Hi Team DL, I have the camera and filter wheel installed on the CDK and my first impression is speechlessness. Very happy with my early results. Camera's been trouble free, installed without an issue and up and running quickly. The sensitivity of the sensor is simply insane. The background on my images so far have a strange wavy pattern in them that I haven't seen before. Can you tell me what's going on here? Does it calibrate out? If so....how? The background pattern is in all of my light frames I've calibrated with the usual bias, darks, flats. Master calibration frames were all created in MaximDL using default settings in the Set Calibration module. The images I'm calibrating are unfiltered - as in no filter in the C slot. The pattern is in all light frames and doesn't change from one image to the next. Any thoughts on what's going on here? Uncalibrated and calibrated FITS are attached. Thanks, Mike
That is called Etaloning caused by a spectral line in your background sky glow interacting with the thickness of the back side illuminated sensor. Basically a similar effect to Newton's rings. Here are some articles: https://www.azom.com/article.aspx?ArticleID=14009 https://andor.oxinst.com/learning/view/article/optical-etaloning-in-charge-coupled-devices When I've used backside illuminated CCDs in the past, I've found the simplest way to mitigate them is to calibrate using sky flats. That way the background sky glow spectrum is the same in then flats and lights, so it tends to remove the effect.
Mike here is another good read. The biggest thing I get out of it is that Etaloning is most apparent at wavelengths longer than 800nm. I too own the 47-10 and haven't seen any etaloning effects, although I am using standard LRGB/SHO filters. https://www.aavso.org/sites/default/files/etaloning- Princeton instruments.pdf
Thanks to both of you for the quick, informative responses. They don't seem to calibrate out so well with my unfiltered frames (aka clear) even with sky flats. I don't (yet) own a clear filter in a 36mm size, and perhaps the solution here is to get a true luminance filter that has a cutoff on the red end of the spectrum. My main use for the clear filter is light curves to determine rotation periods of minor planets, luminance shouldn't have any meaningful effect on the results and would clean up the background. As you'd expect, the etaloning doesn't show up in a Johnson B filter frame. I'll shoot a handful of frames in V, Rc and Ic tonight as well as sky flats for Rc and Ic and see how the results turn out. Thanks again for giving me some direction on this one. Mike
Doug, is etloning a CCD specific issue, or do the new back illunimated CMOS chips have the same effect?
It can happen with any sensor that has a thin detection layer. I haven't had a chance to evaluate the 4040BSI as yet... constant bad weather.
This is one of those cases where the thickness of the layers is close to the wavelength of light. (e.g. 3.5um thick silicon, vs 0.5um wavelength of light).
Hey Doug, Any pointers on getting proper sky flats to calibrate out the etaloning? Attached frames are my before/after and the master flat I used. The master flat was during evening twilight a couple of hours ago with the telescope pointed near zenith. Being that this happens beyond 800nm wavelength, my Ic filter (and perhaps Rc) should be the only thing affected, but I'd like to get to a calibration method that addresses the etaloning when using that filter. Thanks, Mike
Here's the thing... what you really need is the same spectrum as the night-time sky background, which actually means that twilight flats aren't ideal. It takes more work, but it's quite possible to build night sky flats. Some suggest turning off the clock drive and letting the stars trail, but IMHO that just makes lots of bright streaks that are harder to remove. Instead, take a series of exposures of slightly different fields. The images need to be deep enough to get a good background, and you need enough frames to remove the noise. You don't have to move a whole field every time, just a chunk of one. MaxIm DL's Mosaic feature could be used for this. Once you have a bunch of frames you can median combine the frames to get rid of the stars. The best ever sky flat I ever did was for the Desktop Universe all-sky mosaic (later included in Starry Night Pro Plus). I selected several thousand frames outside the Milky Way (we had tens of thousands of images - it was the entire sky after all), zeroed out stars above a certain ADU limit (dropping them out of the stack), and then median combined the rest of the data. It was sooo smooth!
Thanks for the guidance Doug. I'm currently doing time series photometry on multiple asteroids and they're all in relatively unbusy fields this time of year. Similar to your zeroing out stars above a certain ADU limit, I used StarXterminator to remove stars from 750 frames over the last 3 nights in 8 or 9 different locations in the sky. Then, I median combined all of those frames into my "EtalonFlat.fits" file. This is a promising move forward as it gives me a frame of just the etalon affect on the images. The .fits file is attached. I applied the EtalonFlat to an image and I'm pretty happy with my first results. One thing I noted is that I needed to subtract (NGC3189 - EtalonFlat) instead of dividing. Does that make sense, or am I still missing something here? My basic calibration on this was: ((LightFrame - Bias - Dark) / FlatFrame) - EtalonFlat Either way, happy with the progress and grateful for your support. I think a luminance filter over a clear filter is still a good path forward for full spectrum photometry like this, and creating a true flat using a relatively blank star field, a lot of images and an abnormally high dither in between them will likely get me in a good place for images through my I filter. Thanks again, Mike PS - This camera is insane. In every good way.
Your image looks great, and the "EtalonFlat" is kinda cool! This is one of those "tinker until it works" type of thing. Subtracting it sounds reasonable. Glad to help. And yes, those sensors are insanely sensitive.
