search

nanograv / PINT

PINT is not TEMPO3 -- Software for high-precision pulsar timing
Other
117 stars 100 forks source link

sun_angle and solar wind need better testing #893

Open aarchiba opened 3 years ago

aarchiba commented 3 years ago

PR #892 introduces a sun_angle calculation but the testing is basic; it should be compared to some known and trusted values, that specifically include the sun's offset from the solar system barycenter. Likewise we should have some way to test solar wind models built on that.

paulray commented 3 years ago

It looks like the definition of the scale factor is off between Tempo2 and PINT by a factor of 2. When I use zima to create TOAs with SOLARN0 10 and then fit them with Tempo2, it gets a value of 5.0. But the shape and location of the cusp is fine, so I think the angle calculations are correct.

paulray commented 3 years ago

Spoke too soon. There is a small difference in the solar angle computed by PINT and Tempo2.

sunanglediff

paulray commented 3 years ago

Something is funky with the solar wind and PINT. I create a set of TOAs like this: zima sw.par sw.tim --ntoa 365 --duration 365 --error 0.05 --plot

Then I try to fit with PINT and it does not get 0 residuals, but has a strange wiggle.

This is the par file (rename to sw.par): sw.txt

Making this more difficult for me is that something in all the new changes has broken pintempo and pintk so they are giving nonsense.

scottransom commented 3 years ago

I'm going to try and fix the pintk (and maybe pintempo) issues today.

aarchiba commented 3 years ago

Be aware of #906 which fixed subtle bugs with solar angle

paulray commented 3 years ago

Possibly relevant paper: https://arxiv.org/abs/2012.11726

aarchiba commented 3 years ago

I had a student who looked at solar wind modelling, and indeed it's much messier than our usual assumptions. But surely at least we can know where the Sun is and what this model predicts...