Start with the shadow crossing the disk
Imagine watching a distant star that the telescope cannot resolve into a large disk. Its total light stays almost steady. A planet moves between us and the star, hiding a small patch. The recorded light drops, reaches a minimum, then returns as the planet leaves. This passage is a transit.
For a uniformly bright star and a planet wholly inside its apparent disk, the lost fraction is the ratio of their areas. A planet one tenth as wide as its star hides one hundredth of the light. That is a useful first ruler. It gives a radius ratio; a planet in kilometres still needs the star’s radius in kilometres.
Actual discoveries add evidence beyond a ruler. The WASP-126 discovery used repeated dips, follow-up photometry and stellar velocity measurements to establish a planetary companion. The real TESS observations later in this sequence revisit its light curve. We begin with the geometry that makes those measurements interpretable.
Reading: Mandel & Agol 2002, Analytic Lightcurves, arXiv v1 · Maxted et al. 2016, Five WASP-South Planets, arXiv v1
The center and edge do not shine equally
The visible surface of a star is a glowing atmosphere. Near the disk center we see into deeper, hotter layers. Near the edge our slanted view reaches a cooler layer. The edge often looks dimmer. Astronomers call this limb darkening.
Now move the same planet from a central path to a path near the edge. It hides a different amount of light per unit area. The transit also becomes shorter because it crosses a shorter chord. Close enough to the limb, part of the planet remains outside the star even at the deepest moment. This is a grazing transit.
A grazing planet can make a small, rounded dip despite being quite large. The shape of ingress and egress, when the planet enters and leaves, helps distinguish that geometry from a smaller planet crossing fully. Those short intervals deserve attention; the minimum alone has discarded much of the available information.
A camera adds light over time
A telescope does not usually measure an infinitesimal instant. During one exposure the planet keeps moving. Near ingress, part of the exposure may see an unobscured star and part may see an obscured one. The saved value averages those moments.
A long exposure rounds the sharp changes and can make the passage look longer. It also blurs the clue to the path contained in ingress and egress. In a time-resolved observation, shortening the exposure can recover that clue at the cost of collecting fewer photons per measurement. The appropriate choice depends on the transit and the available light.
The TESS example has two-minute cadence. Its file separately records frame timing and live-time information. A time stamp and an exposure length describe different things: the former locates the measurement, the latter tells us what interval the model should average.
Reading: Kipping 2010, Finite Integration Time, arXiv v1 · TESS Science Data Products Description, Revision F
Ask whose light entered the measurement
Suppose a neighboring star adds as much light as the host. The planet still hides the same patch of its own star, but the total baseline light has doubled. The normalized dip becomes half as deep. Interpreting that diluted depth with the area ruler would make the planet look too small.
There is another possibility: the dimming happens on a different, blended star. Two stars eclipsing each other can produce a planet-like shallow signal after their light mixes with a brighter target. Comparing odd and even events, looking for a second eclipse and locating which image pixels change can test that explanation.
The useful question therefore develops in stages. First identify the changing source. Then use the stellar brightness, the path and the exposure to measure a radius ratio. Finally bring in the star’s independently estimated radius. Each step adds information that the little dip by itself did not contain.
Reading: MAST TESS Data Product Overview · Morton et al. 2016, False Positive Probabilities, arXiv v1