Variable star CCD measurements and field rotators

Hi Derek, it's an interesting point, I presently use a classical Gem but if I decide to go to some larger scope I would certainly consider  an Alt Az mount. I know a colleague here in France using an Alt Az mount and an 8" Newtonian. He doesn't use a rotator but stack relatively short exposures ( let say 15 sec.). At such short exposure there is no rotation issue and his photometry looks good. We could make the calculation of the resulting rotation trail depending the position in the sky. In the past I made a small software for it, if I remember well, in most cases 30 sec were ok for imaging. But photometry doesn't need perfectly round stars, most of the time we defocus, and I don't expose more than one minute due to saturation risk.  It should work but we have a problem at zenith, annoying as it's the best sky place for photometry !  

Next what about rotator and flat ? Maybe not so critical, the true issue is dust, but normally dust at level of primary pupil is no issue for photometry, they are a problem only near the sensor. At F/4 (my case) and one inch or two from the sensor, the resulting modulation depth is well below one mmg for the largest possible dust. Next is the vignetting, this is very depending of the scope type, but one more time most of the vignetting occurs at level of the focuser and some at the secondary. Then it's a problem only if the optics is not right of axi-symmetry. A refractor should be ok, a newtonian not sure, would need some good alignement of the primary and secondary. Anyhow a rotator is a cost and some constraint, I would bet on relatively short exposure and averaging results (what I normally do with the Gem). 

Clear Skies !

Roger (PROC)

Hi Roger

Thank you for your comments.  When doing lightcurves, one tries to keep the star on the same pixels in order to minimize any variations and flat fielding issues.  A good rotator and good guiding should be able to keep the stars in the same place on the detector.  You are right about most of the dust issues are close to the detector and should rotate with the detector.  You are also right the the vignetting  will likely not rotate and that may cause issues over the course of the night.  Good flat fielding will help but are flat fields really ever good enough?

It seems to me that only experience will be able to get an answer for this one.  I cannot believe that this have not been tried at the amateur level.  The question just has to be asked to the correct ear.

Thanks

Derek

I guess this calls for experiments done with your individual hardware, because unless you find someone with the exact same equipment, the vignetting could be different. When making flats you want to make several exposures anyway, so you could take note of the rotation angle(s) and then test whether the disagreement between averaged flats at differnt angles is significantly greater than between averaged flats (same number of course) at the same angle. I guess for that experiment you would want a rather stable (over time) light source for the flats and not e.g. sky flats. Do you have a light box?

CS

HB

Well that is the rub.  I agree, the best way to answer this problem is to try it out and see.  However, I do not currently have the scope and field rotator.  I have the opportunity to participate in a group scope project which involves the purchase of a scope in the 20" class.  We are weighing the merits of various available mounts. The new Planewave L-500 would hold the mount in the Alt Az configuration allowing the use of a smaller building, saving money.  The alt az mount would necessitate the use of a field rotator.  This would be fine for general imaging but I am not sure that it would meet the rigid requirements of exoplanet work.  Mounting the L-500 on a wedge would eliminate the need for the wedge but would take up much more room and require a larger observatory building.  This is where I am trying to probe the experience of others before we decide on which mounting method to utilize. 

Clear skies

Derek

I see. Let me put it this way: If you have the chance to get a 20"ish scope for the group, go for it, don't let the rotator/flat-fielding issue (if any) stop you. This is something to take care of in the "commissioning" phase of the telescope, and at the worst you will be able to resolve this (as discussed above) by either using a set of flatfields done at a bunch of different rotator angles (say every 10deg) or by taking shorter unrotated exposures and rotate in software during stacking. For most targets , with 20" of aperture, you will not want to do very long exposures for photometry anyway to avoid saturation. We are talking about a half-meter class telescope here, after all. Especially if your group includes people interested in spectroscopy, this kind of aperture is very useful and you would not want to compromise on the apetrure just to avoid a solvable problem wrt. photometry. Just my 2 cents.

HBE

As mentioned by Heinz-Bernd and Roger, the primary issue with field rotation is any change in the optical path.  Dust is pretty irrelevant since most telescope optical elements are far enough away that the shadows are very large and diffuse.  The two effects that usually come into play are vignetting and the spider vanes.  Vignetting is the worst, as you have to have very careful collimation, with the central peak exactly placed in the center of the CCD (and the sensor placed exactly on the center line of rotation) in order for flatfielding to work at all angles.  On the WIYN 3.5m telescope, I usually took a set of flats in 8 positions and used the closest set when calibrating my images.  The effect there was a few percent, usually around 2%; important when doing all-sky photometry or a very long time series.  That was a professional telescope, extremely well built, and my guess is that rotation will be a somewhat larger issue with amateur telescopes.  You also need to ensure that the rotator is very robust, has smooth and repeatable action, and can handle whatever weight you decide in the future.

Interestingly enough, you can consider a GEM system as one with a 180-degree field rotation upon its flip.  In this case, the entire optical system flips, but the gravity load stays the same, and the culprit is any flexure in your system modifying the optical path.  With most amateur systems, a GEM flip will change the photometry by a few percent, which is why you often see a break in your time series light curve.  It is worst for those systems with lots of vignetting.  It might be interesting to take flats on both sides of the axis, and see if proper application of the flats makes a difference!

That said, I like fork mounts the best because they don't flip and are compact, and alt-az systems second-best for the same reasons.  The alt-az adds the complexity of the field rotator, and often adds complexity in autoguiding (you have to include rotation, and it is actually better to use two guide stars to handle both rotation and translation).  Alt-az is far easier to use near the horizon and towards the poles than even a fork mount, though the zenith can be a challenge.  With modern equipment, these issues are far less important than when I had to write the software!  I wouldn't worry about poor photometry with the L-500.  Test, and if you do see any low-level effects that impact your science, experiment to find the set of calibration frames that solve the problem.

Arne

Hello Arne

Thanks for your answer to this question.  Are the scattered light effects included in your definition of Vignetting?

Gary

Hi Gary,

Scattered light is included in the vignetting function, but it is a very challenging problem for photometry.  For flat-fielding, the incident light is uniform across the field of view, and so the scattering occurs uniformly on whatever surfaces are prone to reflection.  Usually scattering is cylindrically symmetric, and so you get a brighter peak at the field center.  Vignetting gives fainter corners, so the effect is the same and it just looks like stronger vignetting.  If you have a single surface (like a mirror clip) that is causing scattering, and it is not cylindrically symmetric, then field rotation will obviously change the location of that scattering and systematically contribute to the photometric error.

However, when viewing the night sky, you will get scattering from many sources, not only the target star.  For example, the moon may be out of the field, but its glow is scattered into the camera.  This night-time scattering primarily contributes flares or gradients in the image, which in most cases can be subtracted out as part of the sky background, as they are not transmission errors for the target star.  You see this very often when observing near full moon, as the dust gets illuminated differently and leaves donuts in a flattened image.

It is best to carefully look at your images, and your optical train (eyeballing things from the focus tube works pretty well), to eliminate as much scattering as possible with judicious flocking, painting, and baffling.

Arne

Thank you Arne, Roger and Heinz-Bernd.  This is the response I was hoping for. Arne, I appreciate your experience based reply.  You have given me great information which will allow us to make the right decision. 

I am finding that the study of the sources of error in photometry to be quite fascinating.  There must be a screw loose somewhere!

Thanks again and clear skies!

Derek

Yes, I had toyed with the idea of going to CCD photometry in the past, but remain with visual for these reasons of extreme depth of knowledge and technical complexity needed to obtain high accuracy. I think the average AAVSO photometrist really isn't fully aware of all the confounding factors that reduce accuracy, and why the typical CCD observations submitted via WebObs are not much different in accuracy than those of a good, experienced visual observer!

Mike