Questions before submitting new stars in VSX - [Gaia DR3 and Photometry Filters]

Sat, 11/12/2022 - 16:16

Hi everyone,

I had the surprise to find at least 4 eclipsing binaries not present on the VSX database during a follow up of (65803) Didymos after the impact of DART on Dimorphos. I have a few questions on their submission in VSX in the Gaia era.

The targets are, of course, flagged as variables by Gaia and included in the DR3 eclipsing binary catalog. As far as I know, they are not detected as a variable anywhere else and do not have a light curve outside the epoch_photometry of Gaia (I will check again ASAS-SN sky patrol).

I started and finished a 15 nights observation campain on this field of view (few of them failed because of a focuser issue), I have now the complete phase diagram of all of them with a time step of 1min.

I built their ephemeris from the GAIA DR3 epoch and period. I found an O-C of +1.7h for one of them (an EA star, the rest are EW on time compared to the ephemeris). The Gaia epochs are about 2500 JD old.

My questions are:

  • If I understand well this post:, the credit goes naturally to the Gaia collaboration. I guess I can use the period and epoch they give in the catalog for the variables on time with the ephemeris (right?). But what about the EA star with a +1.7h O-C? Should I use a new epoch of mine and the period I found? 
  • Should I consider writing a JAAVSO paper for the 4 "not but almost" discoveries or add them one by one? 
  • My light curves were made using a Luminance filter from Astrodon ( This pack of filters are built for astrophotography and not photometric studies (Photometry with this particular telescope is quite new, we will add a sloan g filter in the future). The choice of the L filter was made in order to catch stars up to mag 18. One of my new targets is indeed about mag 18. Once they are included in VSX, can I still upload my data (about 1k points per star) using the CV filter with V zero point? Or should I delete my data and start again with a photometric filter? Does anyone have a method to compensate the quantum efficiency effect of the CMOS*? Should I add a note about that in the data file?

(* I mean that two cameras don't necessarily share the same QE, implying two differents magnitudes for the same star)

Maybe the last question belongs to the photometry section, I will post it there if necessary.

Thank you!

Clear skies,

Sacha Foschino


American Association of Variable Star Observers (AAVSO)
Gaia DR3 eclipsing binaries

Dear Sacha,

You already answered correctly to most of your own questions yourself ;)

Let me expand on some points.

About using the Gaia DR3 epoch as the mideclipse epoch.
As far as I can see, they are not providing epochs of mideclipse. This is a common thing with survey catalogues. They may provide epochs but they might not be epochs of minimum so determining an accurate epoch is one of the useful outcomes of a VSX analysis performed by individuals. Since you have your own data, determining an epoch from a single eclipse will be possible. That is the best way to get epochs of minimum, something not usually possible with survey data (excepting TESS or Kepler) due to they survey's observing cadence.
If you used the "TimeRef" date in the Gaia DR3 catalogue, see that then they have the "PhaseE1" and "PhaseE2" columns that will tell you the orbital phase when the eclipses occur. And they are not 0.00 and 0.50. They are just random. You might use the period to determine an epoch converting phase units to days but that doesn't make sense if you have your own data.

So, ignore Gaia DR3 epochs and periods. Just use their data, along with ASAS-SN or ZTF data (depending on magnitude and position), and your own data, and find the best combined solution.

Yes, they should be Gaia discoveries and the primary name should be the Gaia DR3 identifier.

About filters.
For VSX: combine your observations with survey observations, adopting the standard magnitudes in one of the surveys. V is preferred if available. If the star is too faint for ASAS-SN or the data are too scattery, you can adopt the r or g zero point from ZTF (ZTF DR14 is available, not linked from VSX, please go to the IRSA page to get the data.)

For the AID: yes, select CV as the passband if you are using V mags for your comparison stars. When we see CV or CR, we don't expect the magnitudes to be similar to those submitted by other observers because -as you said- different cameras have different sensitivity.

We will wait for your submissions.
Last thursdays of each month we have the VSX Q&A in case you want to ask questions or have doubts.
You can check the schedule here.


Thank you Sebastian for your…

Thank you Sebastian for your quick answer.


I'm sorry, I think I don't understand what you mean here:

"If you used the "TimeRef" date in the Gaia DR3 catalogue, see that then they have the "PhaseE1" and "PhaseE2" columns that will tell you the orbital phase when the eclipses occur. And they are not 0.00 and 0.50. They are just random."

That's true that the TimeRef they give is not the epoch of mid eclipse, so the derived phases are not 0 and 0.5 but it doesn't mean they are random. The difference of phase bewteen E1 and E2 is in the vast majority of about 0.5. This is what Molavi et al. (2022, in prep),, show in the fig. A.1. They also give a confidence on the different models and warn on what population of the catalog we should take with caution. I'm curious to know where you read that?

The phases they give allow to predict when the eclipses occur using their TimeRef, Period and Phases. We can use it to estimate when to start a *first* observation but, yes, I will derived an epoch for the mid eclipse and use it during the VSX submission.


Thank you, I will submit them during this week! :) 

Clear skies!


American Association of Variable Star Observers (AAVSO)

We are saying the same thing :)
The random thing is the reference epoch, which is not a mideclipse time. The difference in phase is usually okay (they warn about some secondary eclipses being artifacts due to poor sampling) so a prediction is possible. It shouldn't be difficult to calculate an approximate Min I using the reference epoch, the period and the Min I phase information given in the catalogue, but that assumes that periods are correct, which it's not always the case, so there will be some risks in following that approach.