things worse?

My analysis, therefore, is that my earlier problems were due to flexure, either caused by my SkyWatcher guide scope bracket not being 'solid' enough, or because of the place where I had attached my cabling.  Flexure appears not to have been an issue last night, as evidenced by my perfectly round central stars.  The elongation I now have appears consistent with field rotation due to the PA not being as accurate as it could be - probably due to my blundering when redoing my set up yesterday.

I'd be very interested to hear what folks think of these new findings and my analysis of them.  If anyone is interested in looking at the FITS files, these can be found here:

https://www.dropbox.com/s/t5u7921md1r4d8e/Test%20ED80%20Guiding_600sec_1x1_L_frame1.fit?dl=0 

https://www.dropbox.com/s/3wy9ehcpnimjn5g/Test%20ED80%20Guiding_1200sec_1x1_HA_frame1.fit?dl=0

Clear Outside is suggesting that I might be able to get out tonight. If I do I will try out the new set up. If I still get eggs, I will rotate the camera 90 degrees (I think that this will help determine if I have tilt). Are there any other tests that I could usefully run if I still get eggs tonight?

Cheers, Steve, so now I'm wondering whether that argues against flexure on the grounds that most of the correcting applies to RA. On the other hand a flex in Dec would throw up a spurious command. Perhaps it doesn't tell us anything. I might be thinking about tilt again, though.

Olly

If my set up was unbalanced in RA, could that explain the elongation?  I have tried to be careful balancing the Mesu but because there is no way of 'declutching' I do find it a little difficult to tell if I am in balance.  On both axes there is a wide-ish range of settings that will allow the scope to remain stationary when I let go of it.  I have tried two techniques - one where I apply pressure upwards (or downward) on a point approximately equidistant to the centre - when doing this I try to get to a state where it appears that equal amounts of pressure are needed on each side to move the mount.  The other technique I used was to move the weight or the scope to a point where it just started to move if I let go.  Then I'd adjust it in the opposite direction until it did the same - then I would set the scope/weight at the midpoint of these two extremes.  

Are there any other (better) techniques for balancing the Mesu?