SkyCruncherA thin line beside a passage says how far its claim is proven, and colored numbers carry the same mark: measured numbers come from my own frames and link the Experiment they come from; no line means stated: general knowledge or another source, not tested by me; illustrative shows an idea with rough numbers; implemented is what our shipped tools do. See how claims are marked.
What you are measuring
A transit is the moment a planet passes between us and its star. The planet blocks a small part of the star's disk, so the star's brightness drops by roughly the planet's area divided by the star's: about 1 to 3 percent for a hot Jupiter, a gas giant on an orbit of a few days. The depth of the dip tells you the planet's size, and the time of its middle tells you where the planet is in its orbit.
A light curve is the star's brightness plotted against time. Catching a transit means recording a light curve clean enough that the dip stands out from the noise, with enough time before and after it to know where the "normal" brightness sits. That stretch before and after is the baseline.
1. Pick a planet your telescope can see
Two things decide whether a transit is catchable: how deep the dip is, and how bright the star is. A brighter star gives more light per frame, so its brightness is measured more precisely. A fainter star needs a deeper dip, or a bigger lens.
For a bright star, the limit stops being light and becomes the air: turbulence makes a star twinkle, and through a lens as small as a Seestar's that twinkling adds noise no processing removes. A bigger aperture averages it down. The Seestar S30 and S30 Pro have 30 mm lenses and the S50 has a 50 mm lens.
The SkyCruncher planner does this sum for you, star by star. It lists a transit only when the dip is at least three times its expected scatter, the star stays above 30° from 30 minutes before the transit to 30 minutes after it and is at least 40° up at mid-transit, and the sky is dark (the Sun at least 12° down) for all of that time. Each transit card shows the dip, the expected scatter and how many times bigger the dip is. (implemented)
For a Seestar, the planner works out those numbers for the telescope you pick; for models other than the S30 Pro it scales them by aperture, and it says so on the page. (implemented)
2. Plan the night
Open the planner, choose your telescope and location, and look for a transit card. For each transit the planner gives a recording window from 45 minutes before the dip starts to 45 minutes after it ends, so you have baseline on both sides. (implemented)
Treat those 45 minutes as a minimum. The fit has to tell the dip apart from a slow drift in brightness as the star rises or sinks, and the "before" stretch is what pins that drift down. Both of my nights started with less than 45 minutes of "before", and that cost me (see what I caught). (measured)
Check the prediction itself, too. A transit time comes from a timetable of past crossings, an ephemeris, and older timetables can drift by many minutes. Take the time from a recently updated source. The planner takes its transit times from ExoClock's ephemerides where they exist and from the NASA Exoplanet Archive otherwise, and shows how far each predicted time may be off by that date. (implemented)
3. Set up
-
Use equatorial mode
On an alt-az mount the field slowly turns as the telescope follows the sky, so the stars wander across the sensor. In equatorial mode, on a wedge, the field stays put, the stars sit on the same pixels and the comparison between stars stays clean. For a brightness measurement longer than an hour, the planner's guide recommends equatorial mode for this reason. (implemented) More in Alt-az or EQ mode.
-
Keep one exposure all night
Pick one exposure length, gain and filter and don't change them during the run, so every frame is measured the same way. Keep the framing fixed: no mosaic and no recentering by hand. Don't let the target saturate: if it is very bright, use shorter frames or defocus slightly.
-
Save every frame
Turn on saving each frame in the Seestar app. The individual frames (subs) are the measurement; the stacked picture the app shows is not. More in what your Seestar saves.
-
Get the time right
The mid-transit time is only as good as the clock that stamped the frames. A clock a few seconds off doesn't matter against an uncertainty of minutes; a clock minutes off does. Check the telescope's time against a reliable source before you start.
-
Start early and record without stopping
Start recording before the planner's window opens and leave it running until well after the dip ends. Note the time of any cloud or interruption.
4. Reduce the frames
Comparison stars. The air changes all night: haze drifts through and the star fades as it sinks into thicker air, by much more than the transit's depth. So you never use the target's brightness on its own. You divide it by the combined brightness of other stars in the same frames, the comparison stars. Whatever the sky does to them, it does to the target too, and the division cancels it. Good comparison stars are steady, not known variables, of similar brightness, uncrowded and away from the frame's edges. Many of them, chosen by fixed rules, beat a few picked by eye.
Check stars. Measure a few stars of similar brightness exactly like the target but leave them out of the comparison set. They should show no dip. If one does, the problem is in the reduction or the instrument, not the planet.
Detrending. Even after dividing by comparison stars, a slow drift is usually left, often following the airmass (the thickness of air the light passes through). The usual answer is to fit that baseline trend and the transit model together, rather than removing the trend first. With a short baseline on either side, the trend and the depth trade off against each other, which is another reason to start early.
I am not making any measured claims about detrending methods here yet; those wait until the work behind them is published.
5. What I caught
Two transits, each one night with one telescope, so both Experiments call their results exploratory until a second night repeats them. (measured)
| Planet | Telescope and frames | Depth | Mid-transit |
|---|---|---|---|
| HD 189733 b | Seestar S30 Pro (30 mm), 544 frames of 10 s, equatorial mode, light-pollution filter | 2.50 ± 0.34% | 1.0 ± 3.6 min before the prediction |
| WASP-12 b | Seestar S50 (50 mm), 597 of 647 frames of 20 s used, equatorial mode, IR-cut filter | 1.63 ± 0.20% | 0.4 ± 4.2 min after the newest prediction |
Depths and times (measured), from the two Experiment pages linked in the table.
HD 189733 b, 2 October 2026
The star dimmed by 2.50% (± 0.34%), and the middle of the transit came 1.0 minute before the prediction (± 3.6 minutes). In that headline fit I measured only the depth and the mid-time; the transit's length and the planet's path across the star came from the published orbit. It uses all three pixel colors with each color cleaned. The reduction I froze before looking, green pixels only, gives 2.83 ± 0.40% and 1.6 ± 3.8 minutes late. This night was used to choose the reduction steps, so it can't confirm them. (measured) Large telescopes measured a depth of about 2.45% for this planet (published values, third-party).
The comparison stars mattered most. Going from 7 hand-picked stars to 98 chosen by fixed rules cut the scatter of the 10-minute averages 2.3 times and moved the mid-transit from 14 minutes late to on time. None of the cleanup steps I tried came close to that. (measured)
I started recording 49 minutes after plan, which left about 20 minutes of steady data before the transit began. The depth didn't mind, but the timing leans on that baseline, which is part of why its error bar is so wide. (measured) Source: Catching HD 189733 b.
WASP-12 b, 10 to 11 October 2026
The star dimmed by 1.63% (± 0.20%), a dip 8.3 times its own uncertainty, with mid-transit 0.4 minutes after the newest published prediction (± 4.2 minutes). Three check stars of nearly the same brightness showed no dip, each within about one of its own uncertainties of zero. (measured)
I started only about 15 minutes before the transit began, and with so little "before", the fit can't fully tell the transit from a slow drift. The fit preferred a baseline that follows the airmass, which gives the headline. A straight-line baseline fits almost as well and gives a shallower 1.38 ± 0.19%, with the same mid-time (+1.2 minutes); that spread is counted in the headline uncertainty. (measured) Published depths for this planet run from about 1.25 to 1.54%, depending on the paper and the color of light (third-party).
Two more lessons from that night. In equatorial mode the Seestar re-centered every few minutes, so the stars jumped up to about 100 pixels; software that assumes the stars stay put loses them. (measured) And I didn't check the telescope's clock against an independent source, so the timing rests on an assumption, not a measurement. Source: Catching WASP-12 b.
What the two nights add up to. A 30 mm and a 50 mm Seestar each caught a known hot Jupiter at about the expected depth and time. One night each, so neither result is confirmed yet. For WASP-12 b, whose orbit is slowly shrinking, a timing uncertainty of about 4 minutes is too coarse to measure that shrinking on its own: the night is one more timing that agrees with what larger telescopes have measured. (measured)
6. Send it on
ExoClock is a project that collects transit timings from small and amateur telescopes to keep exoplanet predictions up to date, in support of the European Space Agency's Ariel mission. It accepts light curves from observers and lists a recommended minimum telescope size for each planet. For WASP-12 b, ExoClock lists a 5-inch minimum, larger than my 50 mm lens (ExoClock's figure, third-party). My dip still came out at 8.3 times its uncertainty. (measured)
Before you submit, check that the dip is clear, that your check stars are flat, and that you know how your frame times were set. The planner's transit guide also names the AAVSO Exoplanet Database as a place to send a clean light curve. (implemented)
You can also share a night of subs through SkyCruncher's Upload page. The files stay private, but the target, its position on the sky, the number of frames and the hours of exposure appear on the public sky map; Privacy says what happens to them. (implemented)
Where to go next
- The planner: transits for your telescope and your sky this week.
- Catching HD 189733 b and Catching WASP-12 b: the two nights in full, with the data.
- Starlight, Measured: how starlight becomes a measurement, step by step.
- All Notes in Learn.