Affiliation
American Association of Variable Star Observers (AAVSO)
Tue, 08/05/2014 - 16:27
I am just beginning to collect Ic data on miras. I am finding a huge difference in Ic and V, with the Ic observations near (but not at) saturation with integrations as fast as 5 seconds (with stacking to avoid, hopefully, scintillation). So, I am wondering if I am collecting bogus data (can delete) or useful data. I would be glad to share the images with any observer with Ic experience who would care to examine them and give me feedback.
Ed
Ed What Object are you targeting! possible Mag?
are you using the Atik?
through Skype is much faster!
Hi Joao:
Yes, its the ATIK, I really like that camera. See my post to Brad about the targets.
Ed
Well, the data being valid or not is a much deeper question but I may be able to shed some light on whether the measurements make sense in general. Miras are often listed in Simbad with B and V values that would give a B-V index somewhere around 1.2 or a bit larger. You have to look up the sources to understand these values. They may be averages or max values. The max V and max Ic values probably do not occur simultaneously. More importantly the relationship between V and I can change a lot between V max and V min because of temp changes and the accompanying formation of TiO bands in the V filter range when the star is cooler. R Tau, for example is about 3.5 mags brighter in Ic than V near maximum which is a flux ratio Ic/V of about 25. Near minimum I would expect the ratio to be larger but in the last 10 years there were no Ic measurements near minimum in the AAVSO database. Mira has a V-I of approximately 3 magnitudes at V max and the difference increases to about 5 magnitudes at V min. That gives a flux ratio near V min of about 100. See the attached Plot from AAVSO Data.
So in general, the big difference doesn't seem at all extraordinary, particularly near V min of the light curve.
By the way, I strongly recommend the AAVSO CHOICE course Variable Star Classification and Light Curves if you haven't taken it already. Among other things, it covers the sort of light curve analysis that can answer the question you posed.
I hope this helps.
Brad Walter, WBY
Very informative, Brad, thanks for sharing the data. My problem is that there are not much data in Ic for the two I have measured so far RY Lyr and RY Oph. 600 day light curves are attached (WEY with "+"). The results look reasonable, but I would love for someone to actually look at the images to keep me in line.
I can see from your data that miras play a different game in Ic, and I appreciate your explanation. I think this makes adding Ic to the data collecting more interesting in my continuing variable star education.
I completed the light curve course last year. Mike did a great job. I am on to VStar this Fall to further my educaiton on light curve analysis.
Ed
First, Let me agree with your conclusion about Ic data I think it would add a lot. By the way, all of that stuff about cooling and TiO bands came right out of the VSC&LC course manual. I'm not that smart, but I sometimes remember stuff, particularly when it makes sense and I think I might need that bit of info some day.
From the small sampling of Ic data other than yours, you can already see the familiar pattern starting to form in both curves. V-I somewhere around 3 mags maybe 3.5 mags near maximum but varying somewhat among cycles and the difference increasing to probably somewhere around 5 near minimum. The difference near minimum is really a guestimate due to the small number of points, but you can see that it is larger near minimum than at maximum.
So How to evaluate whether your camera and filter are performing in IC. We have the lovely Landolt northern standard stars now. Nothing against APASS but the Landolt standards usually have a bit higher precision.
There are two that are well positioned near the meridian in the early evening at the moment, SA 38-326 and SA 38-358. They are relatively close in the same FOV, relatively bright at about with nearly the same V mag (approx 9.9) but very different V-I indexes. Image the field centered between these two stars several times with increasing exposure from relatively shallow image depth to saturation Measure the raw V & I magnitudes of both stars. Check linearity in both filters for both stars, and Check the color indexes of both stars. I think you should be able to take the images in a short enough time that you don't have to worry about changing airmass. Even if the airmass does change too much using two stars with very different V-I values will allow you to correct for first and 2nd order extinction. You might even wait a few hours till they get out to airmass 2.0 and repeat some of the exposures (short mid range and just before non linearity begins to get a longer base line for your extinction corrections, if needed. At what max pixel value do the stars start to saturate in the two filters (it may not be at the same level for both filters)? Do the V-I indexes stay constant (within error limits) until the CCD starts to show non linearity in one of the filters? This will let you know where saturation starts in Ic as well as V so you can stay below the lower of the two. It wouldn't hurt to repeat with all filters to have the same info for the others as well. If everything looks consistent, the filter and camera should be giving good data as long as you stay below the significant non-linearity starts to shop up in any of the filters you plan to use for your observations. . .
If you want to stay up late and do the acid test, you could point the camera at NGC 7790 after 3:00 a.m. and take a few images near the meridian to calculate transformation coefficients but you should do this on at least three good nights. If Tv is close to zero and the Tvi is close to 1.0 and they are consistent between the nights then I would say the camera and filter are performing well and the data is good. If the transformation coefficients seem too far out for the Ic filter (or the V filter for that matter) you may lose some accuracy for very red stars.Check to see that the filter is clean and check the manufacturer's filter curve. the normalized product of the filter transmission curve and the with your camera response curve, the resultant overall response curve might have a long red toe or it might turn out to be an Johnson I filter rather than a Cousins/Bessel.
Brad Walter, WBY
Excel File of Landolt northern standards is attached.
display
results from simbad
display data for RY LYR
http://simbad.u-strasbg.fr/simbad/sim-id?Ident=%402868126&Name=V*%20RY%…
but trying Vizier may get you the results you're looking for.
the spectral data can be extracted from here
http://cdsxmatch.u-strasbg.fr/gadgets/ifr?url=http://cdsxmatch.u-strasb…
my usual approach is expose closer, and analyse if there is clipping in the Star curve to be studied.
I hope it help's you ED
your data looks good to a novice like me, i think Arne or other senior members are more enlightned in this subject, but brad shed some light on the expectations
best regards
JAC
Ad Astra
Excellent, Brad. Confession: I have not performed a transform on these new filters. I must be getting lazy. I usually do this and, in fact, trained myself on imaging Landolt fields when I first got started. I will do so ASAP.
Some other advice sought. Just how "fast" an integration time will produce good results? I have been setting the limit at 5 seconds with multiple subs to guard against scintillation. Good practice? Bad practice? Even faster integration times??
Ed
I think 5 seconds is plenty long enough to avoid scintillation effects.Are you familiar with the Dravins papers. the Auto correlations for a 20 CM telescope pretty much hit the minimum (at La Palma) in under 100 milliseconds. Even for a 60 CM telescope they have pretty much disappeared in the same time frame. I think shutter effects may be more important for our sized telescopes below about 5 or 10 seconds. If your camera uses an iris type shutter you need to know the closing and opening times of the shutter to determine the minimum exposure time you need to make systematic error from shutter operating times insignificant compared to other noise factors. Suppose the opening and closing time are equal at 0.01 seconds, The combined effect of opening and closing between center and edge of the chip could be 0.020 sec. if you want to keep the systematic error from shutter effect to 0.1% (about .001 mags), then your minimum exposure needs to be about 20 seconds. I have no idea as to operating times for different shutters. Iris tend to be the slowest and the SBIG rotating "photometric" shutters keep any differential very by the shape of the rotating window. I recall Arne recommending in an e-mail a minimum exposure time for flats of 10 seconds for shuttered cameras other than SBIG (or cameras with a similar rotating shutter mechanism). DSLRs, like film cameras, have negligible shutter effect because they are designed to provide even exposures across the detector surface for exposures of a small fraction of a second.
Figures 14-16 from Dravins, et. al. 1997 PASP 109 - 173D, the first of the three Dravins papers on scintillation, are attached.
Brad Walter, WBY
Ed,
Are you aware of the Castelaz paper on Mira spectra,
http://iopscience.iop.org/1538-3881/120/5/2627/fulltext/980514.text.html ? If you double click on any of the small spectra icons it links you to much larger images and then if you clock on the text hot spot below individual images you can get a high res image of that particular one. Of course, an individual spectrum will be dependent on the phase at which it was taken but for manystars there are multiple spectra taken at different phases. IOP has done a really nice on-line presentation of these to allow you to see them and download them at different resolutions. If the AAVSO data base has decent light curves of these stars you will be able to determine the phases at which they were taken. It seemed to me this imight pretty interesting stuff for a Mira admirer.
Brad Walter, WBY
thanks Brad
quite interesting the variation accordingly with the period, do you know of any study applied to eclipsing binaries where one could study the diference in the spectrum of the two stars. in theory one can diferenciate the spectra during ocultation and transit of one of the members to the other.
i'll take look later on.
but was glad to see your hint.
Best regards
CJGB
I am sure that there is a large literature on the spectra of eclipsing binaries. They are, however, much more of a melange than Miras. One would expect Miras to have much greater similarity between objects and, therefore, more global and generic studies would exist. I suspect there will be a wide range of spectrographic studies one would have to review for eclipsing binaries, each focused on a particular narrow subset. a review of spectrographic studies of the various types of eclipsing binaries in Vizier should give you an idea of the scope of the available literature.
Brad Walter, WBY
Brad: Thanks for all the feedback. I really appreciate it.
Ed
Interesting thread guys, lot's of very helpful info.
I've just selected a few Mira's, some with Humps and a few other Mira's to image in B,V and I'm going to try I too.
Douglas.