Beginner seeking advice on selecting exoplanets for detection (transit)

[Craig Towell]

Hello,

I have consumed quite a few online guides and tutorials about beginner exoplanet transit detection, but the one thing I’m missing is what mag star and transit depths I should aim for with my equipment.

Can anyone give me a realistic indication of what I should be looking for in the chart here ar.astro.cz/en/Exoplanets/TransitsPredictions with regards to star magnitude and transit depth?

For info I’m using a 102mm refractor and OSC CMOS sensor (for now).

Will HAT-P-20 b be a realisitc target for aperture?? (mag 11.3 and 2% depth)

Or should I be looking deeper transits and/or brighter stars?

Thanks in advance for any help received

Craig Towell

[Dr Paul Leyland]

I have no personal experience with a 102mm aperture but my 400mm can do useful work on a 1% depth at 13th-14th magnitude. Simple scaling says that you have 1/16 the collecting area, which corresponds to three magnitudes in brightness.

Accordingly, I expect that you may have a chance with HAT-P-20 but it’s not certain. Certainly well worth giving it a try.

Regardless, it’s easy enough to test the capabilities of your equipment. Choose a star of a particular magnitude and see how long an exposure is require to reach a SNR of at least 500 and preferably 1000. If it is a couple of minutes or less, and preferably less, you can measure a transit of depth 2% or better. Repeat with fainter stars until you determine your limiting magnitude.

Good luck!

Paul

[Craig Towell]

Thanks for the response Paul, that is very helpful.

I also have a 305mm/1500mm Newtonian I could use, however the FoV would be tiny (0.28 deg x 0.16 deg), not sure if that would be enough to get any comparison stars in shot.

[Dr Paul Leyland]

Craig: that FOV is comparable to my 400mm, which is 16.45 * 13.18 arcmin, or 0.27 x 0.22 degrees. I have not yet had any problem finding a suitable comparison star but, to be fair, stars do become much more common at fainter magnitudes.

Once more: suck it and see is my recommended approach. Remember that you can do all this sort of experimentation at your convenience – you do not need to wait until a transit is imminent. If you do so for all the stars which you are likely to observe over the next few weeks you can choose which telescope to use as appropriate.

[Craig Towell]

Thanks Paul that is very helpful. I had a short clear spell when I got home from work and managed to capture a series of images of varying durations of a star field.

I’m just looking at the images now and identifying the stars in the image and their magnitudes.

What do you use to measure the SNR of a particular star? I have pixinsight and I can see the ADU values of the centroid and the background, is it simply a case of taking the ratio of those two values?

Thanks again
Craig

[Dr Mark Salisbury]

Hi Craig. I would add that it may be useful taking a look at ETD or Exoclock to give you an idea what observers with similar equipment are achieving. On the ETD planner you can also filter possible targets by mag and transit depth which may help with selecting targets. On paper HAT-P-20b looks like a good target but it’s worth noting it has a mag 12.5 star just 7arcsec away which you’d need to seperate in your photometry.

Mark

[Craig Towell]

Thanks Mark, I hadn’t considered nearby stars to the target star so that is a very useful tip for me.

I’m having a little trouble with establishing what the SNR is on any given star. Tring to do it in Pixinsight, and to be honest as there is so much conflicting information online I’ve no real idea what I’m doing!

I took a calibrated 2 min sub, debayered and extracted the green channel (OSC camera).

The I created a preview box right around a star, and another on a patch of background sky.

Then I used the e statistics function to look at the mean ADU value of the star (3420) then divided that by the stddev of the background (57) to get 60. So is that a SNR of 60? Pretty flow for a star close to saturation!

I have attached a pic of the preview boxes I used, coupled with a .Fit file of the green channel I was inspecting.

Any help greatly appreciated!

Edit: seems like the .fit is too large

• This reply was modified 1 month ago by Craig Towell.

Attachments:

1. 

   IMG_5192.png
   

[Craig Towell]

Right so I was woken at 3am by the youngest sprog and couldn’t get back to sleep, so I did a load of internet searching and reading around SNR, AstroImageJ and aperture photometry (totally normal behaviour, right?).

Stumbled across this useful looking calculator…
https://mirametrics.com/sn_calculator_mvn.php

Plugging in the pertinent data, or estimates of, and it thinks using a 102mm refractor operating with 2 minute exposures, and assuming my minimum acceptable SNR to be 500, then it looks like I should be aiming for stars of at least mag 12.5 with a 2% transit depth.

Edit: Plugging in numbers for my 300mm/1500mm Newtonian says it should be good down to mag 14 ish with a 2% depth.

• This reply was modified 4 weeks, 1 day ago by Craig Towell.
• This reply was modified 4 weeks, 1 day ago by Craig Towell.

Attachments:

1. 

   IMG_5198.png
   

[Dr Paul Leyland]

Craig Towell wrote:

<p class=”wp-dark-mode-bg-image”>Right so I was woken at 3am by the youngest sprog and couldn’t get back to sleep, so I did a load of internet searching and reading around SNR, AstroImageJ and aperture photometry (totally normal behaviour, right?).

<p class=”wp-dark-mode-bg-image”>Stumbled across this useful looking calculator…<br class=”wp-dark-mode-bg-image”>
https://mirametrics.com/sn_calculator_mvn.php

<p class=”wp-dark-mode-bg-image”>Plugging in the pertinent data, or estimates of, and it thinks using a 102mm refractor operating with 2 minute exposures, and assuming my minimum acceptable SNR to be 500, then it looks like I should be aiming for stars of at least mag 12.5 with a 2% transit depth.

<p class=”wp-dark-mode-bg-image”>Edit: Plugging in numbers for my 300mm/1500mm Newtonian says it should be good down to mag 14 ish with a 2% depth.

Perfectly normal behaviour, IME.

Your numbers sound about right to me and consistent, though not identical, with what I posted earlier.

Your noise will depend greatly on your sky brightness, of course. Don’t expect as good results from a brilliantly moonlit Bortle-7 location as a new moon B4.

I would recommend installing a proper photometry package (I use and recommend APT, Aperture Photometry Tool) but there are several others available. Any of them will give a good measurement of the SNR of any given star in any given image.

Incidentally, and this doesn’t apply to exoplanetary transit work, stacking many unsaturated images in addition mode will increase the SNR by a factor of the square root of the number of images. By adding the images you preserve the photometric accuracy and let the counts per pixel exceed the detector’s saturation limit.

• This reply was modified 2 weeks, 2 days ago by Andy Wilson.

[Dr Paul Leyland]

Craig Towell wrote:

<p class=”wp-dark-mode-bg-image”>I took a calibrated 2 min sub, debayered and extracted the green channel (OSC camera).

Something worth remembering: if all you are interested in is the change of brightness as time passes, it matters relatively little what filter(s) you are using.

If you extract all four Bayer layers (1xR, 2xG and 1xB) and then add them together you will have more signal and relatively less detector noise in both star and sky. If all channels have equal signal, which they won’t, the SNR will be increased by a factor of sqrt(2) or approximately 1.4 which is well worth having.

[Craig Towell]

I tried to install and use APT but just could not get it to work. I did manage to get AstroimageJ up and running which has allowed me to measure SNR on my images, so I did a little bit of testing on a previously captured image.

With my 102 frac, colour camera and Lum filter, I took a 2 min exposure of a Starfield near Polaris.

Looking at that image in AstroimageJ I can see stars at mag 12.10 have a peak value of 35,520 and an SNR of 550. Very usable I hope.

Edit: Interestingly the SNR calculator I linked to earlier predicted an SNR of 595 for that star and capture parameters. Not a million miles off.

• This reply was modified 3 weeks, 1 day ago by Craig Towell.

Attachments:

1. 

   IMG_5259.jpeg
   

[Dr Paul Leyland]

Good!

Sounds promising.

Note that a luminance filter removes the near-IR which, if present, gives an unusual colour balance. You are (or should be) not interested in what it looks like but only how many photons you have available. My advice is to run unfiltered in future and let the IR photons through so that they may contribute to the overall signal.

• This reply was modified 3 weeks ago by Dr Paul Leyland. Reason: Add final para

[Craig Towell]

Thanks Paul, that sounds like good advice. I am using a doublet refractor though, and whilst the colour correction is good, it’s not perfect. Would that make any difference to the decision to go with an IR/UV cut filter over a clear filter.

Just thinking the stars with a clear filter will probably be quite bloated. But I guess that won’t matter much for photometry?

[Dr Paul Leyland]

Ah, I forgot about chromatic aberration. 🙁 Too much time spent with reflectors. That said, my Dilworth despite being similar to a classical Cassegrain though with spherical mirrors, has a series of achromatic transfer lenses. I have never noticed any chromatic aberration when imaging.

Again, the best way is to suck it and see. If the unfiltered images are not too bloated (and they may well not appear larger at all), set the photometry aperture perhaps just a little bit larger to compensate and measure the same field with the same exposure on the same night (i.e. consecutive images) filtered and unfiltered so that the sky conditions are as similar as possible. Use the version which gives you the better SNR.

As always: SNR and cadence is what is important; aesthetics doesn’t get a look in.

[Craig Towell]

Well I think I just about managed it… some scudding clouds led to data gaps and cut short the post egress ‘tail’. The data is not pretty but it shows what its supposed to (just).

Interestingly I measured ingress a bit later than predicted. Will need to check those timings.

Thanks for all your help!

Attachments:

1. 

   HAT-P-32b-4th-Feb-2025.png
   

[Dr Paul Leyland]

Craig: that is pretty damn good!

To be honest, I thought you would have difficulty with a 100mm aperture, given that recommendations tend to start at 200mm, or 150mm at a pinch. You surprised me!

Your data looks entirely sufficient to make a scientifically valuable measurement. Don’t worry about the predicted time to be different from your measurement. That is the whole point of the exercise: to test the predictions. Please report it so others can learn from it.

I am impressed, and please keep measuring other transits with either or both of your telescopes.

Paul.

P.S. You should boast about this one in other astronomical fora such as SGL, CN and Bluesky.

[David Nicholson]

That is truly impressive, thanks for sharing!

The quality of equipment and knowledge that amateurs have available to them is incredible. Once I have got back into visual variable observing I intended to start pulling together some tracking (a mount) to do some DSLR observations. I then intend to follow this with a telescope (my dob being unsuitable) and a dedicated astro camera.

This thread has already been inspirational for one potential variable/exoplanet observer!

All the best,
Dave

PS I agree that you should boast about this on other fora!

• This reply was modified 2 weeks, 1 day ago by David Nicholson. Reason: Typo and added info

[Craig Towell]

Thanks for the kind words Paul and David. I must say it is intensely satisfying seeing the data points plot out for the first time. Addictive too. Not great for the wallet though as I’m already eyeing up larger scopes and a monochrome camera!

Craig

[Author]

Posts