Page 1 of 2

Issues with COAST

Posted: Thu Mar 30, 2017 1:11 am
by rjmodic
COAST has been up and running over three weeks now and its great to see how productive the telescope has been following a long drought with BRT/ART.

However, there are some issues with the new telescope that users should be aware of, especially those using it for photometry.

There is a calibration issue involving incorrect bias and dark subtraction. The bias is being over-corrected. Images with this problem have a high number of pixels with 0 ADU values. These images cannot be properly flat fielded. Doing so results in images that are severely over-corrected.

The COAST camera also has a bad case of RBI (Residual Bulk Image) causing ghost images from previous jobs to appear in images even 20 to 30 minutes later.

The filter wheel has been occasionally selecting the wrong filter. I've gotten an R frame instead of I, a V frame instead of and R, etc. This has happened with BVR jobs as well as single filter ones.

The images from COAST are mirrored left to right. Not sure why this is since I believe the scope is just a two-mirrored system.

I've e-mailed the staff about these issues and hopefully they can be resolved soon. In the meantime, users doing photometry with COAST images need to use caution. I recommend using only comp stars near the variable stars to minimize the errors due to unflatted images. Also try to verify that the filter is indeed the correct one (the filter info in the FITS header may be incorrect). Using red/blue pairs of comp stars and looking at the residuals from your photometry software can help to spot this problem.


Re: Issues with COAST

Posted: Fri Mar 31, 2017 12:43 pm
by garypoyner
Hi Bob,

I've had an R image of OJ287 when my selected filter is V. Fortunately it's easy enough to spot once photometry has been done. This has only happened on one occasion though, despite a number of my jobs being in V.

The focus is a problem too, and today's (March 30/31) batch of images have been particularly affected. The camera orientation is a pain, but it is a known fault which hopefully the team will correct on their next site visit.

I agree completely - it's great to see COAST busy again. Well done to the COAST team!


Re: Issues with COAST

Posted: Tue Apr 04, 2017 9:10 am
by vega4176
Hello I did take some photos of M101 but I noticed that the pictures are very blurry. Is just to warn you that there are focus issues. Regards Carlo

Re: Issues with COAST

Posted: Fri Apr 07, 2017 10:38 am
by vega4176
Hi yesterday I asked to photograph M13 with the blue filter for 60 seconds. The work was performed (see J291076) but the image is very blurry. You will notice that there is still the problem of blurring. Thanks Carlo

Re: Issues with COAST

Posted: Mon Apr 10, 2017 11:12 am
by garypoyner
I believe there will be a site visit sometime soon (after Easter), so hopefully the focus, camera orientation, filter selection and drift in RA & Dec will be looked at.


Re: Issues with COAST

Posted: Thu Sep 07, 2017 11:20 am
by ThomasK
What I understand there is no flats applied to the images from COAST. It is visibly at current with the strong moon shine that the images suffer from vignetting. It is also visible in the photometry where I have to avoid comparison stars at the edges of the images because they deviate to much from comparison stars in the center.

Maybe there also are problems with the darks as the many hot pixels are not subtracted from the images or maybe the temperature of the camera should be looked over.

With the great number of images that COAST now produce it is a pity that usage of flats and darks seems not work correctly. It would make the good images great with not to much work.

From time to time there also is problem with the focusing of the telescope, as I have written in a previous post.

I don't know if anyone from OU is reading this forum, there has been very little response or activity in the forum the last year. That is also a pity, if I was an owner of a such a great telescope used by many engaged amateur astronores around the world I would love to hear what they think and how they use the images. Is there any other forum that OU are using to be in touch with their users of COAST, please tell.

Thomas Karlsson

Re: flat-field & hot-pixel compensation of COAST images

Posted: Fri Sep 15, 2017 4:04 pm
by RossW
Well Thomas, it's not that difficult to make a "Synthetic Flat" to compensate your images.
I first tried this technique on the BRT Galaxy Camera some years ago, and it worked rather well.

This is how I did it for my recent comet images taken through COAST's "Clear" filter:
First I searched through the Gallery of recent images and chose thirteen (the more the better) with exposure times of >=2 minutes taken with the Clear filter around the time of the last Full Moon. I carefully selected images which didn't contain any saturated stars or bright extended objects, but where the scattered Moonlight provided a good illumination of the sensor.
Then I normalised this sequence and used a Median addition to stack them. This "averaged out" the stars in the individual images and left just the background illumination, showing up the vignetting and dust-spots:
flat.jpg (13.8 KiB) Viewed 1089 times

This Synthetic Flat is quite crude, but works surprisingly well: here's an example of an image before & after flat-field division (contrast-stretched to accentuate the effect):
flat_comp.jpg (37.6 KiB) Viewed 1089 times

As for the hot-pixels, I found that I could remove almost all of them using a 3x3 pixel median filter: the stars are spread over a number of adjacent pixels, so they aren't affected.

I used Christian Buil's IRIS program for all this processing.

With best wishes from Ross Wilkinson
Bolton Astronomical Society

Re: Issues with COAST

Posted: Mon Sep 18, 2017 8:06 pm
by ThomasK
Thanks for your response Ross. On the old BRT telescope I also used a synthetic flat that I made in IRIS. I have not experienced that the vignetting on COAST is that severe that I have needed to use a home-made flat, it usually is enough to remove one or two comparison stars.

It may seem like a luxury problem, but with the phenomenal production rate that COAST have I usually got 15-20 images a day and it take me a few hours each day to process them, so every time saving action counts. At the end of the BRT lifetime the team managed to do a usable flat that were applied automatically. As the technique seems solved I think they should do the same as they did at BRT.

For hot pixels I'm not sure that 3x3 median filter is usable when you do photometry. I don't care for hot pixels that are not near any variable or comparison star. If it is on any comparison star the easiest way is to exclude that star. If a hot pixel is on a variable star it is worse. I have some times used IRIS to repair the pixel if it is in the outskirts of the star, but if it is in the centre I don't think you could regenerate the correct value. It is not often this happens, but it do happen some times. So any measure that can be done to avoid hot pixels should be appreciated.

Don't missunderstand me, I'm very pleased with the images from COAST and the telescope deliver a lot of useful results. Only that the images could be even better with not that much work.

Thomas Karlsson

Re: Issues with COAST

Posted: Thu Sep 21, 2017 2:22 pm
by Ed

I did create a synthetic flat for COAST some time ago an load it into the system...only it seems I forgot to change the file extension from .fit to .fits after it came out of IRIS and the system didn't find it.

Fixed now.

Re: Issues with COAST

Posted: Thu Sep 21, 2017 10:17 pm
by rjmodic
I too have made synthetic flats for COAST using 50-100 images for each filter as I did for BRT. However, these flats don't work as well as the BRT flats did. For COAST images taken when the sky was moonless, the corners of the frames appear over-corrected. This is because the dark subtracted images have many negative pixel values and, when saved in 16 bit .fits format, the negative values are truncated to zero. This truncating effect is most noticeable in B frames. I have suggested to the COAST staff that the procedure for making master darks and dark subtraction should be reviewed to see if it could be improved, but I haven't heard of any progress with this issue. An improved dark subtraction routine could also help with the residual bright and dark pixel problem mentioned by other observers.