Flats

Can someone tell me if the dark band across the top 1/3 of these flat frames is typical ?

The camera is a SBIG STXL16200 with a FW8G-STXL filter wheel. You will see a dark band running across the top of all of the flat frames regardless of the filter.

These are "t-shirt" flats taken indoors, and I experimented with the positions of the light sources to make sure there was not a shadow or gradient, but the band is always there.

Could this darker band be from the shadow of the pickoff mirror in the FW8G-STXL ? I have the mirror adjusted to where I cannot see a shadow visually on the images, but maybe the flats are more sensitive.

Any other comments will be appreciated. All of the flats have been dark subtracted with a set of darks taken at the same exposure time that was used for the flat frame.

Mike

https://www.dropbox.com/sh/hv861e5pshovzvm/AAD7aVv35nPDsLslaL5qV-3Da?dl=0

 

Hi Mike.

It could be the prism as you suggested, just on the edge of the main camera sensor.

One tip I learned with setting up the position of an OAG is to trigger a long exposure, ten minutes or so, and while the camera shutter is open look into the front of the OTA and shine a flash-light inside, you will be able to see the prism of the OAG in front of the main camera sensor while the shutter is still open, now, move your head over to one side of the OTA so that your eye-line is right at the edge of the OTA opening and you will be able to see the beam path from your eye to the edge of the OAG prism and beyond to the camera sensor.

As you move your head (and eye-line) around the full circumference of the opening of the OTA you will see at one side of the OTA aperture that the prism appears far from the edge of the main sensor while at the opposite side of the OTA aperture the prism is much closer to the main sensor.

The correct position of the OAG prism for an un-shaded main sensor is when you can move your head (and eye-line) right around the full 360 degrees of the OTA opening and at no position do you see the edge of the prism overlapping the edge of the main sensor.

In some cases though it is impossible to find a fully un-shaded position for the OAG prism, you might need to push the prism so far out from the optical axis to achieve that that you end up with very distorted stars in the OAG camera because you are positioning the prism where the image field is strongly curved and distorted.

The important thing is that when you apply the flats to the lights does the shaded band calibrate out? if so, it's probably not worth over worrying about.

One thing I did notice in your flats is that the vignetted field overall seems somewhat off-centre, with the brightest part of the vignetted field close to the bottom of the frame and not central to the frame, so maybe the collimation is a bit off or there is some droop in the focuser/OAG/camera attachment.

HTH

William.

 

Thanks for the replies, William & Colin.

I took a picture down the OTA with the camera shutter open. Here is what I see. From what I can tell by following your instructions, the pick-off mirror is well away from the main sensor.

The scope is an Astro-Physics 180 EDT with a Quad TCC telecompressor. I sent it back to the factory a couple years ago to have the lens checked out and to retrofit it with a 3-1/2" focuser. This is also when I got the Quad TCC. I am assuming that they checked the alignment of the new focuser.

I have never noticed a dark band across any of my light images.

You may recall commenting on some images of mine a couple of months ago and suggested that there may be some tilting. If the off-center vignetting that you can see in my flat frames is due to tilting would that explain what you noticed in my images ?

Mike

 

Hi Mike.

It’s not really possible to tell much from the photo, although the innermost black ring which is possibly the focuser draw tube does not appear to have a concentric ring pattern that matches the same pattern as seen in the OTA baffle rings, but attempting to make a diagnosis from a photo down the tube is just guessing as its so dependent on camera position.

For smaller and lighter tubes its quite easy to put the OTA in a V cradle on a bench with a collimated laser in the focuser pointing to a wall ten metres or so away, as you rotate the OTA in the cradle the laser dot on the wall should stay in the same place, if the laser dot describes a circle on the wall as the OTA rotates in the cradle then you’ll know that the optical axis is not aligned with the mechanical axis and can begin looking for the cause. For a big heavy refractor like yours its not something I’d like to try on my own.

Have you got a telescope vendor nearby with an optical bench who could check the alignment for you?
The problem with sending things off by shipping is that they get thrown around during transit and a perfectly collimated OTA that leaves the manufacturer can be badly mis-collimated by the time you get it back.

You could also send some flats and lights images to AP and ask their opinion.

As for the band, if it’s not a shadow it could just be a natural variability in the response of the sensor to the lighting source used for the flats, you would need to test again with a known full-spectrum light source such as sky flats.
Even if the band remains, provided it is fully removed when the lights are calibrated then it can be ignored unless it shows signs of getting worse over time.

If you have real concerns about the camera you could take the camera off the telescope and just point the camera at a white screen about a metre away, very dimly illuminated by natural daylight just before dawn is breaking, or just after sunset, and take a few images, if the banding is still visible you’ll know at least that the banding is not an artefact of the OTA.

I’ve never been a fan of T-shirt flats myself as I’ve seen too many flats images taken that way with obvious weave patterns and other artefacts caused by the fabric.

Wait to see what Colin thinks, he’s much more experienced with these things than me and can offer better advice.

William.

 

Hi William

Here is a better picture. It is in better focus, and it is possible to see where the shadow of the pick-off mirror falls. It doesn't look like the shadow is touching the sensor.

I have posted the same question on the Astro-Physics forum. Hopefully, Roland Christen will comment. It was he who checked out my 180 EDT. Initially, with just the lens cell, but when I decided to upgrade to the 3.5" focuser I sent them the whole OTA. I do not want to do that again. The shipping cost was very high.

I live within driving distance of Texas Nautical Repair. They are the North American supplier for Takahashi. Maybe they have the ability to check the collimation of my scope.

I will see what I can do about a better way to take flats. This is admittedly the first time I have ever tried to take any flat frames with this or any other camera. Until now, I have relied strictly on dark frames for my image reduction.

Mike

 

Hi Colin

The weather is supposed to clear off later this week. I will send you something then.

Mike

 

Hi Colin. I have attached a link to some images of M27 that are the most recent I have. These are from mid-October. There is one each of L, R, G, and B along with a Master dark frame that I recently made. The area around M27 has as uniformly dense a starfield as I can think of.

Mike

https://www.dropbox.com/sh/8pkumbavv22856s/AABc851K-sQaRPRuPyhjSD9Oa?dl=0

 

Hi Mike, Colin.

I had a quick look at your M27 images while waiting for my own images from Portugal to download this morning, I'm not sure that much can be determined from this particular dataset because there is no consistency across the four images re: tilt angle.

For meaningful CCD Inspector results the data set should ideally contain only stars and no nebula, and have an exposure time that produces a large number of unsaturated stars to calculate an accurate measurement.

It's not a good idea to use colour filtered exposures from a refractor because there are always some differences in star spot sizes and shapes across the colour spectrum, however good the optics claim to be.

To produce meaningful results with CCD inspector point the telescope to a rich star field at the zenith, take half a dozen exposures through the luminance filter and put those in one folder, then repeat on a different target at an altitude of around 45 degrees and put those images in another folder, finally, if using a GEM mount, flip the mount across the meridian and take a final half-dozen images at an altitude of 45 degrees.

With those three data sets you can determine the basic tilt angle when the OTA is pointing upwards and gravity is supposedly pulling all the optical components into a straight line, while the 45 degree tube angle images from either side of the meridian will show you if something is loose in the image train as the measured tilt angles will be markedly different either side of the meridian flip.

Attached is a snap-shot of the analysis of your four M27 images where you can see the large variation across the colour filters, even though the OTA elevation angle was not that different between them.

Of the four, the only one I would have any faith in is the Luminance image which shows a tilt of 12 degrees in the star shapes across the field, this should not be confused with mechanical tilt though, a very small tilt in mechanical angle produces a much bigger tilt in FWHM star shapes.

Trying to interpret CCD Inspector results is an art-form in itself and not one I'm fully comfortable with..

William.

 

Thanks to both.

I will try to take some images this weekend. It looks like we will finally get some clear weather.

If I understand William's suggestions correctly, you are looking for Luminance images of dense starfields with little or no nebulosity looking directly overhead, and at 45 degrees on either side of the meridian with exposure time so as not to saturate the stars.

Figuring out if there is tilting in my imaging train is just the beginning. Figuring out what to do to correct it will be the next challenge.

Mike

 

So many Software choices! Do you know if the license fee for the CCD Inspector software is a one-time fee or an annual subscription? I have been thinking about getting PixInsight for processing, but I don't think I can get both PixInsight and CCD Inspector. Not so much because of the money, but because I don't think my brain can handle trying to learn two new software systems.

There was an excellent post on the Astro-Physics forum a couple weeks ago that talked about measuring tilt using the Hocus Pocus plugin for the N.I.N.A. software.

I looked through my archives of images, but I don't have any recent images using this exact imaging setup. Earlier this year I switched back to my original STXL-FW8G filter wheel cover so that I could use the integral guider. My older images were all taken with the FW8S cover. I should be able to get some images this weekend.

Mike

 

Just a quick update. Due to a combination of bad weather and extended power outages on both Saturday (6 hours), and Sunday (10 hours), I was unable to take any images over the weekend. It has been that kind of year.

Mike

 

The skiers are happy and think the Wx is great owing to Northern Utah mountains having a 60" (1.5m) of snow dump.

 

At last, I have some images to look at for tilt analysis. These were all taken on December 31. It wasn't a particularly good night, and by 8:45 pm the fog started coming in. By 9:00 we were completely fogged in.

I captured luminance images ranging from 4 to 15 seconds as follows. All but the 15 second images were unguided. Some of the images had auto-dark subtraction.

• NGC869 & 884 This would represent a directly overhead orientation with the scope pointed nearly vertically.
• Deneb region: This would represent a less than vertical orientation facing west.
• M37: This would represent a less than vertical orientation facing east. I also included a guided 15 second exposure of M37.

I was going to try to get something a little closer to the eastern horizon, but I ran out of clear sky and had to shut down.

Mike

https://www.dropbox.com/sh/a0y8z2r91laon6o/AABdl1Uc4JwYGd7hwrZAgF_Da?dl=0

 

Thanks for looking at these for me Colin.

Do you suppose that the exposure time has anything to do with the better results ? These recent images were all very short exposures between 4 and 15 seconds while the older images were all 600 second guided exposures.

I have often wondered about my guiding performance. When I guide I will typically get an error about 0.3" to 0.5" RMS, but over the course of taking several long exposures I can see that the position of the stars in the images will shift.

Point taken about the declination of NGC 869. I live at 30 Degrees latitude. I must have been tired when I was typing :>)

Mike