Affiliation
American Association of Variable Star Observers (AAVSO)
Tue, 03/05/2013 - 16:48

With a clear filter we can usually get better S/Ns than with a V filter and it seems to be recognized that reporting CV measures (clear using a V comp) is useful. It seems to me that computing a CVI, say, transformation can certainly be done (my spreadsheet wouldn't know I was lying to it!) and, presumably, we could report CV measurements but claim they are transformed. BUT, would that be 'better?' Some of my recent targets are quite red (V-I ~ 3)(eg, V2492 Cyg) and the V images need much longer exposures than the I images to get reasonable S/Ns. Getting CV images at the same S/N would save considerable time (for more targets).

When I compute transformation coefficients I transform the instrumental values of the catalog stars to estimate how well it is working. For example, in my BVI values I find the average correction for 20 comp stars is 0.012(0.027) in B, 0.004(0.010) in V and 0.002(0.004) in Ic. From this I take it that my V and I filters are pretty close to "standard" and the B isn't too far off. Next chance I get (cloudy this time of year) I will re-run my M67 images using a clear filter in place of V and see what happens.

Comments?

Jim Roe [ROE]

CV is not V

Hi Jim,

If you do that test you'll find that the stars with red colors are far brighter than they should be if you use V-band values for the comparison star magnitudes.  This is because most of the light coming from the star is at wavelengths where CCDs are very sensitive.  A clear magnitude of a red star won't give you meaningful physical information.  It would make more sense to compare it to an R or I comparison star magnitude, but then you're already getting that information (with better constraint) by taking an I-filter observation.  Signal to noise isn't the only consideration -- you can use a filter and give me a less precise measure of a well-defined bandpass, or you can run without and give me a precise measure of something whose spectral properties aren't obvious.  For some kinds of astrophysics it's better to have the former rather than the latter.

Again, this applies to red objects like your extreme case of (V-I)=3.  For something that's intrinsically blue, you still have the issue of an ill-defined spectral response, but not to the same degree.

Matt

Affiliation
American Association of Variable Star Observers (AAVSO)
CV observations

Matt's posting is correct; I just wanted to add a couple of points.  Red star spectral energy distributions are dramatically different over the unfiltered range of your CCD sensor, with lots of molecular bands and a rising continuum as you head towards the red.  What you should do is exactly what you planned: take some M67 images through the clear filter and see how well you transform.

Secondly, note that your "standard" stars don't go nearly as red as your target star.  That means how the transformation behaves for really red stars is unknown - most systems will show strong nonlinearities for these red stars, so you cannot assume that a linear transformation will get you anywhere close to reality.

Finally, if you do submit some "transformed" unfiltered magnitudes of red stars, use the LCG on these stars to see how well your data matches those observers with real V filters.  A couple of good stars this time of year for comparison are S Ori and R Leo.

Unfiltered observations work fine for most cataclysmic variables; they are strongly advised against for red stars.  Signal/noise is not everything!

Arne