Beginning I-filter imaging. Help with image quality

Affiliation
American Association of Variable Star Observers (AAVSO)
Fri, 08/01/2014 - 01:27

I attach two images. One is a raw integration and the other is calibrated (bias, dark flat). The I-flat was 120 seconds and looks OK (ca, 24000 ADU). The dark is 120 seconds scaled to the image. The image integration is 90 seconds. Airmass was OK (ca 1.2). Note the broad streaks. This streaking seems to be integration-dependent. 15 second integrations -- hardly noted. 10 second integrations almost absent. Observatory is in the suburbs in ca. 4th mag skies. Both B and V looked fine.

I am new to I-filters. Can someone on the forum give me some advice on this streaking? I am assuming that the photometry will be off, so I am not reporting any data with I-filters until I figure this out.

Thanks, Ed

 

 

 

 

Affiliation
American Association of Variable Star Observers (AAVSO)
Ic-band

Hi Ed,

Interesting.  Remember that Ic is outside the eye's range of wavelengths.  Did you see these effects at Rc?  Have you looked for light sources in your observatory - such as a webcam with IR illumination?  Do the streaks stay constant, no matter what direction you point?  What camera and filter vendor?

What you see is not common for Ic.  The images should look the same as for other filters, though you have less throughput and there are atmospheric emission lines (mostly OH) that make things noiser.

Arne

Affiliation
American Association of Variable Star Observers (AAVSO)
Ic-band

[quote=HQA]

Hi Ed,

Interesting.  Remember that Ic is outside the eye's range of wavelengths.  Did you see these effects at Rc?  Have you looked for light sources in your observatory - such as a webcam with IR illumination?  Do the streaks stay constant, no matter what direction you point?  What camera and filter vendor?

What you see is not common for Ic.  The images should look the same as for other filters, though you have less throughput and there are atmospheric emission lines (mostly OH) that make things noiser.

Arne

[/quote]

Thanks for the reply, Arne. I only have BVI, no Rc. I do have a Red non-photometric filter I can try. Yes, there is a laptop in the observatory with a red screen. I will attenuate the light on it as much as possible tonight.  In fact, I will cover it, not needed as I control from the house with TeamViewer9. The guider has a Meade Pro, nothing to be done about that. I have not checked direction spefically, but the bands appear faintly in other variables (taken at lower integrations) that are some distance away from this particular target. I will perform some direction experiments tonight with similar integration times. The BVI set are the newer Astrodon filters.

Let's see what happens tonight.

Ed

Affiliation
American Association of Variable Star Observers (AAVSO)
What type of CCD do you have?

Ed,

Any chance you have a backlit CCD?  These look a little like the fringing affects I have on a new FLI backlit CCD.  They only show up with I (and also Clear) filters.

Gordon

 

Type of CCD

Is your CCD front-illuminated or back-illuminated?  This doesn't look like classical interference fringing but that is the first thing that popped into my mind.  I-band fringing due to atmospheric night-time OH emission depends on the construction of your CCD.  It should be pretty constant no matter where the scope is pointed and it is more commonly produced in back-illuminated CCDs.  

This article details some of the issues and soluitons regarding fringing in the I-band:  http://arxiv.org/abs/1201.2336

My other thought is that it could be scattered light from a bright star a few degrees away.  The other night I had an odd hozizontal ghost in some of my images and when I experimented with it, I found it was caused by off axis reflections in my scope associated with a third magnitude star 2.5 deg to the north of the field.  If that is the case for you then the effect shoud change if you move the location of your field relative to the bright star.

 

Affiliation
American Association of Variable Star Observers (AAVSO)
I think the problem is solved

Thanks to all who replied.

Arne: Yes, it appears to be "observatory contamination." Two possible sources where screened. Laptop covered. Christmas lights (little red buggers) turned off. I suspect the Christmas lights. At any rate, a 90 second exposure of M13 showed no banding. I am going to spec the Christmas lights and the computer screen next time I get out the Aply600.

Ed

Affiliation
American Association of Variable Star Observers (AAVSO)
lights

Hi Ed,

I always have more problem with observatory scattered light at Ic (and redder).  Things like electrical tape become transparent, things that look black aren't, lights are much brighter than they are at eye wavelengths.  The exception was when I was doing NIR imaging (JHKL); I could use a white LED flashlight anywhere in the dome without problem, since the LED does not emit in the NIR.  My preference whenever having to tape anything these days is to use the metallic duct tape, which is not transparent at any wavelength.

So I always check my observatory light sources first when I see patterns like this.  There are other possibilities, especially if you are using a back-illuminated CCD as suggested by others, but they have specific features/patterns that make it easier to identify.

Arne

Affiliation
American Association of Variable Star Observers (AAVSO)
Xmas Contamination lites

Hello Ed

Your suspicion about the Xmas lites is well founded.  My experience with the red rope lights that are sold at Xmas time, is that even with a dimmer, they have lots of Ic in their spectrum.  Even if you can just see the red, there is maybe 10x in Ic--never measured. 

I do have a back illuminated chip, and this does not look like my Ic pattern.  Its more of a random swirl.  Of course, as they say, your results may vary.

I also cover my laptop with a deep red plexiglass cover. 

 

Gary

Affiliation
American Association of Variable Star Observers (AAVSO)
Xmas Contamination lites

Thanks for the comments, Gary. Once I turned off the Xmas tree lights and covered the computer screen the problem disappeared. I use a red screen on the computer but figured I would blank it out also; not needed as I control from the house using TeamViewer9. My ccd is not back-lit, its a standard ATIK 314L+. Amazingly, even with ABG, its linear out to 60000 ADU.

Arne: thanks for the additional comments. I now have a deeper appreciation for scattered observatory light!

Ed

Affiliation
American Association of Variable Star Observers (AAVSO)
Icron Ranger light contamination

Hello Ed and All

On the subject of light contamination, I have an Icron Range, which has a blue, an orange, maybe red and a green LED for status lights.  The blue one is really bright.  Last week, I put 2 layers of Dymo Label Tape over them.  I can still see the colors, but they do not light up the observatory.  The REX was on the scope and the LEX was at the computer. 

 

Gary

Affiliation
American Association of Variable Star Observers (AAVSO)
Sculer I filters

James, 

Because of your comment about transforming the schuler filter measurements, I guess you are aware that with a CCD these will give results that vary considerably from the Cousins standard for I particularly for very red stars or objects with emission lines in the infrared outsied of the standard Ic filter band. . They use a glass formulation that has a very long red end tail that  CCDs don't cut off but PMTs do. It's not nearly as bad as the difference between the Johnson I and the Cousins I filters, but it will probably give you a significantly larger transformation coefficient than you get for your other Schuler filters. 

If you aren't aware of this you can get an idea by going to the Astrodon site and comparing the response spectra that is still posted for the Schuler filters to the response curves for his current U B V Rc Ic filters designed to match CCDs to the Johnson-Cousins standard. The red end of the Schuler I filter furve is cut off in the Astrodon Graphic. The complete one is attached. 

Brad Walter, WBY

Affiliation
American Association of Variable Star Observers (AAVSO)
Schuler Is

Hi Brad,

If you multiply that filter transmission (Schuler Is) times the QE response of the typical front-illuminated CCD, you come very close to the traditional Cousins Ic.  The original BSM uses the Schuler/Bessell colored-glass filters and does pretty well, with small transformation coefficients.  It is true for very red objects that you will get different results than Cousins would have gotten with his equipment, primarily because his photomultiplier had less red response than does a silicon CCD, and that the dielectric filters now sold by Astrodon and others will give you an even closer match to the standard system.  However, it is usually a small effect for most stars.

All wide-band filter systems cannot be defined by a filter alone, as you need to fold in the detector response, and on occasion, the atmospheric response.  The best standard systems are defined by a total system response curve that includes all of these effects.  Then you can tailor your own system to match.

Arne

Affiliation
American Association of Variable Star Observers (AAVSO)
Bessel Ic filter with CCD vs. Standard Ic Response Curve

Arne, 

Before purchasing filters for my new camera with a KAF 6303 front illuminated CCD, I did a comparison of the response I would get compared to the Rc and Ic standard response curves. I found on a normalized basis that there was a significant difference between the standard Ic response curve and the one I calculated for the Bessel filters with this camera. Attached is a PDF of my comparison. I used the standard response curves from the Optec Filter Monograph but I also compared the curve in the monograph to table 1.II in the Bessel 1979 PASP paper. They are consistent. 

As you can see from the attached file, the CCD with Bessel filters has peak sensitivity as about 75 nm shorter wavelength than the standard response curve and has s significant "red tail" beyond 900 nm where the RCA 31034A PMT has dropped to zero response. The "standard" RCA PMT response actually drops to essentially zero a bit above 900 nm and the CCD and filter combination still has normalized sensitivity of 32% at that wavelength is still above 15% at 950 nm. 

I haven't calculated transformation coefficients for Bessel formula filters for a front illuminated CCD. I have only done them for the Astrodon interference type. However, unless the tail and the shift of the peak toward the blue end of the spectrum counteract each other somehow, it seems to me that transformation coefficients for the Ic filter will be farther from zero for a single color transform and transformation coefficients for color indexes including Ic will be farther from one than for the other Bessel formula filters with a CCD. 

What am I missing?