Exoplanets

[Xilman]

Anyone else here interested in starting to observe exoplanets, or have already begun?

The field is no longer the exclusive preserve of professionals; amateurs can make observation of real scientific value. There are two principal means, each of which require measuring the brightness changes of their host stars. Neither require large telescopes (anything from 15-20cm upwards is fine) but they do require cameras and careful data analysis. Most, if not all, VS observers are already well versed in careful data analysis.

The transit method uses measurements of the dip in brightness of the host star as a planet passes in front of it. The light curve gives information about the size of the exoplanet and the size of its orbit.

The microlensing method uses measurements of the increase in brightness as a nearer massive body gravitationally lenses the light of a background star as it passes by. If the background star also has at least one planet the light from them will also be lensed and can be detected in the light curve. The light curve gives information about the mass of the exoplanet.

I have made a start with both of these methods. I can give newbie advice (remember I am a newbie in this respect) and I can give pointers to much more information about  the subject and its techniques, including some written by and for pro-am collaboration. Such information includes predicted transit dates, times, duration and depth, and stars which are likely being lensed.

[Guest]

Sorry I missed this post but yes I'm interested and just starting to find my way around. I've been doing variable star photometry for about a year but for the coming season I'll be looking at exoplanets and asteroids. I see that AAVSO had a feature on exoplanets just recently. 

I'm using a SW200PDS with a canon DSLR.  The AAVSO guide to exoplanets seems to discourage DSLR's for exoplanets but I don't see why you shouldn't use them.  There are a couple of points I'm not sure about.

The recommendation is that you should do all the measurements in the same waveband, V or TG for example. But it seemed to me that was a waste of light and I couldn't understand why you can't add the TB+TG1+TG2+TR counts since we are simply looking for a change in the total flux from the star.

The other thing I was interested in is that some of the light curves I've seen plot magnitude against time. But I thought it would be better to plot flux or total intensity against time.  Magnitude is a logarithmic value so I figured it would make the dip in flux smaller and harder to see.  We are mainly interested in the duration of the transit and the time between transits. Also the formula I have is that the ratio of the dip in flux and total flux is equal to the ratio of planet radius and star radius squared (basically ratio of areas). So we are not really interested in magnitude but flux which is proportional to intensity (I think).

At present I'm trying to improve my set up by adding guiding to my scope.

Cheers

Steve

 

[Xilman]

1 hour ago, woodblock said:

Sorry I missed this post but yes I'm interested and just starting to find my way around. I've been doing variable star photometry for about a year but for the coming season I'll be looking at exoplanets and asteroids. I see that AAVSO had a feature on exoplanets just recently. 

I'm using a SW200PDS with a canon DSLR.  The AAVSO guide to exoplanets seems to discourage DSLR's for exoplanets but I don't see why you shouldn't use them.  There are a couple of points I'm not sure about.

The recommendation is that you should do all the measurements in the same waveband, V or TG for example. But it seemed to me that was a waste of light and I couldn't understand why you can't add the TB+TG1+TG2+TR counts since we are simply looking for a change in the total flux from the star.

The other thing I was interested in is that some of the light curves I've seen plot magnitude against time. But I thought it would be better to plot flux or total intensity against time.  Magnitude is a logarithmic value so I figured it would make the dip in flux smaller and harder to see.  We are mainly interested in the duration of the transit and the time between transits. Also the formula I have is that the ratio of the dip in flux and total flux is equal to the ratio of planet radius and star radius squared (basically ratio of areas). So we are not really interested in magnitude but flux which is proportional to intensity (I think).

At present I'm trying to improve my set up by adding guiding to my scope.

Cheers

Steve

 

Only a quick response to two of your points.

The one about filters I raised with the professionals in Exoclock. The trade-off is signal to noise. The sky brightness is markedly lower in the red and near IR than it is in the blue and green. Remember that an illuminated sky is blue.  It is still blue at night because it is still illuminated, especially if the moon is above the horizon , just a very much darker blue than in the daytime. (Incidentally, the moonlit sky still looks blue to me tonight as I have unusually sensitive eyes after suffering measles badly as a kid, but I digress). The response I received is that a Cousins R or Sloan r' filter is better for reproducibility with other's observations but if you get markedly better cadence unfiltered at the same SNR, go for it. It is easy enough to test this for yourself. Ingress and egress timings vary slightly with wavelength but mid-eclipse timings are essentially unaltered.

As for flux versus magnitude: you are quite right, the "correct" thing to plot is flux. However, it doesn't make any significant difference for dips below 0.1 mag or so.  A fractional 0.02 fall in flux is extremely close to a magnitude dip of 0.02 magnitudes. if you did calculus at school, you will remember that the Taylor series for log(1+x) is x - (x^2)/2 + (x^3)/3 - ... and if x is a very small value  (-0.02 in my example)  all but the first term is negligible and we have log(1+x) = x to a very good approximation.

More later.

Paul

Edited July 22, 2021 by Xilman
Clarify cadence vs SNR

[Xilman]

2 hours ago, woodblock said:

Sorry I missed this post but yes I'm interested and just starting to find my way around. I've been doing variable star photometry for about a year but for the coming season I'll be looking at exoplanets and asteroids. I see that AAVSO had a feature on exoplanets just recently. 

I'm using a SW200PDS with a canon DSLR.  The AAVSO guide to exoplanets seems to discourage DSLR's for exoplanets but I don't see why you shouldn't use them.  There are a couple of points I'm not sure about.

I get the impression that AAVSO are often rather sniffy about using DSLR for variable star observing. Observing an exoplanet transit is just photometry on a VS with very small range.  I  can see why they take this view. There are several reasons why DSLRs are inferior to a CCD camera. For a start, the Bayer mask (and this applies to OSC CMOS cameras too) means that the R and B detectors only receive 1/4 of the photons each whereas every detector in a mono-camera receives its full complement of signal. Further, consumer DSLRs filter out the IR response, again at loss of sensitivity. Terrestrial photographers almost always have a lot of light available and are completely obvious to dark noise. The real killer, though, is that dark noise is very temperature sensitive and DSLR cameras have to be used at close to ambient temperature. A cooled astro camera locked at, say, -20C has very little thermal noise and that is relatively easily measured. A DSLR being used for several hours (a typical transit duration) will vary in temperature with the atmosphere and will gently warm up from internal power dissipation. All these effects make estimation of noise, and hence SNR, harder to measure. When you are looking for a transit depth of 0.015 magnitudes or less (alternatively a flux drop to 0.985 that outside transit) you have to be quite careful not to lose the signal in the noise.

Having said that, DSLRs can result in very good science. You just need to accept that the more difficult targets (fainter stars, smaller transit depths) should be left to the cooled CCD and CMOS astro cameras.  That still leaves far more stars to observe than there are observers observing them.

The great advantage of DSLR cameras is that they are a mass-market device. Economies of scale mean that they don't cost as much as dedicated astronomical cameras and, because they are so much more versatile, they are much more likely to be found in the possession of people who wish to take up photometry and astrometry, not to mention those who enjoy taking pretty pictures. If you've got it, use it!

[Guest]

I must have this wrong. I looked at a star from one of my variable star photometry images. The star had a magnitude of 12.38 and an intensity of 15801. The star image was not saturated. This was a 30s exposure at iso400. I figured a fall in magnitude of 0.02 to give a magnitude of 12.40 (increase in magnitude due to a fall in brightness). So according to my calculations that would correspond to an intensity of 15512 doing the inverse logarithm. A fall in intensity of 15081-15512=289 which is 1.89% fall. But the fall in magnitude of 0.02 is only 0.16% of 12.38. My idea was that the plot of magnitude would appear to show a smaller dip as a percentage than a plot of intensity.

Can I ask where you post your results?

Cheers

Steve

 

[Xilman]

1.89% is very close to 0.02 ...

[Guest]

Yes but if you plotted magnitude against time you'd see it fall from 12.38 down to 12.40 which is a tiny step down. You'd  hardly see it. If you plot intensity against time it would fall from 15512 to 15081 which although still small is much bigger than the fall in magnitude. If you plotted them as percentage say.

What I was thinking of is that say the orbital period is 5 days and the transit time is 3 hours and you were looking for a change in magnitude of 0.02 it would be hard to see. I know from my variable star photometry that there would be loads of noise in the measurements. The dip would be hidden in all the noise. But if you plotted intensity the change would be ten times bigger and you'd have a better chance of seeing it.

 

[Xilman]

Not true, unfortunately. By your measure the noise would also be ten times higher. Remember that the noise is also a fractional change in intensity.  Noise at a fraction of 0.001 of the intensity corresponds to a noise of 0.001 magnitudes by precisely the same argument as above.

 

You may have to convince yourself by observation and measurement.

[Guest]

Yes thanks I get it now but I will try it out just for the hell of it.

Do you have to use software to detect the transit?  I've seen a few charts which show the scatter of points with a line drawn through showing the transit and I think you'd be hard pressed to do it by eye.

 

[Xilman]

4 hours ago, woodblock said:

Yes thanks I get it now but I will try it out just for the hell of it.

Do you have to use software to detect the transit?  I've seen a few charts which show the scatter of points with a line drawn through showing the transit and I think you'd be hard pressed to do it by eye.

 

IMO software is essential to get useful results.

I have not yet used it myself but HOPS has a very good reputation for being easy to use and it feeds directly into an important research project.

Head over to https://www.exoclock.space/ for information about the Exoclock project and the HOPS software.

I am not using it myself because I am more interested in gravitational microlensing than in transits.

[Lariliss]

Thank you for posting an interest in the exoplanets observation.

They impose a real interest:

Paying attention to stability of star-planet systems (orbits, spins);

Data for possible planet information (not only mass, but temperature, so forth);

Moonmoon systems.

 

I would also agree on the flux, that generally it is more correct. But depending on observation, is it not more vulnerable to the data collection (though during the time it may have just insignificant fluctuations)? Magnitude looks more stable for observance, and at the same time gives more information on steady changes, opposed to snapshot of flux.

 

Actually, any additional data that might have been missed will be significant for new simulations of exoplanets investigation.

[Xilman]

45 minutes ago, Lariliss said:

I would also agree on the flux, that generally it is more correct. But depending on observation, is it not more vulnerable to the data collection (though during the time it may have just insignificant fluctuations)? Magnitude looks more stable for observance, and at the same time gives more information on steady changes, opposed to snapshot of flux.

Given that a magnitude is just a logarithm of a flux it doesn't matter which is recorded: both have exactly the same information content.

To be precise, a magnitude difference between two sources is the logarithm of the ratio of the fluxes coming from those sources. It makes no difference to the information content. In essentially all astronomical photometry the observer measures the target of interest and two or more comparison stars. What is recorded is the flux. The magnitude of the target is computed from the ratio of its flux to that of the comparisons which are assumed to have known magnitudes.