Affiliation
American Association of Variable Star Observers (AAVSO)
Thu, 04/24/2014 - 20:10
I am geeting what seems to be a large deviation in Color Index Coefficient (Tbv) for my internal filter wheel in my SBIG 402ME: Tbv = 0.752. My Tvi seems OK (Tvi = 1.024). So, I am wondering if I have a bumb B filter. Anyone else run into this kind of Tbv with these filters?
Thanks, Ed
Too bad we can't edit: Geeting = getting
Ed
Egg on my face, I reversed the axes! Sorry for the string of posts.
Tbv = 1.2371 on the last run. (multiple comps, pretty good spread of colors; good SNR by stacking)
Using Jim Roe's spreadsheet and single comp (different target stars, again good SNR). Thanks for sharing Jim!
Tbv = 1.340
I don't know enough about transform coefficients to know how to handle such matters as selecting the comps (one or many). I do know that picking different colored targets is critical. It might be very helpful if newbies like me had access to an AAVSO-approved "do these stars for comps and targets for M 67 and these for ..." Or perhaps that is simply not possible? If possible it might produce some uniformity and confidence for us newbie transformers.
Thanks, Ed
Ed,
Just curious as to what you calculated for Tv?
Brad Walter
[quote=WBY]
Ed,
Just curious as to what you calculated for Tv?
Brad Walter
[/quote]
Hi Brad:
Using Jim's calculator:
Tv =
-0.039
Tvi =
0.966
Tbv =
1.340
Tv in this calculation is taken V-v vs V-I
V-v vs B-V: Tv is -0.064
B-b vs B-V: Tb = 0.811
Thanks for answering and any additional comments,
Ed
My question was just a matter of curiosity. Tv is normally quite small and your's is. I am recalculating mine. I had a lot of scatter in the data and I am redoing the photometry of NGC 7790 with a much smaller aperture (about 1/2 of what i would normally use) which resulted from Arne's suggestion in another thread.
Then I have to calculate them again, (using M67 this time) because I added a corrector to my Dahl Kirkham. It essentially eliminates the distortion I was getting away from the FOV center after upgrading to an STXL 6303 camera from my faithful old but small ST7E, which is now functioning as a guider.
First. my Tb is certainly not Tb = 0.811. That was a mistake (the R-squared not the slope).
Tb = 0.189 determined on B-b versus B-V
Question: Is it legit to perform B-b versus B-I? If so, I get a much "better" result although the R-squared is still high so there is scatter in the b measures.
Tb = 0.098
Ed,
On my SBIG ST402me camera with the SBIG internal BVIc filter wheel I had the following transformation coefficients the last time I measured:
Tv -0.065
Tbv 1.210
Tvi .041
I don't know if that's helpful or not!
Frank (SFRA)
It is helpful, Frank. I see that your Tvb is high, not as high as mine but certainly higher than 1.0. If this was on M67, I would be interested to know how you went about taking the instrument magnitudes.
I guess I am wondering: Is Tvb = 1.3 and Tb = 0.189 considered deviant? Of is that what one expects?
Thanks, Ed
Here is another set of ST402me transformation coefficients for comparison:
Tv = -0.072
Tbv = 1.160
Tvi = 0.939
This is with refractive optics. For reflecting telescopes, our local mirror coater recommends the standard, not enhanced, coating for photometry because of its relatively uniform reflectivity at the wavelengths of interest. His enhanced coatings are more reflective in the middle of the visible spectrum, at the expense of the UV and IR ends of the spectrum. Was the high value of Tbv measured with enhanced mirror coatings by any chance?
John Centala
Hi Ed,
I assume you mean Tbv, not Tvb. Usually, magnitude coefficients (like Tv, Tb, Ti, etc.) should be zero - that is, if your filter was exactly the same as the standard bandpass, then there would be no systematic variation with color of your measured objects. The farther from zero that you get, the poorer the match between your system and the standard system. The coefficient can be positive or negative, as your filter can be redder or bluer than the standard bandpass. Usually values in the range -0.1 < Tx < 0.1 are considered excellent; values of say -0.3 < Tx < 0.3 are usable, perhaps requiring a quadratic fit; anything greater than this can be used, but is suspicious.
For the color index coefficient, such as Tbv or Tvi, you are fitting a color index vs. a color index, so the slope/coefficient should be near 1.0. Again, ranges of 0.9 to 1.1 are great; 0.7 to 1.3 are probably acceptable with care.
The most deviant ones for amateur use are Tb and Tbv, because they involve the B filter. Most amateur CCDs have poor blue response, and so cut off the bluest part of the bandpass, pushing the filter response redwards. This results in Tb becoming more positive (say, 0.3) and Tbv more positive (say, 1.3). With a thinned, backside illuminated CCD, you can get Tb close to zero, and Tu in the 0.2-0.3 range.
Most ST-402 cameras do a pretty good job of matching the standard system.
Arne
Yes of course I mean Tbv not TVb. Fumbling fingers trike again. Even with the bad scatter in my data Tv was well within the range of 0.1 >Tv>-0.1. I just don't trust it with the amount of variance in the data so I am doing it again with the NGC 7790 info based on your comments in another thread.
For my Astrodon VBRcIc filters, the first pass gave me.
0.005
Tv
1.006745
Tbv
1.05031
Tvr
1.019056
Tvi
0.987654
Tri
Brad Walter
Well I realized I had a math mistake in calculating the above transformation coeffcients. The final values comprised of the averages of three nights observation of NGC 7790 This winter are:
Tv
0.0185
Tbv
1.0216
Tvr
1.0951
Tvi
1.0586
Tri
1.0188
I will post the methodology I used in the thread "Transformation Coefficients, NGC 7790 & Steve Howell's Growth Curve Method."
Brad Walter
Hi Ed,
It is "legit" to perform a fit vs. any color; what you are looking for is a systematic trend that can be modeled with a linear or higher-order least squares equation. However, I recommend that you don't use a color index that spans a wide range in wavelengths, especially when the filter that you are fitting lies at one end or the other of the index. The other reason I don't recommend (B-Ic) is just that it is an uncommon index, and you will get more help/assistance if you use something that everyone knows and understands.
APASS, for example, uses two CCD/OTA combinations, one for B,g',Zs,Y and one for V,r',i'. If, for some reason, one of the cameras is not working, you can use the filter combinations available for a single camera to transform/calibrate the data. We have a number of such nights, and may go back and add these nights into the final solution if their transformation works with nonstandard indices.
Arne
Many thanks Arne. I still have much to learn about these calculations and your comments help a great deal.
Ed
Arne, I assume the v-i index is OK to use even though it spans multiple filter passbands. Can you offer any suggestions when v-i is preferable to v-r or v-r preferable to v-i? I suppose if you were interested in changes in the hydrogen alpha line you would want to look at v-r rather than v-i. Aslo if your setup is limited to BVI filters the choice id obvious.
Are there other situations, disregarding limitations of specific equipment setups, that favors one index vs the other?