There may be another way to identify Over Contact (OC) and Semi-Detached (SD) eclipsing binary stars by their light curves (LC). What appears to be emerging, with the initial finds, is shallow eclipsing binary stars with a different type of pattern (not currently discussed at PH) in the LC that are later classified through the periodogram. Kianjin has been performing the analysis.
Subsequent to this article being written, forum member Nighthawk_black posted information on a recently published science paper by S. Faigler et al (10 Oct 2011) that discusses seven newly discovered shallow transit **non eclipsing binary systems** using a newly developed BEER algorithm. There is a possibility that many of the Type G stars being collected fall under this category. You can read their paper [here](http://arxiv.org/abs/1110.2133 "").
Below are two light curve analyses performed by Kianjin:
APH21101216
Kinajin's comments; "The light curve is a little ragged and noisy, possibly because this is a faint star (mag 15) in a somewhat busy neighborhood, but no close stars appear to be contaminating it specifically. The waveform does look like an OC, with alternating minimas and maximas. The period is not too short for an OC, the shortest period OC listed by Prsa is easily half of this period."
Thus far eight possible EBs have been identified looking for both an enhanced amplitude in the light curve and also minor vertical spacing in the individual light bars. Some, as the one above, are more pronounced. Another common feature with these LCs is that the stars have all been spectral type G dwarf. The only exception is one type K star found by members tucanico and ggccg.
Here is another LC example. In this case the periodogram actually shows a drop.
APH23068012
"This one definitely has the alternating minimas of an OC EB. Just that the eclipses are so shallow, just 0.05% to 0.1%!. Light curve is quite ragged too, but Skyview shows no contamination."
Keying simply on amplitude enhancement in the LC is not enough though it is one of the criteria. I have only found three visual commonalities; amplitude, spectral type and vertical light bar spacing. There are many misses using this method but even so, finding eight potential EBs is encouraging, roughly one out of every 1500 to 2500 (more data needed) general light curve classifications. One last insight in searching potential LCs for these eclipsing binary stars is that one must move from the LC "Classification" page at Planet Hunters to the "discuss this star" page in Talk to look at the LC in more detail. For some reason the LC pattern is easier to discern. I also look at the "View Star" section in Talk to compare with other LCs through the quarters for collaborating tells before deciding to run the Kepler Identification number in the periodogram service.
Below is the link to the collection that will be updated as new finds are made. With the help of Kianjin and other Forum participants, we may be able to not only flag these specific EBs in the future but also develop a workable algorithm for screening the Kepler data. There may now be enough information in this light curve collection to connect the dots, pun intended.
[OC-SD shallow eclipsing binary star collection](http://talk.planethunters.org/collections/CPHS0002p8 "")
Here is the link for checking out your finds using the Kepler Identification (KID) number and generating a periodogram:
[Periodogram Generator](http://nsted.ipac.caltech.edu/applications/ETSS/Kepler_index.html "")
There may be another way to identify Over Contact (OC) and Semi-Detached (SD) eclipsing binary stars by their light curves (LC). What appears to be emerging, with the initial finds, is shallow eclipsing binary stars with a different type of pattern (not currently discussed at PH) in the LC that are later classified through the periodogram. Kianjin has been performing the analysis.
Subsequent to this article being written, forum member Nighthawk_black posted information on a recently published science paper by S. Faigler et al (10 Oct 2011) that discusses seven newly discovered shallow transit non eclipsing binary systems using a newly developed BEER algorithm. There is a possibility that many of the Type G stars being collected fall under this category. You can read their paper here.
Below are two light curve analyses performed by Kianjin:
APH21101216
Kinajin's comments; "The light curve is a little ragged and noisy, possibly because this is a faint star (mag 15) in a somewhat busy neighborhood, but no close stars appear to be contaminating it specifically. The waveform does look like an OC, with alternating minimas and maximas. The period is not too short for an OC, the shortest period OC listed by Prsa is easily half of this period."
Thus far eight possible EBs have been identified looking for both an enhanced amplitude in the light curve and also minor vertical spacing in the individual light bars. Some, as the one above, are more pronounced. Another common feature with these LCs is that the stars have all been spectral type G dwarf. The only exception is one type K star found by members tucanico and ggccg.
Here is another LC example. In this case the periodogram actually shows a drop.
APH23068012
"This one definitely has the alternating minimas of an OC EB. Just that the eclipses are so shallow, just 0.05% to 0.1%!. Light curve is quite ragged too, but Skyview shows no contamination."
Keying simply on amplitude enhancement in the LC is not enough though it is one of the criteria. I have only found three visual commonalities; amplitude, spectral type and vertical light bar spacing. There are many misses using this method but even so, finding eight potential EBs is encouraging, roughly one out of every 1500 to 2500 (more data needed) general light curve classifications. One last insight in searching potential LCs for these eclipsing binary stars is that one must move from the LC "Classification" page at Planet Hunters to the "discuss this star" page in Talk to look at the LC in more detail. For some reason the LC pattern is easier to discern. I also look at the "View Star" section in Talk to compare with other LCs through the quarters for collaborating tells before deciding to run the Kepler Identification number in the periodogram service.
Below is the link to the collection that will be updated as new finds are made. With the help of Kianjin and other Forum participants, we may be able to not only flag these specific EBs in the future but also develop a workable algorithm for screening the Kepler data. There may now be enough information in this light curve collection to connect the dots, pun intended.
Tom, I will dig through my collection bin tonight I might have a couple more worth examining per this discussion.
I agree contamination is not likely an issue with the examples above as far as the raggedness--I wonder if it's possible the OC situation is introducing a little more variability/instability into one or both components than we normally might see for their type...?
Tom, I will dig through my collection bin tonight I might have a couple more worth examining per this discussion.
I agree contamination is not likely an issue with the examples above as far as the raggedness--I wonder if it's possible the OC situation is introducing a little more variability/instability into one or both components than we normally might see for their type...?
I tried to wrap my head around this new method for screening OC and SD EBs but failed to come to any satisfactory conclusions. I then realized that I was trying to break down this method into an 'algorithm' or recipe that can be boiled down to a few steps.
Not so. I realized after a while that my problem was trying to think too much with the left brain. What I think is that Tom has found is a way to use our superior perception and pattern matching abilities to find certain signatures of OC/SD EBs in a raw light curve. The problem is, a lot of this pattern matching happens in our subconscious mind. We can't reverse engineer what our brains can do so well. (cf. for example [the art of chicken sexing](http://cogprints.org/3255/ "") (!) which relies on this subconscious pattern matching through perceptual cues which can be taught)
Most of these light curves look like a long band of fuzzy noise, but look closer and there is some structure there. And since I'm organized in a left-brained way due to my past lives in science and engineering, I took the phased curve of KID 8654079 and tried to boil it down. Here's what I found:
You can see that there is a clustering of data points around the midpoint to slightly above the 1/3 level where it then fades off quite abruptly. The lower third of the light curve is divided into a slightly less thick density band and the last 20% of the light curve shows a more gradual fade out. The histogram essentially takes the various parts of the phased curve and organizes it by the clustering of data points. Since the shape of an OC remains quite consistent, it's possible to use this then as a guide to pick out light curves like this.
Compare this to the light curve of a DSCT - it's generally a noisy fuzz that's not differentiated. That's because the waveform is usually a sinusoidal curve that's itself quite spread out.
I tried to wrap my head around this new method for screening OC and SD EBs but failed to come to any satisfactory conclusions. I then realized that I was trying to break down this method into an 'algorithm' or recipe that can be boiled down to a few steps.
Not so. I realized after a while that my problem was trying to think too much with the left brain. What I think is that Tom has found is a way to use our superior perception and pattern matching abilities to find certain signatures of OC/SD EBs in a raw light curve. The problem is, a lot of this pattern matching happens in our subconscious mind. We can't reverse engineer what our brains can do so well. (cf. for example the art of chicken sexing (!) which relies on this subconscious pattern matching through perceptual cues which can be taught)
Most of these light curves look like a long band of fuzzy noise, but look closer and there is some structure there. And since I'm organized in a left-brained way due to my past lives in science and engineering, I took the phased curve of KID 8654079 and tried to boil it down. Here's what I found:
You can see that there is a clustering of data points around the midpoint to slightly above the 1/3 level where it then fades off quite abruptly. The lower third of the light curve is divided into a slightly less thick density band and the last 20% of the light curve shows a more gradual fade out. The histogram essentially takes the various parts of the phased curve and organizes it by the clustering of data points. Since the shape of an OC remains quite consistent, it's possible to use this then as a guide to pick out light curves like this.
Compare this to the light curve of a DSCT - it's generally a noisy fuzz that's not differentiated. That's because the waveform is usually a sinusoidal curve that's itself quite spread out.
This paper doesn't deal with with G type SD or OC stars directly, but it talks about a new algorithm for finding short period non-eclipsing binaries:
[Seven new binaries discovered in the Kepler light curves through the BEER method](http://arxiv.org/abs/1110.2133 "")
> We present seven newly discovered non-eclipsing short-period binary systems with low-mass companions, identified by the recently introduced BEER algorithm, applied to the publicly available 138-day photometric light curves obtained by the Kepler mission. The detection is based on the beaming effect (sometimes called Doppler boosting), which increases (decreases) the brightness of any light source approaching (receding from) the observer, enabling a prediction of the stellar Doppler radial-velocity modulation from its precise photometry. The BEER algorithm identifies the BEaming periodic modulation, with a combination of the well known Ellipsoidal and Reflection/heating periodic effects, induced by short-period companions.
As far as I could tell none of the seven have Talk entries yet; four of the three I would guess ppl are classifying as Pulsating.
KID 10848064 folded
_*Edit- adding additional KIDs referenced to BEER detections_
8016222
9512641
7254760
5263749
4577324
6370196
This paper doesn't deal with with G type SD or OC stars directly, but it talks about a new algorithm for finding short period non-eclipsing binaries:
We present seven newly discovered non-eclipsing short-period binary systems with low-mass companions, identified by the recently introduced BEER algorithm, applied to the publicly available 138-day photometric light curves obtained by the Kepler mission. The detection is based on the beaming effect (sometimes called Doppler boosting), which increases (decreases) the brightness of any light source approaching (receding from) the observer, enabling a prediction of the stellar Doppler radial-velocity modulation from its precise photometry. The BEER algorithm identifies the BEaming periodic modulation, with a combination of the well known Ellipsoidal and Reflection/heating periodic effects, induced by short-period companions.
As far as I could tell none of the seven have Talk entries yet; four of the three I would guess ppl are classifying as Pulsating.
KID 10848064 folded
*Edit- adding additional KIDs referenced to BEER detections
Wonderful paper NHB! I asked Meg of the science team if they would review this collection. My hunch is that quite a few of the collected LCs here are **eclipsing and/or non eclipsing binary stars** by their periodogram and analysis by Kianjin. Some are more obvious than others when viewing the light curve. Ggccg and Tucanico flagged some of them.
None of these stars are of the seven listed in the paper. I do think we are on to something very interesting.
Wonderful paper NHB! I asked Meg of the science team if they would review this collection. My hunch is that quite a few of the collected LCs here are eclipsing and/or non eclipsing binary stars by their periodogram and analysis by Kianjin. Some are more obvious than others when viewing the light curve. Ggccg and Tucanico flagged some of them.
None of these stars are of the seven listed in the paper. I do think we are on to something very interesting.
I went ahead and made periodograms for the seven shallow **non eclipsing binary stars** from the S. Faigler et al paper. One thing that seems clear is that all the periods are greater than one in Lomb Scargle mode. When you covert to Plavchan you can see sharp curves with alternating max and min in the majority of the graphs.
Plavchan
Plavchan
Plavchan
Plavchan
Plavchan
Plavchan
Plavchan
In the article I posted, you can see in the periodograms that Kianjin analyzed that they are very fast periods, much shorter than one. This, in itself, is very interesting and unique.
I went ahead and made periodograms for the seven shallow non eclipsing binary stars from the S. Faigler et al paper. One thing that seems clear is that all the periods are greater than one in Lomb Scargle mode. When you covert to Plavchan you can see sharp curves with alternating max and min in the majority of the graphs.
Plavchan
Plavchan
Plavchan
Plavchan
Plavchan
Plavchan
Plavchan
In the article I posted, you can see in the periodograms that Kianjin analyzed that they are very fast periods, much shorter than one. This, in itself, is very interesting and unique.
There may be another way to identify Over Contact (OC) and Semi-Detached (SD) eclipsing binary stars by their light curves (LC). What appears to be emerging, with the initial finds, is shallow eclipsing binary stars with a different type of pattern (not currently discussed at PH) in the LC that are later classified through the periodogram. Kianjin has been performing the analysis. Subsequent to this article being written, forum member Nighthawk_black posted information on a recently published science paper by S. Faigler et al (10 Oct 2011) that discusses seven newly discovered shallow transit **non eclipsing binary systems** using a newly developed BEER algorithm. There is a possibility that many of the Type G stars being collected fall under this category. You can read their paper [here](http://arxiv.org/abs/1110.2133 ""). Below are two light curve analyses performed by Kianjin: APH21101216   Kinajin's comments; "The light curve is a little ragged and noisy, possibly because this is a faint star (mag 15) in a somewhat busy neighborhood, but no close stars appear to be contaminating it specifically. The waveform does look like an OC, with alternating minimas and maximas. The period is not too short for an OC, the shortest period OC listed by Prsa is easily half of this period."  Thus far eight possible EBs have been identified looking for both an enhanced amplitude in the light curve and also minor vertical spacing in the individual light bars. Some, as the one above, are more pronounced. Another common feature with these LCs is that the stars have all been spectral type G dwarf. The only exception is one type K star found by members tucanico and ggccg. Here is another LC example. In this case the periodogram actually shows a drop. APH23068012   "This one definitely has the alternating minimas of an OC EB. Just that the eclipses are so shallow, just 0.05% to 0.1%!. Light curve is quite ragged too, but Skyview shows no contamination."  Keying simply on amplitude enhancement in the LC is not enough though it is one of the criteria. I have only found three visual commonalities; amplitude, spectral type and vertical light bar spacing. There are many misses using this method but even so, finding eight potential EBs is encouraging, roughly one out of every 1500 to 2500 (more data needed) general light curve classifications. One last insight in searching potential LCs for these eclipsing binary stars is that one must move from the LC "Classification" page at Planet Hunters to the "discuss this star" page in Talk to look at the LC in more detail. For some reason the LC pattern is easier to discern. I also look at the "View Star" section in Talk to compare with other LCs through the quarters for collaborating tells before deciding to run the Kepler Identification number in the periodogram service. Below is the link to the collection that will be updated as new finds are made. With the help of Kianjin and other Forum participants, we may be able to not only flag these specific EBs in the future but also develop a workable algorithm for screening the Kepler data. There may now be enough information in this light curve collection to connect the dots, pun intended. [OC-SD shallow eclipsing binary star collection](http://talk.planethunters.org/collections/CPHS0002p8 "") Here is the link for checking out your finds using the Kepler Identification (KID) number and generating a periodogram: [Periodogram Generator](http://nsted.ipac.caltech.edu/applications/ETSS/Kepler_index.html "")There may be another way to identify Over Contact (OC) and Semi-Detached (SD) eclipsing binary stars by their light curves (LC). What appears to be emerging, with the initial finds, is shallow eclipsing binary stars with a different type of pattern (not currently discussed at PH) in the LC that are later classified through the periodogram. Kianjin has been performing the analysis.
Subsequent to this article being written, forum member Nighthawk_black posted information on a recently published science paper by S. Faigler et al (10 Oct 2011) that discusses seven newly discovered shallow transit non eclipsing binary systems using a newly developed BEER algorithm. There is a possibility that many of the Type G stars being collected fall under this category. You can read their paper here.
Below are two light curve analyses performed by Kianjin:
APH21101216
Kinajin's comments; "The light curve is a little ragged and noisy, possibly because this is a faint star (mag 15) in a somewhat busy neighborhood, but no close stars appear to be contaminating it specifically. The waveform does look like an OC, with alternating minimas and maximas. The period is not too short for an OC, the shortest period OC listed by Prsa is easily half of this period."
Thus far eight possible EBs have been identified looking for both an enhanced amplitude in the light curve and also minor vertical spacing in the individual light bars. Some, as the one above, are more pronounced. Another common feature with these LCs is that the stars have all been spectral type G dwarf. The only exception is one type K star found by members tucanico and ggccg.
Here is another LC example. In this case the periodogram actually shows a drop.
APH23068012
"This one definitely has the alternating minimas of an OC EB. Just that the eclipses are so shallow, just 0.05% to 0.1%!. Light curve is quite ragged too, but Skyview shows no contamination."
Keying simply on amplitude enhancement in the LC is not enough though it is one of the criteria. I have only found three visual commonalities; amplitude, spectral type and vertical light bar spacing. There are many misses using this method but even so, finding eight potential EBs is encouraging, roughly one out of every 1500 to 2500 (more data needed) general light curve classifications. One last insight in searching potential LCs for these eclipsing binary stars is that one must move from the LC "Classification" page at Planet Hunters to the "discuss this star" page in Talk to look at the LC in more detail. For some reason the LC pattern is easier to discern. I also look at the "View Star" section in Talk to compare with other LCs through the quarters for collaborating tells before deciding to run the Kepler Identification number in the periodogram service.
Below is the link to the collection that will be updated as new finds are made. With the help of Kianjin and other Forum participants, we may be able to not only flag these specific EBs in the future but also develop a workable algorithm for screening the Kepler data. There may now be enough information in this light curve collection to connect the dots, pun intended.
OC-SD shallow eclipsing binary star collection
Here is the link for checking out your finds using the Kepler Identification (KID) number and generating a periodogram:
Periodogram Generator
Tom, I will dig through my collection bin tonight I might have a couple more worth examining per this discussion. I agree contamination is not likely an issue with the examples above as far as the raggedness--I wonder if it's possible the OC situation is introducing a little more variability/instability into one or both components than we normally might see for their type...?Tom, I will dig through my collection bin tonight I might have a couple more worth examining per this discussion.
I agree contamination is not likely an issue with the examples above as far as the raggedness--I wonder if it's possible the OC situation is introducing a little more variability/instability into one or both components than we normally might see for their type...?
I tried to wrap my head around this new method for screening OC and SD EBs but failed to come to any satisfactory conclusions. I then realized that I was trying to break down this method into an 'algorithm' or recipe that can be boiled down to a few steps. Not so. I realized after a while that my problem was trying to think too much with the left brain. What I think is that Tom has found is a way to use our superior perception and pattern matching abilities to find certain signatures of OC/SD EBs in a raw light curve. The problem is, a lot of this pattern matching happens in our subconscious mind. We can't reverse engineer what our brains can do so well. (cf. for example [the art of chicken sexing](http://cogprints.org/3255/ "") (!) which relies on this subconscious pattern matching through perceptual cues which can be taught) Most of these light curves look like a long band of fuzzy noise, but look closer and there is some structure there. And since I'm organized in a left-brained way due to my past lives in science and engineering, I took the phased curve of KID 8654079 and tried to boil it down. Here's what I found:  You can see that there is a clustering of data points around the midpoint to slightly above the 1/3 level where it then fades off quite abruptly. The lower third of the light curve is divided into a slightly less thick density band and the last 20% of the light curve shows a more gradual fade out. The histogram essentially takes the various parts of the phased curve and organizes it by the clustering of data points. Since the shape of an OC remains quite consistent, it's possible to use this then as a guide to pick out light curves like this. Compare this to the light curve of a DSCT - it's generally a noisy fuzz that's not differentiated. That's because the waveform is usually a sinusoidal curve that's itself quite spread out.I tried to wrap my head around this new method for screening OC and SD EBs but failed to come to any satisfactory conclusions. I then realized that I was trying to break down this method into an 'algorithm' or recipe that can be boiled down to a few steps.
Not so. I realized after a while that my problem was trying to think too much with the left brain. What I think is that Tom has found is a way to use our superior perception and pattern matching abilities to find certain signatures of OC/SD EBs in a raw light curve. The problem is, a lot of this pattern matching happens in our subconscious mind. We can't reverse engineer what our brains can do so well. (cf. for example the art of chicken sexing (!) which relies on this subconscious pattern matching through perceptual cues which can be taught)
Most of these light curves look like a long band of fuzzy noise, but look closer and there is some structure there. And since I'm organized in a left-brained way due to my past lives in science and engineering, I took the phased curve of KID 8654079 and tried to boil it down. Here's what I found:
You can see that there is a clustering of data points around the midpoint to slightly above the 1/3 level where it then fades off quite abruptly. The lower third of the light curve is divided into a slightly less thick density band and the last 20% of the light curve shows a more gradual fade out. The histogram essentially takes the various parts of the phased curve and organizes it by the clustering of data points. Since the shape of an OC remains quite consistent, it's possible to use this then as a guide to pick out light curves like this.
Compare this to the light curve of a DSCT - it's generally a noisy fuzz that's not differentiated. That's because the waveform is usually a sinusoidal curve that's itself quite spread out.
Just added the 19th OC SD candidate to this collection http://talk.planethunters.org/collections/CPHS0002p8Just added the 19th OC SD candidate to this collection http://talk.planethunters.org/collections/CPHS0002p8
This paper doesn't deal with with G type SD or OC stars directly, but it talks about a new algorithm for finding short period non-eclipsing binaries: [Seven new binaries discovered in the Kepler light curves through the BEER method](http://arxiv.org/abs/1110.2133 "") > We present seven newly discovered non-eclipsing short-period binary systems with low-mass companions, identified by the recently introduced BEER algorithm, applied to the publicly available 138-day photometric light curves obtained by the Kepler mission. The detection is based on the beaming effect (sometimes called Doppler boosting), which increases (decreases) the brightness of any light source approaching (receding from) the observer, enabling a prediction of the stellar Doppler radial-velocity modulation from its precise photometry. The BEER algorithm identifies the BEaming periodic modulation, with a combination of the well known Ellipsoidal and Reflection/heating periodic effects, induced by short-period companions. As far as I could tell none of the seven have Talk entries yet; four of the three I would guess ppl are classifying as Pulsating.  KID 10848064 folded _*Edit- adding additional KIDs referenced to BEER detections_ 8016222 9512641 7254760 5263749 4577324 6370196This paper doesn't deal with with G type SD or OC stars directly, but it talks about a new algorithm for finding short period non-eclipsing binaries:
Seven new binaries discovered in the Kepler light curves through the BEER method
As far as I could tell none of the seven have Talk entries yet; four of the three I would guess ppl are classifying as Pulsating.
KID 10848064 folded
*Edit- adding additional KIDs referenced to BEER detections
8016222
9512641
7254760
5263749
4577324
6370196
Wonderful paper NHB! I asked Meg of the science team if they would review this collection. My hunch is that quite a few of the collected LCs here are **eclipsing and/or non eclipsing binary stars** by their periodogram and analysis by Kianjin. Some are more obvious than others when viewing the light curve. Ggccg and Tucanico flagged some of them. None of these stars are of the seven listed in the paper. I do think we are on to something very interesting.Wonderful paper NHB! I asked Meg of the science team if they would review this collection. My hunch is that quite a few of the collected LCs here are eclipsing and/or non eclipsing binary stars by their periodogram and analysis by Kianjin. Some are more obvious than others when viewing the light curve. Ggccg and Tucanico flagged some of them.
None of these stars are of the seven listed in the paper. I do think we are on to something very interesting.
I went ahead and made periodograms for the seven shallow **non eclipsing binary stars** from the S. Faigler et al paper. One thing that seems clear is that all the periods are greater than one in Lomb Scargle mode. When you covert to Plavchan you can see sharp curves with alternating max and min in the majority of the graphs.  Plavchan   Plavchan   Plavchan   Plavchan   Plavchan   Plavchan   Plavchan  In the article I posted, you can see in the periodograms that Kianjin analyzed that they are very fast periods, much shorter than one. This, in itself, is very interesting and unique.I went ahead and made periodograms for the seven shallow non eclipsing binary stars from the S. Faigler et al paper. One thing that seems clear is that all the periods are greater than one in Lomb Scargle mode. When you covert to Plavchan you can see sharp curves with alternating max and min in the majority of the graphs.
Plavchan
Plavchan
Plavchan
Plavchan
Plavchan
Plavchan
Plavchan
In the article I posted, you can see in the periodograms that Kianjin analyzed that they are very fast periods, much shorter than one. This, in itself, is very interesting and unique.