Incorrect planetary Shapiro delays

[Former user]

Hi All,

I'm working with a really good summer student (Josef Zimmerman) on some PINT improvements and detailed checks and we uncovered a problem in the TEMPO2 planetary Shapiro calculations that I think we saw clear back in 2015 when we were starting work on PINT but didn't follow-up on (sorry about that!).

It seems like the Earth->Planet vector that is used in TEMPO is likely actually the Planet->Earth vector, as the returned Shapiro delay is maximum when the planet is on the other side of the sky from the pulsar!

The attached two plots show the problem. They show the PINT/TEMPO2 shapiro delays returned as a function of angular separation of the planet and the pulsar (as computed by Astropy).

I haven't looked at the T2 code, but I suspect there is a very quick bug fix for this.

Cheers,

Scott R, Josef, and the PINT team

[Michael Keith]

This is why PINT is so important I think. We’ll look into it.

[heng xu]

We also found the same problem, adding a minus before planet_earth (it read coordinate of planet respect to earth from jpl_ephem, should labeled earth_planet, however it is labeled planet_earth, caused this fault) vectors when calculating “rsa” in shapiro_delay.C should solve this problem.

[Michael Keith]

This should be fixed as of 380c4c4