Alternative Splash image; Double Star orbits

[t00fri]

If you've updated Celestia from CVS recently, there should be a file in the data directory called stars.txt. This is the human-editable file from which stars.dat is built. There are six fields per line: HIP catalog number, RA in decimal degrees, declination, distance, apparent magnitude, and spectral type (same as in spectypesWD.txt.) It's a lot more compacy than the complete HIPPARCOS catalog, but contains all the information you should need. It's also got all of Grant's corrections for errors in the HIPPARCOS catalog.

--Chris

Thanks...I just had remembered stars.txt myself. since spectypesWD.txt was just a little short of what I actually need...

Of course I am always up-to-date with CVS.

Bye Fridger

[granthutchison]

Fridger:
Sorry, I was out of comms for the last 24 hours ... fortunately the Cairngorm Massif has yet to be Bluetooth enabled. But it looks like you've managed without me .
Specific problems I could see with using the 6th catalog were:
1) Retrieving spectral data for the secondary stars.
2) Retrieving postion data for the secondary stars (necessary to generate a good barycentre for widely separated binaries).
3) Deriving masses for both components in order to generate a barycentre (presumably a look-up table against spectral class).
4) Finding catalogue numbers for the secondaries in order to check that we are not dealing with a trinary system or more: for instance, it cannot be established from within the 6th catalog that Hip 73184 is the primary orbited by the double system specified under Hip 120005.

It's also necessary to parse the star names in the 6th catalog to some extent, to check if we are dealing with the orbit of AB, AC, BC or etc.

Grant

[t00fri]

Fridger:
Sorry, I was out of comms for the last 24 hours ... fortunately the Cairngorm Massif has yet to be Bluetooth enabled. But it looks like you've managed without me .
Specific problems I could see with using the 6th catalog were:
1) Retrieving spectral data for the secondary stars.
2) Retrieving postion data for the secondary stars (necessary to generate a good barycentre for widely separated binaries).
3) Deriving masses for both components in order to generate a barycentre (presumably a look-up table against spectral class).
4) Finding catalogue numbers for the secondaries in order to check that we are not dealing with a trinary system or more: for instance, it cannot be established from within the 6th catalog that Hip 73184 is the primary orbited by the double system specified under Hip 120005.

It's also necessary to parse the star names in the 6th catalog to some extent, to check if we are dealing with the orbit of AB, AC, BC or etc.

Grant

Indeed, Grant,

your 4 points match my concerns I went to bed with yesterday;-)

Notably the issue of finding the orbit's barycenter. There is no way out, we need the masses for that!

Do you agree that the coordinates given in the 6th catalog are those of the primary?

In that context, what confuses me a little is this:

Clearly the observations are done in another frame, where the primary is sitting in the origin, the observer is "comoving" and the secondary's position is characterized by (rho, theta), see their ephemeris table. From there, they deduce --by fitting-- all the standard orbit parameters. My confusion refers to the question to what extent barycentric considerations might have been already folded into that analysis somehow?

At least there are very few explanations referring to these delicate issues in the 6th catalog.

Next, we indeed do not seem to have a systematic source for spectral classes of the secondaries, to be exploited on the level of "mass production", i.e. automatization via Perl. Too bad...

Next, unlike the catalog of spectroscopic doubles I was mentioning above, in the 6th, there are no masses given. Full stop.

So you see me ready to complete my script (more than half is already coded...), but what to do with these issues? Clearly in the mass=0 approximation of the seconday, we are almost done...

I also have implemented already the cross correlation with stars.txt to extract the distance.

Bye Fridger

[granthutchison]

Do you agree that the coordinates given in the 6th catalog are those of the primary?

I believe so.

My confusion refers to the question to what extent barycentric considerations might have been already folded into that analysis somehow?

My understanding is that these are elliptical elements centred on the the primary ... they're commonly used to derive angular separation and position angle for a given epoch. The barycentre is of little relevance to most observers, and would anyway require precise measurement of absolute (rather than relative) position, which is much harder to do.

Grant

[buggs_moran]

You guys make me feel stupid.

But, sincerely, thank you for all of your efforts. I teach Astronomy and Algebra through Pre-Calculus to a bunch of guys who go "WOW" every time I show them things in Celestia. I cannot wait for your work on binaries. I just wish I had it in me to go back to school to learn half of the things you know. (Especially the transparencies Fridger.)

[t00fri]

buggs_moran:

in view of the principal problems we are facing right know I feel like you do


Grant:

do you think it makes sense that I complete the script for the m_secondary/m_primary->0 limit, just to have some 0th approximation in hands? Perhaps it's not the most stupid approach, to get going. Any further wisdom we might acquire thereafter could easily be incorporated in further iterations....

This week I have "working vacations" meaning an amazing program ahead of me: selecting & buying a new car, a new computer, DSL 3000 Mbit/s (at last!) and many more items that are more than overdue. That means also setting up another compi with SuSE 9.1, so who knows how much time will be left for the Perl script

--------------ignore if uninterested----------------
As to the computer, I broke for the first time my "holy rule" to always buy the coolest of a given CPU family. My wife urgently needs a new machine for preparing a big-size lecture for hundreds of students. So I was faced NOW to buy the hottest of the Prescott gang (3.2. GHz). I had to stick to the 478 socket, since otherwise I had to throw away my great FX 5900 Ultra/256 MB graphics card...After I had worked out that even a gigantic Zalman copper cooler would have to turn on maximum rpm together with 4 (!) other fans with 400 Watts of power supply, I gave up . Frightening: 250 Watts converted into heat

This morning, in the last minute, my dealer found me a left-over (boxed) Northwood 3.2 GHz that cooks 60 Watts less than the corresponding Prescott at full load and is even slightly faster than the Prescott...

In my new machine I put 3GB (!) of pretty fast RAM. So -- after my new DSL line works (2 weeks from today) you may expect to find a few gigantic new textures in my old TextureFoundry
---------------------------------------

Bye Fridger

[granthutchison]

do you think it makes sense that I complete the script for the massless limit of the secondary, just to have some 0th approximation in hands?

This is as far as I got myself ... I never actually built the ssc (as it would have been at that time), since it was going to be just full of plain white "placeholder" objects, nothing that could really be used in Celestia.
If you feel you have the time to complete the script it can't do any harm, and of course the basic structure will then be available for building on, as you say.

(Meanwhile, I've just placed the last star in my model of local space out to 25 ly, combining some automation with a bit of old-fashioned hand-crafting. There are some nice trinary and quaternary systems out there.)

Grant

[Evil Dr Ganymede]

Notably the issue of finding the orbit's barycenter. There is no way out, we need the masses for that!

And note that masses can't be easily tied to spectral class - an M3 V star will have a very different mass to an M3 II star, for example. You'd have to tie it to the size too, but the giant stars can be arrived at via different evolutionary paths so they're not unique to a given mass.

I think the best solution would be (guess what) to be able to specify masses manually for each object in the ssc. So far it sounds like Chris is unwilling to allow that, but I can't see how you can make meaningful barycentres without it. If any barycentres are going to be unrealistic because they are not tied to mass, then there seems to be little point in adding them to Celestia.

[t00fri]

Notably the issue of finding the orbit's barycenter. There is no way out, we need the masses for that!

And note that masses can't be easily tied to spectral class - an M3 V star will have a very different mass to an M3 II star, for example. You'd have to tie it to the size too, but the giant stars can be arrived at via different evolutionary paths so they're not unique to a given mass.

I think the best solution would be (guess what) to be able to specify masses manually for each object in the ssc. So far it sounds like Chris is unwilling to allow that, but I can't see how you can make meaningful barycentres without it. If any barycentres are going to be unrealistic because they are not tied to mass, then there seems to be little point in adding them to Celestia.

Dr. Evil

but remember, we are talking here about "mass production" of binary orbit data for Celestia from a given catalog, rather than hand editing individual binary systems.

I you can quote me any catalog sources for the masses in question, I promise I'll be real nice to you in the future

Bye Fridger

[chris]

Notably the issue of finding the orbit's barycenter. There is no way out, we need the masses for that!

And note that masses can't be easily tied to spectral class - an M3 V star will have a very different mass to an M3 II star, for example. You'd have to tie it to the size too, but the giant stars can be arrived at via different evolutionary paths so they're not unique to a given mass.

I think the best solution would be (guess what) to be able to specify masses manually for each object in the ssc. So far it sounds like Chris is unwilling to allow that, but I can't see how you can make meaningful barycentres without it. If any barycentres are going to be unrealistic because they are not tied to mass, then there seems to be little point in adding them to Celestia.

You can still manually compute the barycenters using a mass estimate--there's no need for Celestia to know anything about the mass of the stars. While I may introduce a mass field for stars, I don't plan on using it for orbit calculations. I'd rather limit the amount of processing Celestia has to do on data files.

--Chris

[granthutchison]

You can still manually compute the barycenters using a mass estimate--there's no need for Celestia to know anything about the mass of the stars.

That's right. You just need to do a simple sum to place the barycentre for a given system when the masses are known, or reasonable estimates are available.
The problem Fridger and I have is obtaining mass estimates for the 1,800 systems in the 6th catalog, when the 6th catalog doesn't even contain a catalogue number or spectral class for the secondary star. Being able to "add the masses by hand" just isn't relevant to the problem.

Grant

[ANDREA]

You can still manually compute the barycenters using a mass estimate--there's no need for Celestia to know anything about the mass of the stars. While I may introduce a mass field for stars, I don't plan on using it for orbit calculations. I'd rather limit the amount of processing Celestia has to do on data files.--Chris

Hello Chris, Evil, Fridger, Grant, don't you think that, starting from Fridger's Oct 07, 10.55 PM post, the original subject
-Download an alternative Splash Image for celestia 1.3.2-
is changed in a VERY different one?
If you agree, I would suggest to move the given posts in the Physics and Astronomy department, with its own subject (Bynary Orbits Barycentre, e.g.).
This could make easier to find it in the future, using search motor.
By

Andrea

[t00fri]

You can still manually compute the barycenters using a mass estimate--there's no need for Celestia to know anything about the mass of the stars. While I may introduce a mass field for stars, I don't plan on using it for orbit calculations. I'd rather limit the amount of processing Celestia has to do on data files.--Chris

Hello Chris, Evil, Fridger, Grant, don't you think that, starting from Fridger's Oct 07, 10.55 PM post, the original subject
-Download an alternative Splash Image for celestia 1.3.2-
is changed in a VERY different one?
If you agree, I would suggest to move the given posts in the Physics and Astronomy department, with its own subject (Bynary Orbits Barycentre, e.g.).
This could make easier to find it in the future, using search motor.
By

Andrea

Hi ANDREA,

certainly I agree with you that from the point of view of systematics and "discipline", we should harbor this discussion in a new thread. Already several days ago
I was about to start moving at least my posts from after when I switched topics in my response to Chris:

I am enjoying the new "binary orbit" developments very much...

If there is demand, I could easily hack another Perl script to generate at once thousands of binary orbits in the adequate Celestia input format from the official double star catalogs. As you know, Perl can calculate very well and accurately. It's easy to make the script decide on the fly what parameters to use to derive the orbits in each case...

to an independent and more sensibly headed thread.

But then I realized ...

that I also liked the way such discussions tend to drift and develop into something new....

Also I was not sure what would be the appropriate department. The issue is definitely about binary orbits in Celestia, so Physics and Astronomy seemed not too appropriate. Development is also not perfect and much less read...

So all in all, perhaps it's too late now . Certainly, if Chris wants to move us all, it's fine with me.

Since I have been already "locked" today in another thread, being "moved" would be another exciting experience . A clearcut manifestation of the dynamics of this forum...

Bye Fridger

[t00fri]

...
You can still manually compute the barycenters using a mass estimate--there's no need for Celestia to know anything about the mass of the stars.
--Chris

Chris,

could you share your wisdom about estimating the masses? Perhaps one could make something usable from it.

OK, we certainly know Kepler's 3rd law:

Code: Select all

T^2/a^3 = 4 Pi^2/G*(m1+m2)


and T,a, are known from the 6th catalog.

Let d_i =|x_i -x_b| be the distances of the two stars i=1,2 from the barycenter x_b, then

Code: Select all

d1/d2 = m2/m1


Knowing x_b would allow to solve these equations for m1,m2. From the empirical data, the movement of the barycenter x_b = b*t +c can in principle be determined if the observations have enough precision and exist over long enough times...

That's all I could think of.

Bye Fridger

[chris]

...
You can still manually compute the barycenters using a mass estimate--there's no need for Celestia to know anything about the mass of the stars.
--Chris

Chris,

could you share your wisdom about estimating the masses? Perhaps one could make something usable from it.

I don't have any special advice on estimating star mass that you and Grant aren't already aware of. I was just pointing out that if you do have the masses and a primary focused orbit, then you can easily compute the barycenter and semimajor axes for a binary system. Determining the masses is certainly the hard part

--Chris

[ANDREA]

If you agree, I would suggest to move the given posts in the Physics and Astronomy department, with its own subject (Bynary Orbits Barycentre, e.g.).
This could make easier to find it in the future, using search motor.
By Andrea

Hi ANDREA, Also I was not sure what would be the appropriate department. The issue is definitely about binary orbits in Celestia, so Physics and Astronomy seemed not too appropriate.

Can you explain me why?
You four are speaking of the orbital properties of two or more bodies one around the other and around their common baryicentre, and this in Celestia, obviously, but the first goal of all the involved people, like you, is to make things in Celestia as real as possible, so along the Phisics and Astronomy laws.
So I cannot agree that this is not astronomy, sorry.

Development is also not perfect much less read...
So all in all, perhaps it's all too late now
Bye Fridger

You started this post, you are the boss.
Anyhow I was not particularly pushing to move it to Phisics and Astronomy or any other department, but my interest was only to split two subjects that in my opinion are truly not compatible.
By

Andrea

[Evil Dr Ganymede]

You can still manually compute the barycenters using a mass estimate--there's no need for Celestia to know anything about the mass of the stars. While I may introduce a mass field for stars, I don't plan on using it for orbit calculations. I'd rather limit the amount of processing Celestia has to do on data files.

--Chris

How many people do you think are going to know how to calculate the barycentres of a system accurately (especially one more complex than a binary) in order to put them in?

Celestia doesn't need to be processing the barycentres of every multiple system in the universe, it could surely just calculate the parameters of the system that's being looked at?

As for finding the masses, well, you'd need some kind of lookup table depending on the star type and size. Though as I said, this becomes a problem when you're looking at giants that could have evolved from a wide range of main sequence stars. Plus main sequence stars change their spectral type as they get older. You'd have to make some assumptions, clearly.

There are mass lookup tables in some scifi roleplaying games (Classic Traveller, for example) but they're not tremendously accurate, in part because the game is over 25 years old now. I have thought about this problem for many years now and have been trying to get stellar masses sorted out for a realistic system generator for that game and lately I've ended up using the Geneva Stellar Evolution grids as a base, but that's only going to tell me how stars with a given mass and metallicity evolve (that's how I could see that most giants could evolve from a range of main sequence stars). It is certainly not an easy problem to solve - that's why I think it's much easier to enter the masses by hand in the STC and have Celestia calculate the orbital distances for the stars, instead of hoping that the values you have in the lookup table are correct. Or at the very least include a separate 'calculator' program with Celestia that allows people to figure these things out (no, don't ask me to do it. I can't program in anything other than Fortran, and I have no time or inclination to learn another language).

As I see it, the problem with NOT having any mass field at all is that the resulting system will most likely not be realistic. And if you're going to the trouble of adding barycentres, then what's the point of having them if you can plonk in any value you like with no reason to care about whether it's actually real or not? It seems that this would reduce Celestia's educational potential somewhat, for starters.

[t00fri]

...
Can you explain me why?
You four are speaking of the orbital properties of two or more bodies one around the other and around their common baryicentre, and this in Celestia, obviously, but the first goal of all the involved people, like you, is to make things in Celestia as real as possible, so along the Phisics and Astronomy laws.

So I cannot agree that this is not astronomy, sorry.

It is of course astronomy, but in this department we tend to discuss preferably more general matters, not specifically concerned with Celestia development.

But really, I am more concerned that this important discussion takes place at all than where it takes place.

Bye Fridger

[selden]

Fridger,

I hope you don't mind, but...

I took the liberty to change the title of this thread to be

Alternative Splash Image; Double Star orbits

Of course, you can change it to whatever you want, too.

[t00fri]

Fridger,

I hope you don't mind, but...

I took the liberty to change the title of this thread to be

Alternative Splash Image; Double Star orbits

Of course, you can change it to whatever you want, too.

Excellent Selden. At least this hints at the right content of the thread. It does not stand particularly for my ability to create catching-on logically coherent thread titles, though .

Bye Fridger