Coathanger, the full half hour...

[ollypenrice]

After making a tilt-test optical jig for the ASI2600 camera and tuning out a bit of tilt, it went back in the RASA 8 last night for a test. It was a disaster, worse than before. We then rotated the camera 90 degrees from landscape to portrait and, for the first time since we started all this, we had perfect corners. Ours is not to reason why. CCD Inspector gave great values for FWHM, tilt and collimation so... if it ain't broke, don't fix it. Portrait it is.

We stacked 10x3 minute subs with no calibration files and I processed the stack in PI and Ps. I was expecting to give the corner stars a cosmetic tweak here and there at 100% but there was nothing to fix. Bliss!

(Note: a smaller chip would have been performing well some time ago. The APSc we are using is more demanding of tilt and collimation at F2.)

Olly

[peter shah]

A result I'd say Olly...round stars are round stars!!

[vlaiv]

It's hard to tell on reduced image, but I'm having a sense that star halos are elongated in diagonal direction (bottom left - top right)?

Is it just me, or is this real?

[ollypenrice]

52 minutes ago, vlaiv said:

It's hard to tell on reduced image, but I'm having a sense that star halos are elongated in diagonal direction (bottom left - top right)?

Is it just me, or is this real?

I could persuade myself that I could see this but I cropped a few stars in centered square frames and I can't really find anything.

Olly

[irtuk]

Beautiful Olly. 'Dunno what you are talkin 'bout, but beautiful

 

[vlaiv]

4 minutes ago, ollypenrice said:

I could persuade myself that I could see this but I cropped a few stars in centered square frames and I can't really find anything.

Maybe this helps?

Or maybe if I rotate it 90° at a time instead of flipping it:

I still think it's there.

[kirkster501]

deleted.

 

Edited September 1, 2021 by kirkster501

[Stu Todd]

1 hour ago, vlaiv said:

It's hard to tell on reduced image, but I'm having a sense that star halos are elongated in diagonal direction (bottom left - top right)?

Is it just me, or is this real?

I agree with vlaiv. This elongation was the "stand out" aspect of the image. It looks like it's just the halos though. 

I haven't had a magnified look at the central orange stars etc.

 

 

 

Edited September 1, 2021 by Stu Todd

[Mike73]

Is it really something we can fix though or do we even need to fix it? I dont know why but I think an off centre halo looks better than a centred one.

Ignoring the micro lensing (which is a whole lotta clone work saved for a cloudy night) and considering Sadr is actually pretty bright the halo in this doesn't bother me too much.

[gorann]

1 hour ago, vlaiv said:

It's hard to tell on reduced image, but I'm having a sense that star halos are elongated in diagonal direction (bottom left - top right)?

Is it just me, or is this real?

A great relief to see that you finally got round stars Olly! Regarding oval star halos I have also noted that in some of my RASA images and suspect it could be related to the curved camera cables, so a kind of soft star spike, but that is just a guess. Just out of curiosity, maybe you can have a look at what angel your cables lie in relation to this slight halo effect. In any case such a small halo distortion is easy to live with.

Edited September 1, 2021 by gorann

[The Admiral]

36 minutes ago, Mike73 said:

Is it really something we can fix though or do we even need to fix it? I dont know why but I think an off centre halo looks better than a centred one.

Ignoring the micro lensing (which is a whole lotta clone work saved for a cloudy night) and considering Sadr is actually pretty bright the halo in this doesn't bother me too much.

 

May be the odd star can be ignored, but I do find that when all the stars seem to have oval halos with their major axes in the same direction, it does detract from the image. Just my take of course, YVMV.

Ian

[ollypenrice]

1 hour ago, vlaiv said:

Maybe this helps?

Or maybe if I rotate it 90° at a time instead of flipping it:

I still think it's there.

Yes, you're right. Another good test!

1 hour ago, gorann said:

A great relief to see that you finally got round stars Olly! Regarding oval star halos I have also noted that in some of my RASA images and suspect it could be related to the curved camera cables, so a kind of soft star spike, but that is just a guess. Just out of curiosity, maybe you can have a look at what angel your cables lie in relation to this slight halo effect. In any case such a small halo distortion is easy to live with.

I also wondered about the cables, since the stars themselves look round and CCD Inspector is happy with them.  A diffraction artifact from the looped cables could well be a smeared out diffraction artifact and the orientation is plausible. I could always try a different cable route in front of the corrector. The present USB3 camera cable is also a thick one. A flat one, edge on, might be a better bet.

I'll have a think about a semi-automated Photoshop fix as well but, even 'as is,' this is not the kind of thing to bug me. Like the Takahashi 'Inverse lighthouse beams,' I don't worry about them.

Edit: A quick wipe with the Burn Tool (set to mid-tones at 6%) over the 'long' side of the halos seems to dispense with most of the effect. It's only necessary to my eye on the brightest stars.

Olly

Edited September 1, 2021 by ollypenrice

[vlaiv]

24 minutes ago, ollypenrice said:

I also wondered about the cables, since the stars themselves look round and CCD Inspector is happy with them.  A diffraction artifact from the looped cables could well be a smeared out diffraction artifact and the orientation is plausible. I could always try a different cable route in front of the corrector. The present USB3 camaera cable is also a thick one. A flat one, edge on, might be a better bet.

It's very hard to guess what could possibly cause elliptical shape of halo. Maybe we should try analytical approach?

Size of diffraction spike depends on length of straight edge that causes it. Circular aperture can be seen as bunch of very short straight edges each at slightly different angle - that creates great many weak diffraction spikes each perpendicular to its own forming edge - these join together to form uniform halo.

In order for halo to be elliptical - diffraction edge in one direction must be longer than in other direction and yet uniformly curving to form almost complete / smooth halo (halo in the image is not completely smooth - it has some spikes - due to segment of cable being straight).

I think that it must be section of ellipse as well.

One simulation later ....

(left - aperture, right PSF blurred with gaussian blur and stretched to show halo).

This only shows that one must cable-manage!

[ollypenrice]

19 minutes ago, vlaiv said:

It's very hard to guess what could possibly cause elliptical shape of halo. Maybe we should try analytical approach?

Size of diffraction spike depends on length of straight edge that causes it. Circular aperture can be seen as bunch of very short straight edges each at slightly different angle - that creates great many weak diffraction spikes each perpendicular to its own forming edge - these join together to form uniform halo.

In order for halo to be elliptical - diffraction edge in one direction must be longer than in other direction and yet uniformly curving to form almost complete / smooth halo (halo in the image is not completely smooth - it has some spikes - due to segment of cable being straight).

I think that it must be section of ellipse as well.

One simulation later ....

(left - aperture, right PSF blurred with gaussian blur and stretched to show halo).

This only shows that one must cable-manage!

What we have at the moment, except for a change in its orientation...

Olly

[The Admiral]

 

6 minutes ago, ollypenrice said:

What we have at the moment

Is the dew shield a little oval? Could that conceivably have an effect?

Ian

Edited September 1, 2021 by The Admiral

[ollypenrice]

4 minutes ago, The Admiral said:

 

Is the dew shield a little oval? Could that conceivably have an effect?

Ian

It is, but it's a good deal larger in aperture than the corrector plate so shouldn't be seen by the optics or camera. Still, worth a look. I also rather wish the camera were black since reflective things close to the light path are never to be approved of.

Olly

[vlaiv]

16 minutes ago, ollypenrice said:

What we have at the moment, except for a change in its orientation...

That is a circular ring that those cables are attached to? Seems to fit the orientation in the image as it is at 45°.

What is the height difference between camera rear and dew shield? Tilted circle is ellipse.

[The Admiral]

9 minutes ago, ollypenrice said:

It is, but it's a good deal larger in aperture than the corrector plate so shouldn't be seen by the optics or camera. Still, worth a look. I also rather wish the camera were black since reflective things close to the light path are never to be approved of.

Olly

OK. Though the fov on the long side is 3.4 degrees, and the corrector plate aperture extends almost to the tube, so it would depend on the dew shield's extension of course.

Ian

[ollypenrice]

1 hour ago, vlaiv said:

That is a circular ring that those cables are attached to? Seems to fit the orientation in the image as it is at 45°.

What is the height difference between camera rear and dew shield? Tilted circle is ellipse.

Yes, it's the edge of a jam jar lid with the flat middle cut out! It does slope up slightly as it is now from camera to dewshiled edge so, as you say, forming an ellipse. I can easily modify it to be parallel with the corrector. Good point.

1 hour ago, The Admiral said:

OK. Though the fov on the long side is 3.4 degrees, and the corrector plate aperture extends almost to the tube, so it would depend on the dew shield's extension of course.

Ian

The dewshield doesn't extend far at all above the top of the camera.  Maybe 50mm.  Making the dewshield short like that arose from a discussion with Goran who found no dewing issues since the camera is warm and draws air in from the back before blowing it out of its sides. The idea is not to trap too much of the cooler's warm air output in the dewshield. There was certainly no dewing issue on the RASA on a morning which saw the guidescope lens very wet.

I think an elongation on the side of the cable circle makes sense.

Olly

Edited September 1, 2021 by ollypenrice
clarification

[GalaxyGael]

Good theories and liklihood in this thread. I also think the oval halos were a little prominent, or just a tad more so than the nice round stars. 

Can we eliminate the turbulence caused by the two camera exhaust vanes, where there would be two regions of hotter air, given the speed of the scope? 

[ollypenrice]

21 minutes ago, GalaxyGael said:

Good theories and liklihood in this thread. I also think the oval halos were a little prominent, or just a tad more so than the nice round stars. 

Can we eliminate the turbulence caused by the two camera exhaust vanes, where there would be two regions of hotter air, given the speed of the scope? 

Now there's a thought! Not more ideas!!! Oh no...

One crisis at a time is the trick when getting a RASA to work.

I do normally pull halos down in post processing but this was a quickie and I also felt they helped define the cluster as the subject of the image.

lly

Edited September 1, 2021 by ollypenrice

[tomato]

Great start, now I know why I went for the QHY version, it’s flat black.

Based on my experience the air inside the dew shield is well mixed and warm,  based on the fact that the cooler cannot maintain the set point temperature.

 Just a thought, what if you have to rotate the camera to get the target (eg M31) across the diagonal?

[wimvb]

You could of course do what Marcel Drechsler does:

(If you need me, you can start searching in the woods 😉 )

[ollypenrice]

8 hours ago, tomato said:

Great start, now I know why I went for the QHY version, it’s flat black.

Based on my experience the air inside the dew shield is well mixed and warm,  based on the fact that the cooler cannot maintain the set point temperature.

 Just a thought, what if you have to rotate the camera to get the target (eg M31) across the diagonal?

I only ever rotate the camera between landscape and portrait. I never feel the need to use a random angle and doing so makes returning to the target some time later a nightmare. M31 won't fit on my widefield rigs anyway without cropping the outer glow, which I would hate to do, so I regard it as a 2-panel target even with a 4 degree-wide field.

Last night we corrected the angle of the camera to accurate Portrait format and star shapes were still good. However, we probably can't use Landscape with the RASA as it is since we get distorted corners that way. We can live with this. It will doubtless do mostly mosaics anyway, which is one of the ironies of widefield setups. You have a wide field and you want wider... I never do mosaics in the TEC.

15 minutes ago, wimvb said:

You could of course do what Marcel Drechsler does:

(If you need me, you can start searching in the woods 😉 )

This is a possibility but I really don't like diffraction spikes, especially on widefield images. I think of it as the 'World war 2 cemetery look.'  Last night's images, after tweaking the cable loop, seems to have round stellar glows.

Olly

 

[Catanonia]

That is why I went for the smaller ZWO183MC with the much smaller pixel size. No such issues for me and my RASA 8 and getting very good resolution / sampling.