KIC 8462852- Minor difference between B and V filter

American Association of Variable Star Observers (AAVSO)
Sun, 11/08/2015 - 23:35

After taking filtered measurements of this star I found that my V band measurements were in line with everyone else

but my B band measurements are only 3 hundredths lower than V Band. How can this be? All the other stars in the image

show a decrease for the most part from 4-5 tenths of a magnitude and my darks and B and V flats were done correctly. I'm

using Maxim DL the same as usual.....

I'm perplexed....

American Association of Variable Star Observers (AAVSO)
KIC 8462852- Minor difference between B and V filter

Some perhaps too obvious questions might be:

1. Is this for sure a photometric B filter or one that was thought to be photometric but is actually one made for LRGB astrophotography?   Perhaps you acquired the filter from someone who said it was photometric but is not?

2. Is the filter scratched even slightly or someone agressively cleaned it?  Closely inspect it by looking through it with very bright light.

Now having asked some obvious things, one issue I just ran into with my old SBIG CFW-8 filter wheel that it stopped indexing to the correct filter.  Coincedentally I was doing V and I photometry on the very same KIC 8462852 and noticed at dawn that my sky flats looked a little different and the comp and check star mags were out of whack.  A whole nite was wasted on that star and one CV.  Something told me I should look into the filter wheel on power-up and while selecting filters using MaximDL.  Though it sounded like the filter wheel was indexing OK and MaximDL was not complaining of any errors, the wrong filter was being selected!!  And during power-up it would occasionally straddle two filters (thus the different flats).  I did all the recommended cleaning of the sensor the O-ring and motor tension etc. but no luck.  Long story short; just today I sent the filter wheel back to SBIG for repairs.  Since ultimately it is the camera that passes commands the filter wheel, I'm hoping it doesn't actually wind up being the ST8-XME camera that needs repairs.  

I'll try to post here what the final verdict was.  

The main lesson I'd like to pass along is look down the bore of the filter wheel visually every six months or so and perhaps as preventative maintenance clean the sensor once a year with isopropil alchol and inspect filters for fogging or buildup of corroision or oxidization (esp. in humid environments).   Also check the camera's linearity once a year too and visually inspect flats for changes and dark and bias frames as well.  It may all sound like unnecessary checks but I've learned otherwise.  Good science depends on this sort of periodic close scrutiny.


American Association of Variable Star Observers (AAVSO)
KIC 8462852- Minor difference between B and V filter

The verdict here in my case was a bad sensor on the CFW-8 filter wheel.   I sent it back to SBIG for repairs.  Note: if you have to do this be sure to pack the carousel very very well.  Especially protect the center pin on both sides.  I thought I had but as a result of rough handling, both center pins on both carousels were bent when they arrived at SBIG so I had to pay $75 for a replacement (down to one carousel now).



American Association of Variable Star Observers (AAVSO)
Filter heights

Another potential issue to look out for, is that on my CFW-10A SBIG filter wheel, there is very little clearance between the threaded inserts holding the filters and the case.  I found that only a few filters are a little longer, and a few do not thread in all the way.  Not sure if its the carosel threads or the filter.  Changing positions of the filters did provide a bandaid.  This suggests a little of both.  This cause the wheel to hang up occasionally.

Not sure the CFW-8 has the same minimal clearance.


B filter issue

As can be seen from the light curve generator, the V band values should be around V=11.8, and the B band values should be around B=12.3.  If you are measuring B only a few hundredths fainter than V, then there is likely an error somewhere.

The first thing to do is to look at the magnitude difference between the various comparison stars, on both your V image and your B image.  For example, the 124 comparison (000-BLS-549) and the 128 comparison (000-BLS-555) are a good pair to work with.  Their standard magnitude differences are:

B: 13.270 - 13.230 = -0.040

V: 12.789 - 12.427 = 0.362

So compare these differences with the differences you obtain from your own images.  If, for example, you get nearly 0.362 as the difference for both your "B" image and your "V" image, then most likely your filter wheel did not move, and both images were taken through the V filter.  If you instead get nearly -0.040 as the measured difference for these two comp stars on either the "B" or the "V" image, then the filter wheel did not move, and both images were taken through the B filter.  This is an easy way to test the filter wheel (and a way that I also test an unknown image to see what filter was used).

Let's assume that you got a comparison star difference of, say, 0.05 for the B image and 0.41 for the V image.  Then you can be pretty much convinced that the images were taken through the correct filters, and you have to look elsewhere for the problem.

The first and most likely problem is that you used the V magnitude for the comparison stars for both the B image and the V image, rather than B magnitudes for the B image and V magnitudes for the V image.  This is an easy mistake to make if you are "rolling your own", using say a spreadsheet or calculator to calculate the standardized photometry.  Usually, using something like VPHOT avoids this problem, as long as the filter is correctly listed in the image FITS header (you should check this as well).

Some other possibiliites:  your filters may not be standard B and V (I've seen mismarked filters before).  This is easy to detect by visually looking through them, as obviously the B filter will be blue and the V filter green.  Don't count on the reflected color, as interference filters reflect light differently than colored glass filters.  You could also have a very large transformation coefficient for the B filter, so that your untransformed values are quite wrong (though this amount of error is very rare).  You could have entered the photometry values of the comparison stars incorrectly in your calculation.  Your B filter could have a very large red leak (also uncommon these days).

The bottom line, when you are very discrepant from others, is to start with the most basic tests and then move to more and more exotic ones.  If you can't find a solution, then it may mean that the data you are comparing against is wrong and you are right, so don't just discard discrepant data.  (However, in this case, you are comparing against many other observers, not just one, and so you are likely to be the wrong data!).