Sat, 08/29/2015 - 20:31
In the photometric tables, the magnitude and color index of the stars come from many different sources. For bright stars (mag. <6), almost the stars come from BSC and Tycho-2. The Tycho-2 is much more accurate than BSC. The first has errors of few hundred of magnitude while the second has errors of 0.1 mag. Is there a reason why it is used BSC? Would not it be better to use just the Tycho-2?
Regards,
Luigi
I believe the reason Yale BSC is used is because Tycho-2 saturates at the brightest stars, and thus is not a complete catalog when you get brighter than mag 2 or 3. Maybe someone on the chart team can verify this?
Mike
When we were populating the Variable Star Database (VSD), the set of comparison stars displayed by VSP, we used a group of about a dozen photometric catalogs to choose from to obtain the best available photometry. We used BSC with a higher priority than Tycho, because it had true standardized B&V values, along with R,I in some cases. Most of that photometry was obtained using photoelectric systems in the mid-20th century. However, the one drawback for the BSC is that it does not contain the photometric uncertainties. Therefore, I used a blanket "0.1mag" for all BSC photometric uncertainties, with the understanding that they would usually be far better than this, but I wanted to be conservative. Now, 10 years later, my feeling is that using Tycho2 transformed B,V magnitudes is probably a better choice than the BSC/Tycho mix that we currently have, and I will most likely deprecate the BSC when the next update is made.
That said, both the BSC and the transformed Tycho2 magnitudes are very good, and better accuracy than most non-professionals can obtain. You can use the mix safely for variable-star observations.
Arne
Hi Mike, Arne,
thanks for your answers.
Arne, recently I have created a spreadsheet (see attachment) that converts the magnitude VT and color index (B-V)T of Tycho-2 in the Johnson's system. I have used spline interpolation method with data that I founded in the article "The Hipparcos and Tycho Photometric System Passbands" of Michael Bessel. I don't know how AAVSO does the conversion but I have tested the spreadsheet on a large number of stars and the conversion is very accurate.
In my photometric observation I would only use the Tycho-2 catalog. I would like to work in this way: in the photometric table, when I find a star that comes from another catalog, I take the VT and (B-V)T from Tycho-2 and convert them in V and (B-V) using my spreadsheet.
Can I do this? What do you think?
Regards,
Luigi
Hi Luigi,
If you use your own calibrations rather than the values listed by VSP in the Photometry Table, you need to place the revised values in the "comments" field for your submission. A researcher will assume that you are using the photometric values that were available through VSP at the date of your observation unless you explicitly tell them otherwise.
A fancy fit is probably not worth it for Tycho. It is reasonably close to the standard system, so something simple like a quadratic or cubic fit is usually sufficient. More elaborate fits really need to take into account the spectral type of the star, especially its luminosity class. For VSP, we use the cubic fits for main sequence stars given by Mike Bessell. You can also use the already-converted values in the Kharchenko (2009) catalog, available through Vizier.
Arne
I don't agree with Tycho-2 being the best source for bright star photometry. For the reason Arne mentioned initially, I'd prefer true UBV values, preferrably from the GCPD where several databases are included. That is the better source for UBV photometry I found. Even HIPPARCOS is incredible consistent for non-red giant stars and matches the UBV catalogue to 0.01 mag. or less using Bessell transformations. The B-V cited in the HIPPARCOS catalogue comes from UBV values when they are available and form Tycho when they are not (Tycho-1, which is better than Tycho-2 for the very bright stars). Tycho-2 has some deviations from the UBV system even using Bessell tables. The bluest stars are redder, the problem is mostly with B but V is also brighter by 0.01 or 0.02 mag. Also yellow stars appear whiter in B-V by up to 0.05 or so in some cases, especially not so bright stars. That is a result of years of checkings of a wider variety stars for old sequence work.
A couple of examples:
tau Sco V=2.825, B-V= -0.259 (GCPD); V= 2.831 (HIP), B-V= -0.206 (HIP from T-1); V= 2.789, B-V= -0.198 (Tycho-2)
sig Cet V= 4.747; B-V= 0.449 (GCPD); V= 4.747 (HIP), B-V= 0.454 (HIP from UBV), V= 4.740; B-V= 0.435 (Tycho-2)
Cheers,
Sebastian
[quote=Sebastian Otero]
sig Cet V= 4.747; B-V= 0.449 (GCPD); V= 4.747 (HIP), B-V= 0.454 (HIP from UBV), V= 4.740; B-V= 0.435 (Tycho-2)
[/quote]
Hi Sebastian, Maybe this example is not the best? Only 7 millimag difference between the V's! We need to remember that even "very constant" main sequence stars like the Sun exhibit short and long term variations. These are due to the solar cycle, transits of large sunspot groups and flaculae, flares, etc. Ellipsoidal stars vary more. Possibility of large planetary transits over the stars, variations in interstellar absorption, etc. Sometimes, by chance, these variations could potentially add up in the same direction too. They can be 4, 5 or more millimags! Thus, the "random errors" we see in photometry of "constant" stars can very well have a significant systematic component, caused in part by real life ;)
See ACRIM results - http://acrim.com/Acrim1%20Results.htm
Mike
Hi Mike,
> Maybe this example is not the best?
Yes, probably, but I chose a random white-yellowish star and the difference is in the sense I've seen in lots of similar stars so I wanted to show that the colors are not exactly matching B-V. It is more pronounced for 7th or 8th mag. stars when the difference may be up to 0.05 mag. or so.
And Tycho-2 is always brighter in V on its good range, not fainter, so the effect is systematic, something that does not happen with HIPPARCOS, when the errors are random and very small.
Also, HIPPARCOS and Tycho data are contemporaneous so the mean magnitude should be the same.
Also, it is enough to see the raw light curve of Tycho data to see how large the scatter is when we go to fainter magnitudes and how large amplitude red variables like miras are much bluer most of the time due to the fact that the Bt-Vt is derived from mean values of all the observations independently and the star is not detected in Bt when it is faint. This leaves us with lots of miras with B-V close to 1.0 or bluer, something which is not real. I've read papers claiming the existence of blue companions to these miras when it is only a problem related with the Tycho values and UBV photometry gives the real red colors.
Cheers,
Sebastian
Hi Arne,
you said "...you need to place the revised values in the "comments" field for your submission."
I do ensemble DSLR photometry, then in the "Comp Label" field I insert "ensemble". I can't indicate all stars that I have used in the ensemble. However a researcher don't knows which stars I have used. The "comments" field is not very convenient for reporting all informations needed.
Regards,
Luigi