Thu, 03/13/2014 - 20:25
I am planning to do relative (not absolute) photometry and I want extremely accurate readings since the variations I am looking for are extremely small. Given that requirement, I would be interested to know what others recommend regarding Flat frames. Should the Flat frames be taken at the same temperature as the light frames? Or is that not necessary? Thanks.
Ed
Hi Ed,
Can we get a bit more information? What kind of targets are you looking at, with what equipment, what filters, what cadence?
Generally, temperature is a 2nd or 3rd order effect on flats. The primary change is in the thickness of the filter, which moves the bandpass. For extreme standardization efforts, keeping the temperature constant (and in fact, placing the filters in a parallel beam rather than a converging one) is important, but not for general differential photometry. Far more important are scintillation, consistent image quality and focus.
Flat field frames always add random noise to any image processed with them. What they remove are all of the systematic effects: vignetting, shutter errors, dust donuts, pixel-to-pixel sensitivity variations, etc. I consider them essential for any CCD observing, but for extreme precision, there are many considerations (such as extreme signal/noise in the flats).
Arne
Hello Ed;
First I would suggest that you study the difference between Very Accurate Photometry and High Precision Photometry. For exolanets, you do not need accuracy, what you need is high precision with minimal systematic errors in the light curves. (for more info, wiki accuracy and precision or resolution).
Second, I would suggest that you observe several known transits of exoplanets to determine how well your observing technique is. First try 2% transits, then 1% transits, then go for 0.5% transits and finally 0.2% transits. This will show what you should expect and will allow you to asses your systematics.
How small are the "minute dents" in the light curves you are looking for?
Sounds like you have already read Bruce Gary's book on exoplanets.
Gary
Hi Ed,
I am not an expert in observing exoplanet transits but I have followed a few transits with more experienced observers. As regards you question about exoplanet transit predictions and database plese see: http://var2.astro.cz/ETD/
You can find all the information needed to follow a known transit, just type your longitude and latitude.
Following an exoplanet transit producing good quality data is difficult and challenging at the same time. While there are few easier exceptions, the light of the hosting star dims generally by only about 0.01-0.02 mag and possibly less. As already pointed out, scintillation and statistical noise are the key factors that conspire against good quality data. There are many opinions among experienced observers on how to reduce those effects. For what I could see defocusing the star so that you can increase the exposure time in the range of 100-200 secs is effective in reducing the effects of seeing as it is effective the use of a photometric bandpass filter, R and V seem a good choice. The choice of good comp stars is another important factor. Many people use ensemble photometry to average out small gradients in the image. You mention the problem of calibration. Well, although good calibration is essential in reducing vignetting and dark current you are supposed to track your target during the whole observing session. If your tracking system is good, calibration is not the determining factor to achieve good results. By the way a few observers have discovered new variable stars following exoplanets transits. In fact small variations of light of the comp stars are relatively easy to detect as they appear to alter the light curve of the star you are following.
Hope this helps
Gianluca
another good source you probably know of already is
http://www.exoplanet.eu/
and another one I really like that you may not already know is
exoplanets.org/
I don't think you want to extend your integration times to any longer than the 100 to 200 seconds you lose too much temporal resolution.
Attached are three papers on Scintillation by Dravins et al. They reference earlier work by Young whick is also excellent
Brad Walter
Hello Ed:
When you say " (for more info, wiki accuracy and precision or resolution)"---I mean to google on the terms and look also at wikipedia--they have a good discussion on the difference between these two terms.
Gary
I think Gary is directing you to the following website on Wikipedia:http://en.wikipedia.org/wiki/Accuracy_and_precision.
Precision is a measure of the "spread" of your sample. The statistic most commonly used is standard deviation when your are dealing with the spread of a sample or standard error if you are dealing with the standard deviation of a statistic (means, for example). Variance is also sometimes used as a measure in some calculations. Variance is the square of the standard deviation (more correctly, the standard deviation is the square root of the variance).
Accuracy is the agreement of your measurement, or the mean of your multiple measurements to the "true" value.
Suppose you measure the length of a table 10 times with a very fine scale tape measure. You may have a precision that is quite good - say a 64th of an inch - but if you recorded your lengths forgetting that you started all your measurements at the 1 inch mark (a systematic error) your accuracy is poor.
On the other hand if you used a tape measure with much coarser divisions that allowed you to estimate your measurement to 1/8", you didn't make the one-inch mistake, and the mean of your 10 measurements was very close to the "true" length of the table, then your measurement would be relatively accurate but your precision would be much lower. Here's the rub: unless your are measuring a standard with a known "true" value, precision limits accuracy because you don't have any way of knowing what the true value is. Your measurements provide the estimate of the true value within their precision PROVIDED THERE ARE NO SYSTEMATIC ERRORS affecting accuracy. Subsequent much more precise measurements that take into account all significant systematic errors might establish the accuracy (or lack thereof) of your measurements through more precsise estimates of the true value. Significant systematic error in this context means tht the total effect of uncorrected systematic errors is much smaller than the standard deviation of the measurements and would not cause the precsion of the measurements is the limiting factor in estimating the accuracy.
I can suggest two excellent references dealing with error analysis in the physical sciences. The first is an "old standby" little reference text Data Reduction and Error Analysis, by Bevington and Robinson. This covers a wide range of error analysis topics and is meant to be a reference more than a teaching text.
The second, An Introduction to Error Analysis, by John Taylor, is a more elementary text but has better explanations.
I strongly second Gary's recommendation that you capture light curves of some known exoplanets starting with bright stars with transit depths approaching 20 millimags (2% flux) and then progress to ones with smaller transit depth. In the physical sciences 3 standard deviation null hypothesis probability is generally considered the minimum acceptable probability to establish more than an indication or possibility. That means as a quick order of magnitude estimate, you need 7 millimag precision for your measurements to establish a 20 millimag transit and 3 millimag precsion for a 10 millimag transit depth.* I have done this with scopes ranging from 250 mm to 0.6 meters and even with the 0.6 meter telescope capturing exoplanet transits below about 7 or 8 millimags with sufficiently fast cadence and sufficient SNR to identify the 4 contact points and capture a good estimate of the minimum is tough. You need good conditions.
Brad Walter
*These values would be correct comparing a single sample against a known standard value (a Z test) or comparing two sample values (t test of sufficient sample size). For a light curve you aare calculating the probability that the coefficients of the best fit model are different from zero or whatever the coefficients are for the comparison curve determined by the out of transit observations. This statistical analysis is fairly complicated for a transit light curve because the model is non linear. That's why its great to have resources such as the Exoplanet Transit Database website that has developed a reasonable model including a simple model for limb darkening to apply to our data.
Sorry if this is redundant. I had it all cued up to send this morning but didn't hit the save button. Of course when I did hit the save button this afternoon, I didn't refresh first or the post would have been lost.
Hello,
when searching 'high precision photometry' as title words in NASA ADS http://adsabs.harvard.edu/abstract_service.html, one gets many search results. A lot of them are not freely accessible, but some are for sure (e.g. those which have arXiv resources (X after the bibcode). For example some older or more important(?) papers are. There are two nice papers I found, their references give more hints to follow:
http://adsabs.harvard.edu/abs/2009MNRAS.396.1023S (Southwoth et. al 2009)
http://www.jstor.org/stable/10.1086/323387 (Everett and Howell 2001)
About precision and/or/vs accuracy, I find very clear the analogy of shooting targets. At least it helps to memorize those two things :-)
http://climatica.org.uk/wp-content/uploads/2012/02/Accuracy-vs-precision1.jpg
Tõnis
I am also looking at exoplanets. I have attached a database you can use with TheSky of all the recently released Kepler known exoplanets. I could not upload the file with the .SDBX extention, so it has the txt extention. Just rename using the .SDBX extenstion before trying to use.
Frank
I think Tom is talking about including the exoplanet data in either SkyX or SkyX's associated CCD control program; most likely in the FITS header. I'm not sure how that would increase the accuracy of your photometry. It would make historical analysis a little easier afterwards.
James
This has been a very interesting discussion. I wanted to add a few words. Ed's system is really quite excellent - you won't find many amateurs better equipped. Detecting an exoplanet transit with such a system is almost trivial. Applying any typical flatfield, and looking at one of the higher-amplitude events, will yield a nice light curve. It sounds like Ed has also read Bruce Gary's book, and so will know the basic techniques of photometry, as well as knowing to observe the host star well before and after the transit to derive a good baseline for analyzing the dip. Other posters have given great websites where you can find empherides to plan each night - there is always an exoplanet transiting somewhere! I've done many exoplanet systems, and it is both exciting to see the results of a planet around another star, as well as a photometric challenge. I encourage all digital imagers to give an exoplanet transit a try!
That said, the CBB filter has use for very few projects outside of exoplanet transits. Please be sure to purchase a few more filters, especially photometric ones (Johnson/Cousins preferred, Sloan afterwards) and consider doing photometry of some variable stars. They can be equally exciting - far more exotic light curve shapes and much more necessary interpretation to understand the underlying star system. In fact, there are several host stars that are intrinsic variables, and those make nice dual targets.
Best of luck, Ed! Always feel free to ask here if you have additional questions.
Arne
I took it off the official kepler page and parsed with TheSky. The Excel spreadsheet is attached that was the source. These are the confirmed exoplanes only. There is a much longer list of all suspected exoplanets.
Frank
http://kepler.nasa.gov/Mission/discoveries/
BTW that site is very cool, if you follow the KOI link you get a table of all the data for the planet. Then if you follow the link at the top labeled "transits", you get a calculator that will give you the timing of the next transits for your location.
Frank
The data available is very complete, and you can see the amount of "dip" anticipated. So yes you should be intellegent in your choice so that it matches your available equipment. For myself I am looking at all the photometry applications to see where i want to focus my efforts, and what equipment I want to invest in. For now it is just playing with the iTelescope sites and the incredible telescopes available there.
Frank
While exploring the Kepler database and looking at the lightcurves is a lot of fun, I would strongly recommend that you do not use the Kepler exoplanet transits for your groundbased program. Look at the scale of variation - Kepler is a unique instrument that can perform micromagnitude precision photometry. Far better for your use will be the ground-based discoveries, such as those from WASP, XO, TrES, etc. HD 209458 is a classic example. Pick those with ~0.02mag dips, and look at the finding charts to ensure that there are nearby equal-brightness stars for use as comparisons.
Arne
Here is a link to Bruce's book. http://brucegary.net/book_EOA/x.htm You can download it in PDF format and convert it to Mobi (kindle) or some other so that you can read it on a tablet or reader if you want to. Caliber is a free conversion program that converts from one format to another. Amazon also sells a version for kindle if you want to pay a few bucks. I would also just suggest emailing Bruce! He is very personable and is very good to answer questions. There is a link on his book page to other pages and somewhere you can find his email address. I got to visit with him several months ago in his home and am very impressed with his level of expertise. He has put up with my inane questions and been very helpful in my projects. Just a note. I am rereading his book for the 4th time right now! It seems as I progress up the learning curve I seem to find new ideas that I missed before. It is a handy reference even for those of us that don't do exoplanet searches! Paul Temple TRO Espanola, NM
In the fall of 2012 I started putting together a list of exoplanets that I thought would be good candidates for amateur observers. The criteria for the "yes" entry in the project list were
1. 14th v mag or brighter
2. Transit depth 0.01 mag or greater
I don't think that I set any limits for orbital period.
The iedea was to get a list together and then start establishing sequences for all of these transiting exoplanets. I was (and still am) a newbie at sequence generation and Sebastian Otero and Arne were helping me get my arms around the task. A lot of checking and verification is reqauired even before you start to generate the sequences.
The list is attached filled in to the point I stopped. I was still working full time then and could not devote the time needed. It is an interesting project. It's a lot of work for one person burt I plan to start working on this again if Sebastian and Arne can stand it. If there is sufficient interest, perhaps a group of us could work together, update the list and complete the sequences.
Brad Walter
I calibrated a set of about 22 exoplanet fields with Sonoita at BVRI, back in 2008-2009. I think the list below is pretty complete, and that would probably be a good starting point for exoplanet observing, and the photometry should be in Seqplot. Most of these fields are included in Brad's spreadsheet. After the field name are the BVRI magnitudes of the host star, plus a short comment when necessary.
Arne
--------------------------------------------------
HAT_P_1 10.317 9.736 9.398 9.097 double
HAT_P_2 9.198 8.714 8.455 8.196
HAT_P_3 12.452 11.527 11.016 10.596
HAT_P_4 11.836 11.208 10.864 10.549
HAT_P_5 12.620 11.957 11.588 11.263
HAT_P_6 10.841 10.405 10.139 9.887
HAT_P_7 10.991 10.475 10.189 9.912
HD17156 8.834 8.187 7.836 7.545
HD149026 8.795 8.200 7,863 7.583
HD189733 8.570 7.672 7.155 6.714
HD209458 8.217 7.647 7.325 7.026
TRES_1 12.603 11.741 11.265 10.869
TRES_2 12.048 11.401 11.035 10.687
TRES_3 13.099 12.364 11.943 11.571
TRES_4 12.112 11.581 11.266 10.979
WASP_1 12.239 11.656 11.312 10.989
WASP_2 12.652 11.783 11.284 10.855
WASP_3 11.055 10.563 10.266 9.977
XO_1 11.874 11.186 10.791 10.428
XO_2 11.965 11.132 10.675 10.301 wide double
XO_3 10.233 9.831 9.588 9.338
COROT_2 13.347 12.454 11.928 11.434