I just want to add stars

[Sadachbia]

Yay, the brown dwarf companion is now rendering! Unfortunately, though I keep increasing the distance it's supposed to be orbiting around the system barycentre, it's still INSIDE the larger star (and exactly in its centre, at that). Do the RA/declination/distance for the two stars need to be different values from the barycentre's RA/declination/distance? (I just used the same numbers.) Also, Star A's orbit around the barycentre right now is 180.5 while Star B's semimajor axis is 810000.2 (as I said, I kept increasing it; if the units are AU, the stars should be way far apart by now instead of superimposed on each other).

I know that for stellar companions of such disparate mass, the barycentre of the binary pair would be located closer to the centre of the larger star (likely inside its radius), with the smaller star following a much larger path around the barycentre. (The massive Star A's orbit should probably be smaller, but nothing is finalised at the moment. This system is also going to have planets eventually, which would also affect the stars' orbits around the system barycentre; however, I want to make sure the stars are good to go before I worry about planets.)

P.S. I love how the brown dwarf texture is a very nice dark pink, in accordance to what I read on Wikipedia about their looking more like "magenta coal tar dye" to human visual perception, thanks to the absorption spectra of Na and K

[selden]

When stars orbit around a barycenter, they don't need their own RA, Declination and Distance values. Only the Barycenter needs them. The stars going around the barycenter should have EllipticalOrbit specs.

Please review the STC description in the WikiBook. It includes an example.

http://en.wikibooks.org/wiki/Celestia/STC_File

[Sadachbia]

I deleted the RA, declination, and distance codelines from the entries for Star A and Star B. Now only the barycentre has the 3 coordinates. My code is as follows:

Code: Select all

500001 # Eazurcha A
{ ...orbital data... }
500003 # Eazurcha B
{ ...orbital data...}

500001 "Eazurcha A:HZ 63996 {"
 SpectralType "F9I-b"
 AbsMag -2.01
}
500003 "Eazurcha B:HZ 63997 {
 SpectralType "T1V"
 AbsMag 14.00
 Texture "browndwarf.*"
 Radius 100000
}


However, the brown dwarf is still located inside the larger star. I also can't select them by name, only by typing in the fake HIP identifiers (500001 or 500003). I know the names in the codelines for the barycentre orbital data don't affect what names Celestia think the stars have, because they're after a "#" in both cases (the "comment" marker that causes Celestia to ignore everything on that line after the marker).

[selden]

Try this:

Code: Select all

Barycenter "Eazurcha"
{
   RA  100
   Dec -60
   Distance 40
}

Star 500001 "Eazurcha A:HZ 63996"
{
OrbitBarycenter "Eazurcha"
SpectralType "F9I-b"
AbsMag -2.01
Mass 1.04
 
   EllipticalOrbit {            
      Period         1138 # sqrt ((a1+a2)^3)
      SemiMajorAxis     5 # mass ratio 0.05:1.04 (m2:m1)
      Eccentricity      0.52
      Inclination      82.22
      AscendingNode    67.30
      ArgOfPericenter   4.02
      MeanAnomaly     200.13
   }         
}

Star 500002 "Eazurcha B:HZ 63997"
{
OrbitBarycenter "Eazurcha"
SpectralType "T1V"
AbsMag 14.00
Texture "browndwarf.*"
Radius 100000
Mass 0.05
   EllipticalOrbit {            
      Period         1138
      SemiMajorAxis   104 # mass ratio 1.04:0.05 (m1:m2)
      Eccentricity      0.52
      Inclination      82.22
      AscendingNode    67.30
      ArgOfPericenter 184.02
      MeanAnomaly     200.13
   }         
}


[Sadachbia]

Thank you, Selden~~the altered code you provided makes both stars render in the proper place. I guess this means I have to calculate the actual variables for both stars of interest anytime I want to create a binary system, though. (At least I don't have to calculate ACTUAL star mass. It seems I can just have the ratios in the codelines, yes?)

So this is clearly the solution to fictional multiple star systems, but if I want to add additional names to known multiple star systems, is there a way to copy Celestia's star data so I don't have to calculate from scratch? Apparently, I can't rename stars just by adding lines of code like "(HIP number):name 2:name 3" into my STC file. (I tried to do this for the two Eazurcha stars, as well as having the names in quotes in the barycentre orbital entries the way your code was, but this didn't allow me to select the stars using anything besides the fake HIP numbers; I even backed up my starnames.dat file and added lines to say "500001:Eazurcha A:HZ 63996" and "500003:Eazurcha B:HZ 63997", but that didn't do what I wanted either, so I restored the starnames file to the backup.) So I must still be doing something wrong >_<

[selden]

Thank you, Selden~~the altered code you provided makes both stars render in the proper place. I guess this means I have to calculate the actual variables for both stars of interest anytime I want to create a binary system, though. (At least I don't have to calculate ACTUAL star mass. It seems I can just have the ratios in the codelines, yes?)

You're quite welcome.

You're right in that to be realistic, the semimajoraxes of the two orbits of a binary have to be in proportion to the stars' masses. Their sum can be any value you want, though -- close together or far apart -- which is what determines the period. The two stars have to be on opposite sides of their orbits, so their ArgOfPericenter values have to be 180 degrees apart. All of the other values in their orbital definitions have to be the same for both stars, but can be any value you want, determined by what you need them to be.

So this is clearly the solution to fictional multiple star systems, but if I want to add additional names to known multiple star systems, is there a way to copy Celestia's star data so I don't have to calculate from scratch?

I've found it easiest to look up the RA and Dec in Simbad ( http://simbad.u-strasbg.fr/sim-fid.pll ), and write down the other values as shown by Celestia's display. One probably could get them all by using a CelX script, but I've never bothered doing that.

Apparently, I can't rename stars just by adding lines of code like "(HIP number):name 2:name 3" into my STC file. (I tried to do this for the two Eazurcha stars, as well as having the names in quotes in the barycentre orbital entries the way your code was, but this didn't allow me to select the stars using anything besides the fake HIP numbers; I even backed up my starnames.dat file and added lines to say "500001:Eazurcha A:HZ 63996" and "500003:Eazurcha B:HZ 63997", but that didn't do what I wanted either, so I restored the starnames file to the backup.) So I must still be doing something wrong >_<

When you provide a replacement Star definition, you have to include all of the Star's parameters in the new definition. The original definition is completely deleted. Unfortunately, unlike SSC defintions, there is no Modify statement for STC catalogs.

[Sadachbia]

When you provide a replacement Star definition, you have to include all of the Star's parameters in the new definition. The original definition is completely deleted. Unfortunately, unlike SSC defintions, there is no Modify statement for STC catalogs.

So is there no way to rename members of a binary star system? What you said means that including the additional names with all the star data necessary overwrites the orbital data and puts the brown dwarf in the middle of the giant. I don't know what else to try.

I can type in the name of the barycentre ("Eazurcha") and go there, where both stars are clearly visible, just not named with anything besides "HIP 500001" and "HIP 500003". Is there no protocol to simply append "A/1" or "B/2" plus the name of the barycentre to the two or more stars orbiting it?

[selden]

Your description of your problem confuses me. Using the example that I provided, each star has three names, all of which are shown and all of which can be used to goto the system.

You don't have to "rename" any stars if you define them yourself from scratch. All of the names should be provided by you in the initial star definition. You don't have to rename an existing star defined in Celestia unless you really want to. If you do, you must specify the Hipparcos number used by Celestia for that star, which is not what you did in your example.

Below is a screenshot of the system, after I used the name Eazurcha A to goto it. Then I used the mouse to select Eazurcha B so B's orbit would be hightlighted.

And here's a Celestia URL to go to it, too.

cel://Follow/Eazurcha%20A/2011-08-29T11 ... rc=0&ver=3

p.s. It would help me to understand your problem if, instead of describing what you're doing in generalities, you would provide a complete example, explicitly specifying each step that you've done and including the full text of the STC file that you're trying to use.

[Sadachbia]

I found the problem. I had a # where it didn't need to be ("500001 # "Eazurcha A:HZ 63996"). I am still a noob with Celestia. ._.;

I've also been following the news on Hurricane Irene, and wherever you are in NY, Selden, I hope you're staying safe. It looks like a boatload of not-fun for everyone on the East Coast.

[selden]

It's good to know you found the problem!

Where I am we only had the kind of rain and wind that we usually get from large thunderstorms, plus a couple of very brief brownouts, also common with thunderstorms. It's a lot worse downstate.

[Sadachbia]

Now for a new question topic: Is there a way to find out what fake HIP number Celestia uses for stars that don't have a real HIP designation? (Such as stars discovered with the OGLE or WASP projects, etc., or double stars in systems where the Hipparcos number is the same.)

[selden]

Celestia does not use fake Hipparcos numbers, since they are no longer required in order to define Stars.

You'll need to search the STC files which are in Celestia's data directory.
charm2.stc nearstars.stc spectbins.stc
extrasolar.stc revised.stc visualbins.stc

[Sadachbia]

They're not required? What about for fictional stars? The walkthrough in "A (not so) Brief Introduction..." says that Celestia needs to identify stars with HIP numbers, whether they're real (up to something in the 100-thousands) or fake (suggesting that fictional stars begin with the number 500000). I've already numbered several of my fictional stars with numbers above 500000, though I guess there's no need to change that since it seems to work all right. Will definitely take a look at the STC files you listed, though.

[selden]

Sorry: I haven't done any significant changes to the page for quite a while. If one is defining stars which have real Tycho numbers, it's appropriate to use numbers which Celestia will translate into the correct on-screen values, of course. It used to be that fake Hipparcos numbers were needed when defining stars which didn't have real ones, but that's no longer the case. Recent versions of Celestia can keep track of Stars without them.

[Sadachbia]

Okay, cool! Thank you :3

[Sadachbia]

Noob math question...

Code: Select all

      Period         1138 # sqrt ((a1+a2)^3)
      SemiMajorAxis     5 # mass ratio 0.05:1.04 (m2:m1)

.....

      Period         1138
      SemiMajorAxis   104 # mass ratio 1.04:0.05 (m1:m2)


What mathematical operation gives me a number I can use from the mass ratios? I feel like I should know how to do this, but dividing m1 by m2 (or m2 by m1) and then multiplying (a1 + a2) by the resultant fraction to calculate one star's semimajor axis doesn't give me the right numbers. (I've always used a lot of trial and error when doing math problems, trying different operations until I got the same solution as the textbook. Not the most elegant way. Example: I added the semimajor axes of the two binary stars in my fictional M31 system, which came out to 109, and the square root of 109^3 is 1137.99; pretty much the same result for the period as in Selden's calculations. I just don't know how the m2:m1 ratio works for dividing up the a1 + a2 value into its components.)

[selden]

(In what I've written below, I've tried to be as clear as possible, including mentioning things I'm sure you already know. However, some people reading this might not be aware of some of them.)

You first have to decide an appropriate separation for the stars. This is the sum of the two stars' semimajor axes.

Typically "a" is used as the symbol for an object's semimajor axis, and one uses an appropriate number to identify the stars. As a result, the total separation between the two stars can be written as (a1+a2).

At any particular time, the distance of one of the stars from the barycenter is inversely proportional to its mass. That is, low mass stars orbit far from the barycenter while high mass stars orbit close to it. One way of looking at is that low mass stars don't pull on high mass stars very hard -- it's like a teeter-totter. To be balanced, someone heavy sits near the axis and someone light sits farther away.

Then, knowing the stars' individual masses (let's call them m1 and m2) and the sum of their SMAs, you can calculate their individual SMAs using the formulae

a1 = (a1+a2) * m2/(m1+m2)
and
a2 = (a1+a2) * m1/(m1+m2)

These are how I'd write them for calculating in a Fortran program. When writing them on paper, they might look more like this:

Code: Select all

               m2
a1 = (a1+a2) -------
             (m1+m2)

and

               m1
a2 = (a1+a2) -------
             (m1+m2)


Does this help?

p.s. In the above, I'm assuming that the semimajor axes (a) are measured in units of AU, the distance of the Earth from the sun, and that the masses (m) are measured in units of the Sun's mass. Otherwise one has to take into account a Gravitational Constant to convert the units appropriately.

[Sadachbia]

Yes, that's very helpful! The equations make sense to me and I can see how you got them. I like when math makes sense to me. (But then, who doesn't?) Thank you for explaining them. I appreciate it a lot.

Binary stars that already exist in Celestia's data files have information about the stars' separation; for example, it says that the two components of UPS Andromedae are about 750 au apart. In that case, I can just use the value of 750 as the (a1+a2), and that should give me the period that the two stars actually orbit in the real universe, right? Not having to make up numbers for some multiple star systems makes my job a little easier, because many of the objects in my addon are going to be real celestial objects with fictional data (names/planetary companions) added.

Thus, for the UPS And system (which I'm naming "Prothoe" in my fictional universe), I get the following:
(a1+a2) = 750.52
m1 = 1.28x Sun's mass (this number is from the Wikipedia entry on Upsilon Andromedae A)
m2 = 0.5x Sun's mass (Upsilon Andromedae B is a star of class M4 V, and doesn't have its own Wikipedia entry, but 0.5 MSol appears to be the average mass for M4 V stars)

Plugging in these values for a, m1, and m2 gives me a SMA for UPS And A of 210.82 au, and one of 539.70 au for B. Calculating P (square root of 750.52^3) comes out to 20560.96, which I assume is in units of Earth years. That's a very long period, but then, these stars seem to orbit fairly far apart. It would be interesting to see if the 3 known planets of the UPS And system are tidally locked to either star (or if they orbit one star of the pair instead of the barycentre) or how the presence of the stellar companion affects the orbital resonance.

Are my calculations okay? Do you get similar answers when you use these values?

[selden]

Another way to check the individual SMA values of the stars is to compare them directly to the stars' mass ratio (which, I realized later is what you'd actually asked). When viewing stars, however, you can see only the sum of the two SMAs. Using the ratio directly works only when you're making up imaginary binaries and can specify an arbitrary value for one of their SMAs.

a1:a2 = m2:m1

so

a1 = a2 * m2/m1

so

a1 = 539.70*(0.50/1.28) = 210.82
which is right.

Since the sum of the stars' masses is somewhat larger than the mass of the Sun, the period of a planet orbiting our own Sun at the same distance would be somewhat slower. The simplified formula for the orbital period of a planet orbiting our own Sun (which has a mass of 1 in this format, while the planet's mass is too small to make a difference) is



P = a^(3/2)
= (538.7+210.82)**1.5
= 20560.96 years.
That's the same value you got, so it seems you left out the sum of the stars' masses when calculating the period. You have to divide (the cube of) the distance between the stars by their total mass:

[edit: Oops. I got my parentheses in the wrong place. Sorry. Fixed.]



P = sqrt( ( (a1+a2)^3 )/(m1+m2) )
= sqrt( ( (210.82+538.7)**3)/(1.28+0.5) )
= 15380.30 years.


p.s. a useful Web page for calculating or verifying some binary orbital parameters is
http://hyperphysics.phy-astr.gsu.edu/hb ... istar.html
in its section on visual binaries.


I hope this helps.

[Sadachbia]

It helps a ton. Thank you again for being so patient with me. I always forget that most of the equations I run across assume a star of 1MSol, and adjustments have to be made if that's not the case.

Does calculating the orbital period/distance of fictional planets (whether they orbit a star with mass the same as the Sun's or different) always ignore the planet's mass because it's too small compared to its primary to make a noticeable difference? Would it make sense in a system with widely-separated components, like UPS Andromedae, to have planets orbiting one star or the other instead of the barycentre? (Orbiting one star would mean I would use just that star's mass in the SMA/period calculation, as opposed to the mass of both stars. Right?)