Hello! Stan is looking at BH CRU values. My B magnitudes are brighter by up to 1/2 magnitudes than another observer, and we are trying to sort through the discrepancy.
As far as my equipment, I use an 8 inch LX200 classic with 0.63 focal reducer, an SBIG ST402 with BVIC filters, and I operate my runs with MPO Connections. Most of my images are taken at 0 degrees C. I take typically take two images and average them after calibration using AIP4WIN, unless one is significantly trailed or out of focus for some reason.
B images are typically at 60 seconds and V images shorter to avoid saturation and to stay in the linar portion of the CCD response curve. The two uploaded images have V=40 second exposures. I use a two stage lightbox (two white plastic opal plates separate the light sources from the telescope opening and each plate is separted by about 10 inches) to obtain flats; the light sources are white LEDs and one incandescant bulb. Flat darks and darks at the same temperature as the images are applied.
The imamages are calibrated in MPO Canopus and analyzed with MPO PhotoRed using ensemble photometry, transformed, and uploaed to AAVSO. I calculate my transformation coefficients about once a year using a number of Landolt fields and calculating the transforms with MPO PhotoRed.
With my FOV and with BH CRU centered, there are three comps: 000-BJK-993 as the check star (B-V = 1.2) with comps 000-BJK-994 (B-V = 0.2) and 000-BJK-995 (B-V = 0.6).
Might the difference be with the B-V for the check star and comps? It looks like BH CRU has a B-V of 3 to 3.5? Additionally, no error ranges are listed for these comps. Might a high error/uncertainty range account for some of the discrepancy?
I'll offset my runs in the future. I should be able to get comps 000-BJK-991 (B-V = 1.4) and 000-BJK-992 (B-V = 0.7) in the frame as well.
I've uploaded two runs - 19-Feb and 6-March. I can provide the raw images and calibrated images as well. Thank you for any suggestions. Best regards.
Mike
Hi,
what kind of B filter you are using? Maybe it has a infrared leak and the star in qeustion (BH Cru) is reddish?
I have had a similar problem with a Astrodon B filter some time ago which had this leak and was exhcanged by Astrodon afterwards.
Regards,
Josch (HMB)
Hi Mike,
Like Josch, I am suspicious of a red leak in the blue filter with this very red object.
Astrodon created several different Johnson B filters, each one having more IR blocking to decrease a couple of red leaks in the dielectric filter. There is a really small leak around 700nm, and then a very bad one at 1.1 micron. If you are using Astrodon filters, look at the label. The most current one says B***. The number of asterisks indicate the "generation" of the filter. Let the forum know what you find.
If you are not using Astrodon filters, then we can go to the next level.
Arne
Hi Arne and others,
The problem is not just this one star but two others - R Cen and RU Sco. The B-V colour of BH Cru varies between ~2.8 and ~4.6, R Cen is reasonably steady at ~2.0 and I know nothing about RU Sco except that Sara Beck found an apparent discrepancy there.
I've been trying to do comparison of colours of R Cen during its changing period using the Auckland Observatory (APOG) measures, those of Giorgio di Scala and Neil Butterworth who uses DSLR very well. There is no real problem there except for the probability that with that degree of B-V extrapolation who can bet on the final result to more than a tenth or two.
The other two values determined for each star, V and I, are good and Mike is not using R. I assumed that with improved techniques the filters were now very reliable - I even said this to Sara - and that red leaks were a thing of the past. Luckily the EMI tubes we used then had a sharp cutoff to zero at 6000A.
Apart from some of Mike's early measures the offset is fairly constant so far. But the B-V colour of BH Cru changes quite a lot and it's close to rising at present. There are several months yet before it gets close to the Sun. The offset seem to be related a bit to colour with BH at 0.7, R at 0.5 and RU Sco too few to determine.
The filter idea hopefully will resolve the problem.
I'd compared the sequence values with the Tycho ones in Guide 9 and realised that they were poor but that isn't at this stage the real problem. If they are photographic then better are clearly needed. Hopefully all this will be resolved as Mike's measures of these interesting stars are extremely valuable. I was hoping that R Cen would have become slightly bluer as the period shortens - there are some changes in shape which may be real but UBV photometry has an inverted U-B curve which suggests an unseen companion. But 0.5 was too much!
So that's the whole background to the problem. Mike, thanks for calling in the experts.
Regards, Stan
Hi Arne,
I've been giving some more thought to this. Crudely, the intensity in B is around twice what it should be. This is hardly a leak but a torrent.
Mike sent me his transformation values - about 6% in V, 5% in B, or maybe the other way round. But if the colour difference is as I suggest then the b-v/B-V plot would show a much greater correction needed. Certainly BH Cru is extremely red at minimum but R Cen at 2.0 and RU Sco a little less are not. And a red leak would be largely overwhelmed by the strength of hotter stars in the B filter, which stars presumably are used in the calibration.
Canopus is used for the reductions. I know a couple of observers who have abandoned it but for no clear reason that I'm aware of - it seems to do many things other than reduce data. But is there anything in the data entry stage which could cause such errors?
Sara suggested a faulty filter but looking at the values of the BH Cru comparisons (in Guide) and the probable R Cen ones they should all be affected fairly much the same by any leak. So can we ask the people who are familiar with this software for suggestions?
Regards, Stan
Hi Mike,
I should also make a few comments about the comparison stars for this field. The three stars you are using (993,994,995) come from Mati Morel's great sequence catalog, but are based on photographic magnitudes and haven't been revised to the more accurate photometry from APASS. You should submit a sequence revision request to the Sequence Team (from the home page, go to Observing->Variable Star Charts->Report Chart Errors). In particular, the Bmags are in error, but in the wrong direction - the photometry table values are too faint by a couple of tenths, which would make your Bmags too faint rather than too bright, so your problem is even worse than you expect!
Second, star 113 (000-BJK-993) that you are using as a check star is a known variable from the ASAS catalog:
ASAS J121601-5614.0
which has an amplitude of about 0.1mag. Luckily, as a check star, it does not influence your target star photometry, but you should probably choose something different. As an LPV of interest, I think BH Cru is a good candidate for sequence revision.
Arne
Hi Arne
I also have used 113 (000-BJK-993) as a reference star in my observations of BH. If it is a known variable, why is it still in the AAVSO list of possible comparison stars?
Robert
HI Arne
When I started using 000-BJK-993 for BH Cru I had a look in the ASAS database and could not see it as a variable. I just had a look in the ASAS database. 993 is 12:16:01.12 -56:14:07.1 . I can see no variable around that point. Am I looking in the wrong database when I am looking to confirm that stars are not variable??
ID ↓
RA
is 12:16:01.12(2000)
DEC
(2000)
Period
[days]
T0
V
[mag]
V Amp
[mag]
Class
Other
ID
Other
Class
120553-5629.3
12:05:53
-56:29:18
1.228652
1874.49
11.24
0.21
EC=ESD
-
-
120758-5600.7
12:07:58
-56:00:42
1.049792
1873.81
10.73
0.47
ESD
-
-
120857-5524.5
12:08:57
-55:24:30
211.2676
2447.5
12.51
0.37
MISC
-
-
121148-5710.9
12:11:48
-57:10:54
2.766366
1880.97
12.75
0.5
ED
-
-
121301-5558.5
12:13:01
-55:58:30
0.382098
1873.09
11.21
0.24
EC/DSCTr
-
-
121322-5617.8
12:13:22
-56:17:48
139.5348
2230
12.28
0.45
MISC/SR
-
-
121411-5456.0
12:14:11
-54:56:00
5.31321
1882.4
11.93
0.46
DCEP-FO
-
-
121418-5628.8
12:14:18
-56:28:48
225.9886
2417.5
12.38
2.13
MIRA
-
-
121500-5449.2
12:15:00
-54:49:12
70
2054.8
8.36
0.33
MISC/SR
V0369~Cen
SRB
121617-5617.2
12:16:17
-56:17:12
518.1422
3292.5
6.9
2.28
MIRA
BH~Cru
M
121653-5501.5
12:16:53
-55:01:30
55.71030
2031.1
10.63
0.37
MISC/SR
-
-
121705-5743.9
12:17:05
-57:43:54
7.551478
1898.4
10.85
0.31
DCEP-FU:
-
-
121742-5637.2
12:17:42
-56:37:12
32.46753
1928.7
9.5
0.17
MISC/SR
-
-
121816-5718.8
12:18:16
-57:18:48
602.0234
2402.6
11.17
0.47
MISC
-
-
121845-5613.8
12:18:45
-56:13:48
142.2310
2219.9
11.69
1.64
MISC/SR
-
-
122003-5607.7
12:20:03
-56:07:42
0.510128
1873.02
12.11
0.47
EC
-
-
122112-5642.7
12:21:12
-56:42:42
1.364502
1872.84
11.69
0.6
ED
-
-
122118-5641.1
12:21:18
-56:41:06
85.71428
2167.7
9.34
0.47
MISC/SR
-
-
122244-5637.5
12:22:44
-56:37:30
52.17391
2037.7
10.98
0.23
MISC/SR
-
-
122349-5528.2
12:23:49
-55:28:12
123.1153
2263.9
13.58
2.15
MIRA
-
-
122352-5514.0
12:23:52
-55:14:00
1.009359
1873.4
10.55
0.15
ESD
-
-
122522-5639.8
12:25:22
-56:39:48
59.16859
2076.4
11.76
0.4
MISC/SR
-
-
Robert,
Yes, you are looking at the wrong database, look at our VSX, there you will find many more variables than if you only look at a specific catalogue.
This star is not in the ASAS catalogue because it was published by M. Kiraga in 2012. It is not part of the original ASAS list.
It is a BY Draconis variable and its total amplitude is 0.3 mag. (there are mean magnitude changes over the years) so it shouldn't be used. It has now been removed as a comparison star for BH Cru.
Cheers,
Sebastian
Thanks Arne and Sebastian
I have used the ASAs and VSX catalogues to check comparison stars before but did only check the ASAS for BH Cru - being slack I had assumed that the Comparison stars chart would be OK. Luckily there are several good comparison stars in the BH vicinity to replace 993.
Thanks again to you both. Now let us hope we can resolve Mike's issues with his Blue filters.
Regards
Robert
Arne and Sebastian
One last thing on these two from me for now; I am removing all of my observations from the AAVSO data base on BH Cru as they were submitted using 993.
I have the images stored and will reanalyze them and resubmit the data. I wil also submit data on 993 as it is obviously in the same field.
Robert
Hi Arne and Sebastian
I have resubmitted my BH Cru data and the V and B values are a much better fit with the other data now I am using 000-BJK-988 instead of 933 as my comparison star.
I have also done some data mining on my BH Cru images and dragged out some 000-BJK-993 observations. They need a lot more work.
Thanks for your help.
Robert
Hello! The filters are the BVIC filters from Custom Scientific that SBIG has made specifically for the ST402 internal filter wheel.
I'll put in a sequence update request.
Thank you for the input to date. Best regards.
Mike
Hello! I forgot to add that my ST402 is over 6 years old, if that might come into play. Best regards.
Mike
Hi Mike,
Of course - I should have read your original posting more closely! My apologies. However, my guess is that the Custom Scientific filters also have a red leak. It doesn't take much of one to affect LPV photometry. There are several ways to test for this:
- do a full spectral scan of the filter, from ~400nm to ~1100nm. That will show the basic Johnson B response, plus any red leaks out to the response limit of the CCD. Dielectric filters typically have red leaks, and you add a special red-blocking layer to remove them. The problem is that you need ~10**(-6) blocking on B in order to get good results. The other problem is finding a lab that can scan your filters.
- find a field where there are very red calibrated stars. Typical examples are SA110 and SA111, both heavily reddened regions studied by Landolt. Image and plot, see if there are systematic deviations as you get very red.
- image a very red star near the horizon, say 10 degrees up. The atmosphere spreads the starlight into a miniature spectrum, so that the blue light can be many arcsec from the red light. A red leak acts like a second filter, so what you get is TWO images of the star, one at the correct location for the blue light, and one that is at the correct location for the red light. You can measure the angular distance between the two images and actually calculate the wavelength of the red leak by calculating the differential refraction.
The final test is to just look at the comp stars for a field. If you use one as the comp, and the others as targets, you can see whether they give correct results while the red variable is deviant.
Let me see if we have some clues in the images you posted. It may take me a couple of days; someone else might take a look in the interim (compare, say, (APASS B - instrumental B) vs. (B-V) for the stars in the field).
Arne
HI Mike
I Have taken a few observations of BH from Adelaide in South Australia of this very red star. Out of interest I have a 10" OTA and a QSI camera and because it is summer in Adelaide -10C. Using that I need a 360 sec exposure for the Johnson B; 10 Second exposure for the Johnson V; and having to mask the scope down heavily to even get a useable reading at 5 sec with Cousins I. I have not submitted any Ic observations because I am not satisfied with the results. This is a very red star.
I use 000-BJK-993 as the reference star and 000-BJK-990 as my check stars and my data is close to other observers. As a check I have used several other stars in ensemble and got the same results.
I agree with the other respondents that your filters may be letting IR light through.
Do you get the same issues with other stars?
Robert Jenkins
Hello! For my transmformation coefficients, I imaged several landolt fields at different altitudes, last done about 5 month ago. I used MPO Canopus/PhotoRed for the analysis. Here are my coefficients.
B = 0.046 and Zero Point = 19.402
V = -0.062 and ZP = 19.756
I = -0.001 and ZP = 18.695
We've ad some clouds and I've been traveling, so I hope to get out this week for imaging. I'll image SA-110 and SA-111. I'm not sure that I can get down to 10 degrees for a low horizon given my observing location, but I'll see what might be available. R LEP may be a good candidate.
Robert, R CEN has the same B mag difference, though V seems consistent with other observers. I don't see any other RU SCO observations in the database to compare it with. RT CRU has a wide variety of values from different observers, and mine aer in teh ballpark, but tend to be brighter than most. Many of my variables do not have other observers, such as R, S, V, WW PYX, Cl CAR, and TT CEN. PNV J17032620-3504140 seems consistent with other observers, though there are not a lot of observations to eyeball.
Perhaps Brian Warner can weigh in on MPO Canopus/PhotoRed. Best regards.
Mike
Hello! I had a brief reply from Mr. Warner about MPO Canopus/PhotoRed, which I thought I would include here for folks discussion. Best regards.
Mike
"Hi Michael,
If using the AAVSO procedures in PhotoRed,
For each selected compstar:
M_target - (m-comp-m_target) + M_comp
where upper case are "catalog" magnitudes and lower case are
instrumental magnitudes. The final target value is the mean of the
several M_target values. The error is the standard deviation of the mean.
The "check" star in the process is actually a "control target", i.e.,
its value is calculated the same way as the actual target. Since its
magnitude is supposedly known and non-variabble, this provides a means
of checking the veracity of the process.
The quality of the final results depends on the accuracy of the catalog
magnitudes - assuming that the comps are similar enough in color to the
target. If not, it is possible to apply pre-determined transforms to
each comp-target pair, thus reducing any errors based on color differences.
In my AAVSO work, I have used - whenever possible - comp stars in AAVSO
approved sequences for a given field. This produced generally very tight
data, i.e., a low standard deviation in the mean.
All this is, as I understand, the way the AAVSO expects ensemble
photometry to be done under the revised system introduced a few years
back with the new CCD format.
You can find more about this starting on about page 265 of the Canopus
manual.
A last thought: no attempt is made to compute errors by a more
complicated process of merging the individual errors of each
measurements, the transforms, catalog magnitudes, etc. since - in my
lowly opinion - all of these become merged in the variations about the
mean and reflected in the standard deviation of the mean. To date, that
presumption has served extremely well. As long as I'm clear about how
the error is calculated, then the end-user can judge the quality of the
data himself and use or not use the data, or assign a weight, as he sees
fit.
I hope this helps."
Hi Mike,
We have 4-6BV datasets each for RU Sco, R Pyx and S Pyx from BSM_South; I'll get those submitted this week. We have quite a bit more BV data for V Pyx, so that may be a better candidate for comparison.
Stan, (B-V) is just a basic indicator of redness for LPVs. (B-Ic) gives you a better handle on how red the star is, and for BH Cru, that value is about 8. Even there, Ic peaks about 850nm, and by the time you get to 1100nm (the sensitivity limit for silicon), the star is several times brighter. Having great red response in a sensor is both a blessing and a problem. With the original Astrodon B filter (no longer sold, and I'm pretty sure everyone has replaced it in their filter wheels), S Ori (which is not quite as red as BH Cru) was actually brighter through the B filter than through the V filter.
See Munari and Moretti:
2012BaltA..21...22M
for more details on characterizing red leaks. I am not 100% convinced that this is Mike's problem, but I've never tested the ST402 filters and it is a good first possibility for such offsets. It could be that everyone else is wrong and Mike is correct, for example!
Arne
Hello! My first test for red leak inmy B filter.
I followed R LEP for a couple of hours, from air mass of about 2 to over 5 at 10 degrees. The star did not appear to change in the multiple images in B and V. I presume that if there was a red leak that the star would appear to split as it descended fromabout 30 degrees to 10 degrees?
I'll check my Landolt fields tonight.
I imaged BH CRU. I'll check the magnitudds with the previous comps and compare those values when I elimiate the variable check star and again when I add in the additional comps I was able to get into the frame by off setting the image a bit from BH CRU. Best regards.
Mike
Hello! I finished my BH CRU analysis from last night.
There does not appear to be a red light leak. As I mentioned above, R LEP did not split in B and V as I followed it from 30 degrees to 10 degrees. I have not checked B-V for SA 110. I will do that later on.
Here are last night's values for BH CRU.
Using 993, 994, and 995 as I did before. B = 13.348, V = 9.393. B-V = 3.955. Berr=0.037 Verr=0.009
Using 994 and 995. B = 13.471. V = 9.394. B-V = 4.077. Berr = 0.026. Verr=0.007
Using the updated chart with 991, 992, 994, 995 with 995 as the check star
B = 13.399. V = 9.418. B-V = 3.981. Berr=0.156. Verr=0.026
Using the above 4 stars but 991 as the check star
B = 13.382. V = 9.401. B-V = 3.981. Berr= 0.137. Verr= 0.014
How do these values look? Best regards.
Mike
Hello! Still no luck tracking down the difference in magnitudes between myself and other observers.
I off-set last nights run. The first run used comps 992, 994, and 995. The second run used bright comps 987, 988, and 989.
First run: B= 13.437, V= 9.455 Berr= 0.127, Verr= 0.021, B-V= 3.982
Second run: B= 13.394, V= 9.426, Berr= 0.109, Verr= 0.020, B-V= 3.968
I would appreciate any suggestions for identifying the difference between myself and two other observers. Best regards.
Mike