Celestia Forums

I'm setting a white dwarf binary to place some of my recent magnetic field models, and I'm experiencing an old problem again with the orbital Period and the RotationPeriod. Whatever what I'm doing, after few years, I get a major misalignment between the stars and their bulges, and also with the CMOD object which should be rotating rigidly with the stars. Here's the STC code I'm using :



Here's the SSC code :



If I accelerate time for some years, the SSC model gets a misalignment relative to the stars, like what is shown below. How can I correct that ?



I'm tired of these kind of problems with Celestia. Something has to be done !

Chris,

there's something I don't understand here. My orbital period is set to 0.0000247325 years. The RotationPeriod is then EXACTLY

RotationPeriod = (0.0000247325)*(365.25)*(24) == 0.216805095 hrs.

There is no roundings here. So why the rotating model doesn't keep up with the moving stars ? Internally, is Celestia using 1 year = 365.25 days or what ?

I think that I've found the problem . . . Celestia is using double precision values for the orbital period, but only single precision for rotation periods. This drift shown up before because orbital periods are normally considerably longer than 13 minutes and so it takes a long time to notice it. I think that using double precision for the rotation period should fix the problem.

--Chris

The problem was in fact that rotation periods were stored internally as single precision floats. I modified some files so that doubles are used instead and the misalignment no longer appears when I accelerate time. I've committed the fix to CVS; if you're able to build the CVS version, please verify the fix when you get a chance.

Thanks for reporting this one. It's subtle, and I confirmed that prior to the fix, the orientation of synchronous rotators in the solar system would drift slowly over long time spans (albeit long outside the span when the orbital theories were valid.)

--Chris

Thanks, Chris.

I'll build a fresh Celestia from CVS today and make a report later.

Report :

Updated.

Recompiled.

Checked.

DONE !

chris wrote:I think that I've found the problem . . . Celestia is using double precision values for the orbital period, but only single precision for rotation periods. This drift shown up before because orbital periods are normally considerably longer than 13 minutes and so it takes a long time to notice it. I think that using double precision for the rotation period should fix the problem.

--Chris

Well, that should solve some of my problems on my accretion disk/gas streams that I was having a while ago... Awesome!