This is an important step towards astrophysical modelling. It requires getting light curve data at all orbital phases - not too hard with a 3-hour EW, but needs luck and patience with a 10-day EA. As a rule, light curve data sets from several nights will be needed, and combined in a period analysis application such as PERANSO, as described on the Light Elements page. The result is a light curve phased or folded on the system’s period. There’s a few tricks to this trade however.The light curve of any eclipsing binary can change over time, just as the period can. So it’s best to try to get the complete LC in one observing season. Even then, starspots wandering around can make LCs at the same phase different week by week. Use the Ephemeris Calculator spreadsheet (in Downloads) to find the times to observe phases not yet covered.For purposes of astrophysical modelling, it is very useful to get filtered data in two or three filters, used in rotation. B and V are quite the best to get, then Ic as an optional extra. Your observational data needs to be raw, i.e. get mags for V and C, not V-C. That’s because they need to be transformed to the standard photometric system, eliminating instrumental colour biases. Obtaining your Transformation Coefficients, and transforming your observational data, are beyond the scope of this introduction; but see how-to articles by Benson and Sarty in Downloads. Incidentally it is best not to use transformed data to measure minima, because time-stamping is less precise since every colour mag data point is derived from two of different colours taken at (slightly) different times.If you’re working with a team to get phased LCs, beware that data from different instrument systems can be difficult to reconcile. You may find that Bob’s LCs are systematically brighter than Alice’s, or eclipses are more stretched. In theory if everyone transforms their data, the discrepancies should disappear (that’s the point of transforming). But in practice they usually don’t, quite. This can be due to not-quite-perfect TCs (they never are perfect enough), coloured atmospheric pollution such as smoke haze, or Murphy and his retinue of gremlins. Even transformed data from one instrument system night by night may need a little shifting up or down to align overlaps. Sad but true. (PERANSO handles that.)Your output from this exercise should be complete phased LCs in B and V, and optionally Rc or Ic. If the application or spreadsheet you use to apply your TCs to your observational data will output a B-V light curve, get that too. From it you can read off the colour index of your binary system (usually one star dominates) and hence - assuming it’s on the main sequence - its effective temperature. Often the B-V LC will be wavy, indicating a temperature difference between the two component stars.
Reporting your light curves
It’s a good idea to keep your light curve files in the Dropbox folder for the variable, accessible to your colleagues in the project. They will be important in any paper you write on period (change) and modelling.Forging aheadYour transformed LC data are the input to system modelling, and any paper you write on your modelling should include the light curves. Any article you publish on the system, refereed or not, is always well graced by light curves; as well as allowing sophisticated readers to see important information in the curves.