CzeV627 Delta Scuti Variable Star
Posted: Wed Jun 17, 2020 10:23 pm
Last night was an all-nighter. I started with WASP-3 b, an Exoplanet in Lyra and ended with CzeV627. a Delta Scuti variable star in Cygnus. Here are the results of the latter.
CzeV627 was discovered by Czech astronomer Bohuslav Hladik in 2014. (In the attached screenshot I included what I believe to be his discovery data.) It's brightness varies by a mere 0.05 magnitudes every 2 hours. It is of type Delta Scuti (DSCT). I am not an astrophysicist but I believe that pulsations are due to the cyclical build up of opaque plasma in the chromosphere followed by ejection. DSCT stars are short period. Many of them you can capture in one session. Some have relatively large amplitude changes like YZ Boo but a great many, like CzeV627, are virtually indistinguishable from a regular, constant-brightness star. Furthermore, low-amplitude DSCT stars are susceptible to changing amplitude and sometimes will go dormant. Catalogs list CzeV627 with a depth of 0.05 magnitudes but I have seen light curves where it is significantly lower followed by an increase. Last night's data indicates that the depth is currently 0.03 magnitudes. By the way, 0.03 magnitudes is just about the upper limit on the depth of Exoplanet transits.
In the attachment the plot of magnitude versus time is upward sloping since my data was not "de-trended" for changes in Airmass. Through my work with astrophotography using a mono camera and LRGB filters, I know that the atmosphere attenuates light at different wavelengths. Red light more readily passes through the atmosphere but blue light suffers most as the distance increases from the zenith. So what happens is that stars suffer from a color shift. Every star shifts red as they grow closer to the horizon. Since hot blue stars have a lot of energy in blue wavelengths they appear less bright in a monochrome camera. Red stars don't suffer that much. In the attachment there is a screenshot from AstroImageJ showing three stars with "apertures" placed on them. The green aperture is CzeV627. The two red apertures are constant brightness "comparison stars". Each of these three stars have a different color index. Ideally I would want to select comparison stars with the exact same color index as the variable star but this is almost never possible. AstroImageJ can perform "de-trending" for Airmass but I did not employ it. Up until now I have not seen this effect since all previous variable stars I have measured have much larger magnitude changes.
You may notice that the stars are severely out of focus. That is on purpose. The exposure is 150 seconds. At perfect focus the stars would have been severely saturated. To mitigate saturation I de-focused. The brightest pixel in the variable star is now just below saturation. Thanks to Robin, I learned that this is a very effective technique for increasing accuracy. Notice the third column in the spreadsheet labeled "Source_AMag_Err_T1". Values are around 0.0024 magnitudes. That's incredible compared to other photometry I've done at perfect focus where the error was 0.05 magnitudes. That is a 20x increase in accuracy by doing two things: (1) de-focus, (2) increase exposure to bring the star just below saturation. Of course, it comes at the price of long exposures so there is no free lunch!
AFTERTHOUGHT:
After reviewing the data from my first exoplanet transit HAT-P-5 b seen here: https://u235-varstar.now.sh/light-curve/hat-p-5-b I now see that no such upward trend in the data exists. What is different? Most everything is similar except for: de-focusing and exposure. Frankly I don't believe that de-focusing is the culprit. That leaves exposure. For HAT-P-5 I decided upon a non-aggressive exposure that put the maximum pixel value at around 30,000 ADU. From tests that I ran a year earlier I knew that the sensor becomes non-linear at about 55,000 ADU. Last night I threw caution to the wind by flirting with that level. And of course one of the hazards of imaging east of the meridian is that exposure determination is made through the thickest airmass. This meant that as CzeV627 approached the meridian the star became brighter which pushed it farther into the non-linear zone. By the end of the session I could see that a single pixel had reached saturation at 65,535 ADU. No worries, I thought, there are plenty of pixels within the aperture that are below saturation. So that is my theory. Next clear night I will lower the exposure to test my hypothesis.
EPILOGUE:
If as I suspect that the culprit is overexposure then this is a lesson to heed for those in the audience who are Lunar and Solar imagers. We all agree that saturation is to be avoided but one should also consider the deleterious effects of running your highlights too hot. Beginning at around 55,000 ADU you will lose detail. Your highlights will become flat and uninteresting. Don't get me wrong. It's OK if you want to show detail in shadow and the mid-range but know that if you don't crop out highlights then it will distract from your composition.
Brian
CzeV627 was discovered by Czech astronomer Bohuslav Hladik in 2014. (In the attached screenshot I included what I believe to be his discovery data.) It's brightness varies by a mere 0.05 magnitudes every 2 hours. It is of type Delta Scuti (DSCT). I am not an astrophysicist but I believe that pulsations are due to the cyclical build up of opaque plasma in the chromosphere followed by ejection. DSCT stars are short period. Many of them you can capture in one session. Some have relatively large amplitude changes like YZ Boo but a great many, like CzeV627, are virtually indistinguishable from a regular, constant-brightness star. Furthermore, low-amplitude DSCT stars are susceptible to changing amplitude and sometimes will go dormant. Catalogs list CzeV627 with a depth of 0.05 magnitudes but I have seen light curves where it is significantly lower followed by an increase. Last night's data indicates that the depth is currently 0.03 magnitudes. By the way, 0.03 magnitudes is just about the upper limit on the depth of Exoplanet transits.
In the attachment the plot of magnitude versus time is upward sloping since my data was not "de-trended" for changes in Airmass. Through my work with astrophotography using a mono camera and LRGB filters, I know that the atmosphere attenuates light at different wavelengths. Red light more readily passes through the atmosphere but blue light suffers most as the distance increases from the zenith. So what happens is that stars suffer from a color shift. Every star shifts red as they grow closer to the horizon. Since hot blue stars have a lot of energy in blue wavelengths they appear less bright in a monochrome camera. Red stars don't suffer that much. In the attachment there is a screenshot from AstroImageJ showing three stars with "apertures" placed on them. The green aperture is CzeV627. The two red apertures are constant brightness "comparison stars". Each of these three stars have a different color index. Ideally I would want to select comparison stars with the exact same color index as the variable star but this is almost never possible. AstroImageJ can perform "de-trending" for Airmass but I did not employ it. Up until now I have not seen this effect since all previous variable stars I have measured have much larger magnitude changes.
You may notice that the stars are severely out of focus. That is on purpose. The exposure is 150 seconds. At perfect focus the stars would have been severely saturated. To mitigate saturation I de-focused. The brightest pixel in the variable star is now just below saturation. Thanks to Robin, I learned that this is a very effective technique for increasing accuracy. Notice the third column in the spreadsheet labeled "Source_AMag_Err_T1". Values are around 0.0024 magnitudes. That's incredible compared to other photometry I've done at perfect focus where the error was 0.05 magnitudes. That is a 20x increase in accuracy by doing two things: (1) de-focus, (2) increase exposure to bring the star just below saturation. Of course, it comes at the price of long exposures so there is no free lunch!
AFTERTHOUGHT:
After reviewing the data from my first exoplanet transit HAT-P-5 b seen here: https://u235-varstar.now.sh/light-curve/hat-p-5-b I now see that no such upward trend in the data exists. What is different? Most everything is similar except for: de-focusing and exposure. Frankly I don't believe that de-focusing is the culprit. That leaves exposure. For HAT-P-5 I decided upon a non-aggressive exposure that put the maximum pixel value at around 30,000 ADU. From tests that I ran a year earlier I knew that the sensor becomes non-linear at about 55,000 ADU. Last night I threw caution to the wind by flirting with that level. And of course one of the hazards of imaging east of the meridian is that exposure determination is made through the thickest airmass. This meant that as CzeV627 approached the meridian the star became brighter which pushed it farther into the non-linear zone. By the end of the session I could see that a single pixel had reached saturation at 65,535 ADU. No worries, I thought, there are plenty of pixels within the aperture that are below saturation. So that is my theory. Next clear night I will lower the exposure to test my hypothesis.
EPILOGUE:
If as I suspect that the culprit is overexposure then this is a lesson to heed for those in the audience who are Lunar and Solar imagers. We all agree that saturation is to be avoided but one should also consider the deleterious effects of running your highlights too hot. Beginning at around 55,000 ADU you will lose detail. Your highlights will become flat and uninteresting. Don't get me wrong. It's OK if you want to show detail in shadow and the mid-range but know that if you don't crop out highlights then it will distract from your composition.
Brian