Affiliation
American Association of Variable Star Observers (AAVSO)
Tue, 08/19/2014 - 23:36
I did a 30 second Johnson V filtered image of V339 DEL with the Sierra Stars 24" last night. Using VPHOT with a six star ensemble and the 126 comp as the check star, I got a magnitude of 12.911 V versus 12.649 V for the 126 comp. I subsequently deleted this star as my check star. Is this star variable? Has anyone else noticed any magnitude divergence of this comp over time? With V339 DEL currently around 12.5 V, this could be an issue for recent, current and future photometry.
Regards, Kevin Paxson - PKV
Hi Kevin,
I took a V band image of V339 Del last night as well, on iT21 at NM Skies. I went back and chose 126 as the check star, and using 124 as the comp I got a magnitude of V = 12.638 for 126.
Not sure what's up with your image, but my quick check is not consistent with what you found. I'm using VPhot for analysis, you might try repeating my experiment (124 as single comp) using VPhot and see what you get?
Let me know what you find...
Brian
Hi Brian/Kevin,
The 124 and 126 comp stars both come from TMO61. Since Kevin indicates the 126 is off, then the 124 is probably off by the same amount, so Brian's test doesn't work.
Let me look at that calibration and see if/why it is off. It may take me a couple of days, as I am traveling back from Flagstaff tomorrow and Thursday is full of meetings.
Arne
Hi Arne,
I have been observing V339 Del for131 nights from end of March 2014 until now. I have been using stars 117 and 120 for comp and check stars, respectively.
My magnitude determination seem to be very consistent and there is a nice diminishing trend of the nova light.
Should I change the comp stars due to the fact the nova is less bright than several months ago.
I also observe an increase in brightness for the past several nights. Could that be a real increase in brightness in V not seen in I band light?
I will keep observing the nova until season ends.
Regards,
Josch
I will take a couple of images tonight if the weather cooperates. I have yet to look at my recent images to investigate this. The light curve looks good with many contributors. However for the first time, the 126 star was a big outlier in my ensemble. I only take an image every few weeks. Kevin
FWIW on 7/27 I ran a test of 126 using 129 as the comp and got good agreement with the published value for 126. However as Arne points out, 129 is also from Tortugas Mountain. In fact every star in that series except 80, 98 and 105 is from Tortugas Mountain. These three are from BSM.
So just for kicks I measured 126 (TM) in about 60 images using 105 (BSM) as the comp. I came up with a value of 12.696 +/- 0.01 which is probably close enough considering the stated errors of 126 and 105.
Jim Jones, JJI
I concur Jim. I did a similar test using data from 2014-06-04 and 2014-06-07. Nothing peculiar about 126.
Michael
[quote=jji]
FWIW on 7/27 I ran a test of 126 using 129 as the comp and got good agreement with the published value for 126. However as Arne points out, 129 is also from Tortugas Mountain. In fact every star in that series except 80, 98 and 105 is from Tortugas Mountain. These three are from BSM.
So just for kicks I measured 126 (TM) in about 60 images using 105 (BSM) as the comp. I came up with a value of 12.696 +/- 0.01 which is probably close enough considering the stated errors of 126 and 105.
Jim Jones, JJI
[/quote]
I say again, examine your images of that star at VPHOT's highest magnification (8x). If any of them look odd, be suspicious. Examine againat low zoom. Adjust max and min in the display so that with a 10 to 15 point spread between maximum and minimum, the image has a "salt and pepper" look to it. The way photometry software repairs images is to either duplicate or average the adjacent columns. Such stars are not suitable for photometry. This is a topic fresh in my mind; so fresh that the egg on my face hasn't dried completely.
Here (I hope) is how it should look with a zoom factor of 1:
Examine your image carefully with a small difference (say 10 to 15) in VPHOT Image display values so that the sky gives a "salt and pepper" appearance. If there is a "zipper" look to it, then there were dead columns there. Don't use any stars for measurement that had dead columns in the image of that star..
Hi Kevin,
I'm back in my office, and so was able to do some checking. Here is a comparison of the photometry table results from VSD/VSP, which use BSM for the 105 and 109 star and TMO61 for the 124, 126 and 129 stars, along with two nights from Coker30 as a cross-check:
Star VSD coker30
105 10.525 10.503
109 10.893 10.851
124 12.417 12.420
126 12.649 12.678
129 12.885 12.895
So I don't see any discrepancy in the VSD/VSP photometry table. I'm wondering if you accidentally measured the 129 star instead of the 126 star.
Arne
I took another 30 second Johnson V image with the SSON 24" last night. I re-looked at my images from last night and two nights ago, which were taken with the same telescope and exposure length. I looked at both images in VPHOT with the same sequences and measurement annuli. The 126 comp was with "normal bounds" last night, yet it was indeed nearly 0.3 mags fainter on the night before. I am lost for an explanation other than a possible asteroid occultation. A one night dip in brightness is puzzling. This issue is closed for now, but I will keep an eye on the this star. My mystery image file is available on request. Kevin Paxson - PKV
Hi Kevin,
That makes the 126 Comp sound like an eclipsing binary. I never upgrade a comparison star to a true variable until I have confirming evidence, preferably on a different night or with a different telescope. We have about 70 nights of data of the field using coker30, so the chances would be good that we would have a similar eclipse recorded. You also indicate, I think, that your faint observation was on UTD 2014-08-19, and we have a dataset from that night centered on 07:15UT where we could compare, if you let me know when your image was taken. Or, there may be another observer with similar datasets.
Arne
Arne, the JD date of the 126 comp star being at 12.911 V was 2456888.77309. Regards, Kevin Paxson - PKV
Hi Kevin,
Unfortunately, the BVRI set that matches your JD the closest had trailing, so I haven't processed it. Here are the individual values from Coker30 from the night before, the night in question, and the night after, using an ensemble:
HJD Vmag
56887.8539 12.655 (the other two measures of this night were trailed)
56888.8077 12.654
56888.8126 12.654
56889.8527 12.628 (trailed)
56889.8576 12.653
56889.8625 12.654
Note that 56888.8077 HJD == 2456888.8014 JD, so about 40 minutes after your measure. I have an earlier dataset at 245688.7965 JD or 33 minutes after your measure; though trailed enough that I didn't process it, the star looks the same brightness as the subsequent set.
So if this was a real event, it had to return to its normal brightness, exactly, within 40 minutes. I would be happy to look at your image and see if I can find any instrumental problem.
Arne
After looking at Kevin's image, it is obvious that the situation described by Lew Cook exists, where a bad column or two on the sensor has been interpolated across by the upstream image processing system to "fill in" on the final image. Those bad columns fell exactly across the 126 comp, and therefore lowered the total flux reported in the image.
I will sometimes interpolate across a bad column, or a hot pixel, if that process only affects the sky background or at most something in the wings of a star profile. Since you don't know ahead of time, the best practice is to leave cosmetic defects in your images as you process them. Then, at the photometric analysis phase, inspect the images to see if any problem pixels fall within the measurement aperture. At that time you can make the judgement as to how to handle them. If you "fix" the pixels before you analyze the image, it is often hard to tell that one pixel only has an estimate for its recorded flux. All images coming from AAVSOnet leave all cosmetic defects alone; the researcher is expected to make the judgement call.
Another way that I have done this in the past is to create a bad pixel mask. This is a mask where good pixels have a value of "0" and a bad pixel has a value of "1". You can use this mask as an overlay, assigning a color to "1" so that the overlay is easily visible on an image, or where the mask can be used in post-processing to either flag stars as being contaminated and therefore rejected, or to interpolate across if you feel comfortable. IRAF of course has tools for creating and using such masks because science sensors tend to have many more defects than what you can buy commercially today.
Bottom line: the 126 comp is just fine, stable to the best of my ability to measure. The mix of comparison stars from BSM and TMO61 yield approximately the same scale from bright to faint.
Arne