I thought I would try something different, some scientific imaging, to see if I could wring enough precision out of collection of cheap telescope bits to record an exoplanet transit.
It took me 5 attempts over 3 months until everything co-operated (mostly).
I suspected that i'd made a timing error when the actual ingress time was out by 12 minutes compared to the exoplanet transit ephemeris predicted ingress.
However, I found results from the 2 meter telescope at Siding Springs Observatory from 2021 that was around 6 minutes early. Maybe TOI-3486.01s year is a bit shorter than believed.
TOI-3486.01 is a little smaller than Jupiter and orbits its star TIC-221861843.01 in around 2.22 days. TIC-221861843.01 is about 20% smaller than our Sun.
I used a C8 on a HEQ-5 and a Player-One Poseidon-C IMX571 based camera.
Processed with ASTAP and AstroImageJ.
What a world we live in, where an old guy in a backyard in Melbourne can detect a planet around another star.
So is that a detection threshold of 0.02 magnitudes you're working with there?
Also did you use the de-focus method, or did you buy one of those special diffuser filters?
Markus
Great questions, Markus.
The transit depth from the ephemeris is 15.05 ppt, which is around 16 milli-mag, approximately 0.016 mag. The star has a magnitude of 11.5. While NASA calculated the transit flux depth at 1.5%, my results indicate a closer to 2% drop in flux.
I did not use a diffuser. The C8/Starizona corrector/IMX571 combination leads to significant oversampling, resulting in a good spread of photons from the star across the pixels. I supplemented this with a slight defocus, extending to about 5.0 HFD. The imaging session lasted about 3.5 hours, during which I allowed the C8 focus to drift (as they tend to), and the session concluded at around 6.5 HFD. This did not pose any issues during processing.
Hi Chris. Awesome result. I tried my hand at asteroid occultations several times, but that was a bit too 'real-time' and I couldn't get fast enough video frame rates and still have a decent amount of signal. Exoplanet transits look to be a bit more of a relaxed pace with no need to video.
What exposure times did you use? And how often did you take images?
Hi Chris. Awesome result. I tried my hand at asteroid occultations several times, but that was a bit too 'real-time' and I couldn't get fast enough video frame rates and still have a decent amount of signal. Exoplanet transits look to be a bit more of a relaxed pace with no need to video.
What exposure times did you use? And how often did you take images?
Thanks Andrew.
It was a pretty relaxed pace, just like any other imaging I've done, with just a few different concepts and setup considerations.
For this transit I used 147 sec exposures at 200 gain, and imaged continuously for transit and an hour before and after, however, these values are determined by the length of the transit and the magnitude of the target star, so every target will be different.
The imaging time (1 hour before transit to 1 hour after) was 3.5 hours. I wanted to get around 30 images per hour to give me enought data points, so I aimed for somewhere around 120 sec. exposures. That would give me about 100 images in total. Thats a nice starting point for processing.
Another thing to consider is the camera gain. I aimed to get the brightest pixel in the target star image to be around 66% of the camera full well value. The target star brightness is going to change during the night, not just from the transit but as you track through different atmospheric thicknesses. 66% of max pixel value gives you some headroom because if any pixels saturate in the target star you're going to be losing data. You don't want that. However, if you see the max pixel value increasing too high during imaging you can try decreasing focus to spread the flux out and this will lower the individual pixel values of the target star. Anyway, I adjust the gain to get the initial pixel value to around 66%.
Don't forget to shoot flats and darks with these setting also, you need them too.
This is fantastic. Really excellent work and congrats on the capture!!
I’ve been really wondering about what you said: “ The target star brightness is going to change during the night, not just from the transit but as you track through different atmospheric thicknesses.” Would you need to calibrate this effect out by identical captures on other nights without the transit?
This is fantastic. Really excellent work and congrats on the capture!!
I’ve been really wondering about what you said: “ The target star brightness is going to change during the night, not just from the transit but as you track through different atmospheric thicknesses.” Would you need to calibrate this effect out by identical captures on other nights without the transit?
You know, when I posted this I wondered if some bright spark would pick up on that.
Yes, you do have to correct for variations in brightness caused by the atmosphere. In the images, along with the target star, you select multiple reference stars. Reference stars should be similar in brightness to the target star and not be a variable star. The software will use variations in the reference stars to correct the target star for variations in atmospheric conditions.
You know, when I posted this I wondered if some bright spark would pick up on that.
Yes, you do have to correct for variations in brightness caused by the atmosphere. In the images, along with the target star, you select multiple reference stars. Reference stars should be similar in brightness to the target star and not be a variable star. The software will use variations in the reference stars to correct the target star for variations in atmospheric conditions.
Chris
Ahh ok I see. Yes that makes a lot of sense I wasn’t aware of that capability. What software were you using for analysis? Just ASTAP or something more specialised?
Ahh ok I see. Yes that makes a lot of sense I wasn’t aware of that capability. What software were you using for analysis? Just ASTAP or something more specialised?
I used ASTAP just to bulk platesolve the images for alignment. The heavy lifting is done by AstroImageJ (AIJ)
Thanks for taking the time to provide all this really useful information. I have a target (TOI 2414.01) picked for tonight, so I'll be giving it a go.
I watched the video you linked. What is not clear from that is what format the light frames are in at the start. They don't seem to be raw images and look to have already been debayered before applying the calibration files. This is the opposite of convention, where you would calibrate first and then debayer the calibrated light frame.
Unless he has first done a green channel extraction on everything (lights, darks, flats etc.). I tried to work it out from his light filename but it doesn't give any clues as far as I can tell.
Thanks for taking the time to provide all this really useful information. I have a target (TOI 2414.01) picked for tonight, so I'll be giving it a go.
I watched the video you linked. What is not clear from that is what format the light frames are in at the start. They don't seem to be raw images and look to have already been debayered before applying the calibration files. This is the opposite of convention, where you would calibrate first and then debayer the calibrated light frame.
Unless he has first done a green channel extraction on everything (lights, darks, flats etc.). I tried to work it out from his light filename but it doesn't give any clues as far as I can tell.
The video assumes you're using mono images and never talks about processing colour subs. However, I only have a colour camera so I had to find a process to debayer and extract luminance after the calibration process.
This is what I finally ended up doing
1. Followed the calibration process in AIJ using my bayered subs.
2. Use pixinsight to debayer all the AIJ calibrated images.
3. Use pixinsight script->Batch Processing->Batch Channel Extraction Script.
select your files and in the Channels tab select "CIE Lab --> L" and
execute.
4. Use pixinsight script->Batch Processing->Batch Format Conversion.
to change the L subs you just created from .xisf back to .fits
5. Use ASTAP to batch platesolve again.
After all that you can continue in AIJ using the luminance files created, just as in Patriot Astros video.
Also I think you can save in step 3 as FITS and step 4 is not needed, maybe give that a try.
I have a bit more information on the star TIC-221861843. I was able to find it in the GAIA catalog as GAIA Source ID 5927041953543630592 and it is located 500 light years away.
Honestly, it still amazes me that amateurs have the technology available to confirm a planet around a star 500 ly or more away.
As for the 12 minutes discrepancy between predicted ingress time and actual ingress time I measured and the 6 minutes discrepancy with Siding Springs in 2021, my best guess is that the this exo-planet is not in a near circular orbit, as the ephemeris assumes, but in a elliptical orbit with the long axis precessing over time, similar to Mercury.
Anyway, thanks for all the comments I received and I hope that this will inspired others to take up the challenge.
