I've uploaded the NGC 7790 standards field photometry from my calibration work at USNO-Flagstaff using the 1.0m telescope. There are something over 300 stars now available, ranging from 11<V<18, and with all UBVRI filters.
From 11<V<14, the data comes from short exposures of the field, designed to not saturate the three cepheids in the cluster. There are about 11 nights of data that were of proper quality. For 14<V<18, the photometry is from longer exposure images on 7 nights.
There are a few complications. First, there were a handful of existing comparison stars in the field. While their photometry was updated with this import, and they are of standards quality, they will not show up when you click "Standards" in the VSP input form. Right now, a star is either a comparison star, or a standard. We will have to resolve that in the next VSP release. However, if the star is listed with a sourceID of 10, it is of the same quality as the others in the field and can be used for determining your coefficients.
Second, there are two NSV stars in the cluster. NSV 11781 is a delta Scuti variable with small amplitude, but I've removed it from the standards file. Note that the current VSX coordinates are off by a couple of arcsec. NSV 26181 is perhaps a very low-amplitude red variable (also with poor VSX coordinates). I've left it in the standards file because it is also the reddest star in the field.
To access this standards field, enter the coordinates 23:58:23.2 +61:12:25 in the VSP form, and click "would you like a standard field chart?".
Enjoy!
Yes all of these are from VSP. If you look at the sequence table you will see that all have AUIDs and everything is in the same order as the sequence table that you download if you set the max magnitude to 14. I just cut and pasted so that all the photometry data was set up to do calculations.
As to printing the file, you had more success than I did. When I printed the PDF none of the numbers or callout lines showed up. PDFs can be strange about printing comments.It treats them as overlay layers and I can never remember what to do to "flatten" the file. Try the attached jpg. They are single layer files and this seems to print just fine. I suggest using about 160% zoom when printing.
Brad Walter, WBY
Hello Brad
Thanks for doing this. Its just what is needed. I assume that the sequence is a VSP Standard Field from AAVSO web site and are Arnes's latest numbers. Thanks for the "xls".
One thing that happens to me, not sure if others are seeing too. When I view the plot on the screen, I can see all 46 numbers. When I print it out, I only get a hand full (05, 02, 09, 23, 17, & 42). Is there something that I can do to get all the numbers? I'll try it from the Library today later. I can get a color print there. I will let you know if I get the same small set of numbers.
Gary
Here are my re-do transform coefficients using Arne's new NGC 7790 BVRI comp mag.'s and PTGP4.0
FROM PTGP V4.0 Program using Instrumental magnitudes:
30-Aug-14
6-Sep-14
Average
Tbv
1.394
1.380
1.387
Tb(BV)
0.308
0.294
0.301
Tv((BV)
0.025
0.019
0.022
Tvr
1.082
1.088
1.085
Tv(VR)
0.047
0.036
0.042
Tr(VR)
-0.028
-0.046
-0.037
Tri
0.984
0.990
0.987
Tr(RI)
-0.032
-0.051
-0.042
Ti(RI)
-0.016
-0.041
-0.029
Tvi
1.034
1.040
1.037
Tv(VI)
0.025
0.019
0.022
Ti(VI)
-0.007
-0.019
-0.013
You can download the spreadsheet I made to identify and reduce the comps stars for processing in AIP4WIN here:
http://www.astroimage.info/files/NGC7790_VSP-Chart%28BritTHAN14.6%20V%292.xls
Here is the .STARS file you can use in AIP4WIN of these 16 selected stars:
http://www.astroimage.info/files/NGC%207790_13617EUQ2.TXT (Save this text to a Notepad .txt file and rename the .TXT to .STAR before use). This file avoids inputting all the UBVRI data for the 16 comp stars I've show on the previous spreadsheet. After processing in AIP4WIN, you'll need to rename the comp stars to the IDs shown in the spreadsheet or PTGP4.0 will not be able process your data.
James
Hi James,
This looks much better, of course. You've gotten rid of a 0.3 offset on both Tb and Tv, so now your results fit well with others that have similar equipment. Tb and Tbv are still a bit farther off than I'd like - can you tell us details about your equipment?
I have two suggestions. First, according to your post, you've eliminated all stars redder than (B-V) = 1.36. There are a half-dozen stars that are redder than this, and they contribute greatly to the coefficient determination. You should include them if possible. Second, you are only using 16 stars. While you've carefully selected them, any specific star has uncertainty in its calibration, and using more helps to eliminate that source of error in the calculation. 300 is not essential, but doubling to 30 would be beneficial. Brad found 46 stars brighter than V=13.9, for example.
For everyone, one of the problems with trying to calculate transformation coefficients from a cluster is the wide range in magnitude, especially when star colors are considered. Including the very red stars means that they are going to be especially faint at B and V, while the blue stars are brightest there. Similarly, in the red, the blue stars will be faint and the red stars especially bright. Most CCDs don't have sufficient dynamic range to handle this. The best solution is to find the exposure in each filter where the brightest star you are using does not saturate, and then take many of these exposures and stack to get enough signal/noise in the fainter stars of the calibration set.
Arne
Arne,
To be truthful, I might measure all 46 but I am confident that I would end up discarding some as outliers because of crowding effects. I am also confident that I would end up with at least 30 or so good ones.
Please see my other post this morning on max apparent B-V limit when selecting stars for this cluster.
Brad Walter, WBY
To: Arne & Brad
I can include stars redder than 1.36. I wanted to pick as many as I could from Arne's list. However, PTGP's list of NGC 7790 are only 45 stars compared to Arne's list of 300 for the standards field option. Stars bluer than 0.55 seem to be under represented in this PTGP list. I also wonder if PTGP is still using the "photometry table on the binaries.boulder.swri.org site?"
Here are the stats on my system: 13.1" apeture (330mm) classical cassegrain with no refractive elements working at 2442mm focal length with an SBIG ST-10XME, water cooled at -16F below ambient (ambient being between 78-67 degrees F). The exposures for NGC 7790 are 180 seconds for each filter at 1x1 binning. My filters (1.25" diameter), for B & V, are Astrodon dichromatic photometric, and for R & I, Schuler Red and IR72 Schuler Is glass filters. I'm using MaximDL 5.23 for image acquisition and AIP4WIN for photometry reduction.
I'm glad my TC number are looking better. I wonder if these are "good enough" to apply to the XZ Cet data I've collected?
James
p.s. I can try to alter the exposures btw filters to get better SNR; my XZ Cet data use 60 second exposures in BVRI to avoid saturation.
The next version of PTGP will be available within a few weeks. It has major enhancements - the biggest being it now goes directly to VSP in real time and downloads the latest standard star values. Also, in addition to the PTGP and AiP4WIN format, it will directly take VPHOT download files - i.e. you can upload your images to VPHOT, load the standard stars, and download the of resulting photometry reports for each filter. It also now accepts AUID's for star identification; it still allows the old "boulder" ids to be used - but we go to VSP and get the most recent standards values. It also will accept a MaxIM input format.
I need to update the documentation and be sure it ports successfully to a Mac. I'm also asking a few people (currently George and Ken) to continue to test it. I plan to write up a paper and hopefully present at the November meeting.
Gordon
Gordon,
It sounds pretty spiffy. those of us who are stick in the mud spreadsheet guys might just be enticed to use it. As I posted earlier You can teach an old dog new tricks if they are easier and result in the same or more treats. Saving hours messing around with spreadsheets means it is not only easier but we can get more results - more treats.
Brad Walter WBY
I'm no greater a cheerleader of spreadsheet use that you guys! However, I'm mostly using them to select the 45 comp stars that are referenced in the PTGPv4.0 database for NGC7790 and re-formating the data from AIP4WIN for use in PTGP. I'm trying to follow Arne's pipeline for Transform Coeffiecient production using the primary programs offered on the http://www.aavso.org/transform page. here are the software packages I'm using, with a generous dose of spreadsheet re-format/averaging btw them:
1. MaximDL 5.23 for Image acquisition/calibration.
2. Downloaded VSP standards magnitdes for the chart at 23:58:23.2 +61:12:25, 15'x15' wide. This is processed using PhotomCap_14 to make a AIP4WIN readable .STAR file, negating the need to input the BVRI magnitudes manually in AIP4WIN.
3. AIP4WIN, load all BVRI images and process by tagging all the comp stars from the .STAR file of comparison stars. A magnitude instrument report is produced; editable in MS Excel for use in PTGP.
4. PTGPv4.0 processing to derive the Transform Coefficients (TCs).
5. Use TransformApplier to convert your AIP4WIN photometry data to standard magnitudes using the TCs from the previous step.
I'll show some of the Transform Coefficient graphs I'm getting out of PTGP once I get a wider selection of stars redder than B-V 1.36 in the selection of my comp stars.
James
Here is my latest TC results from using 22 stars between B-V of 0.382 to 1.742:
FROM PTGP V4.0 Program using Instrumental Comp Magnitudes (22 Stars):
30-Aug-14
Tbv
1.385
Tb(BV)
0.316
Tv(BV)
0.039
Tvr
1.085
Tv(VR)
0.071
Tr(VR)
-0.007
Tri
0.996
Tr(RI)
-0.007
Ti(RI)
-0.003
Tvi
1.042
Tv(VI)
0.039
Ti(VI)
-0.002
http://www.astroimage.info/files/NGC7790_VSP-Chart%28BritTHAN14.0%20V%29%28B-Vto2.0%292.xls
These 22 stars were the max I could find between PTGP's list of NGC 7790 and Arne's; I omitted Star 128 because its variable NSV 26181 with a B-V over +2. My Tr and Ti values seem to tighten-up and my Tv(BV) got larger. I will process my 2 other nights of NGC 7790 data later on today...takes over an hour to be slow/careful as possible in processing with AIP4WIN and picking the right stars to process the instrumental magnitudes as well as taking the data in MS Excel and re-formatting/averaging the values in a way PTGPv4 reads.
James
Gary:
You can overlap rings in Maxim: place the first set, then click and drag them out of the way, place your second set, then drag the first set of rings back to their proper location. Moving the first set of rings will only affect their positioning on the image you're working on and in any case, when you move them back to the original position they should snap to the centroid of the star (if you have snap to centroid option set.)
Rick Wagner
WCR
Hello Rich
Thanks for the tip. I don't use snap to centroid, as it resamples the data, which according to Mike Newbury, is not good. I use the other two options, star matching, and act on all images. Even then, I find that I need to move some of them a little, as they will identify the wrong star, particularly in crowded fields.. You can see this in the light curve plot, as some images will have a spike in Magnitude. When you examine the image, one can see that the wrong star has been identified.
Gary
Hi Gary,
I may be mistaken, but I think you have something confused. Centroiding is exactly what you want to do - you want your aperture to be centered on the star. The resampling of data that Newberry refers to is when you stack images - you want to stack with integer pixel shifts and not partial shifts. Aperture photometry has nothing to do with resampling. You should "snap to centroid".
Arne
Hello Arne
So you are saying that Maxim does a partial pixel centroiding without resampling the data. OK, I see that that could be what is done. You would not need to resample to calculate the centroid to partial pixel precision. You do have to do partial pixel measurements, but you have to do that anyway. Learn something every day.
This does remove some of the "Art" of placing the aperture and rings to miss near by neighbors. :)
Gary
Hello Arne
So you are saying that Maxim does a partial pixel centroiding without resampling the data. OK, I see that that could be what is done. You would not need to resample to calculate the centroid to partial pixel precision. You do have to do partial pixel measurements, but you have to do that anyway. Learn something every day.
This does remove some of the "Art" of placing the aperture and rings to miss near by neighbors. :)
Gary
I couldn't get to my data these last couple of days because of my house's power failure due to back-to-back blown transformers serving our area's electrical during this LA heatwave! (No photometry data taken for UT dates 15&16Sep14). Here are my latest transforms using the present version of PTGP4.0:
NGC 7790 Transforms
FROM PTGP V4.0 Program using 22 Instrumental magnitudes:
AVG'S
S_DEV
30-Aug-14
6-Sep-14
7-Sep-14
10-Sep-14
11-Sep-14
12-Sep-14
Tbv
1.385
1.369
1.372
1.373
1.380
1.375
1.376
5.854E-03
Tb(BV)
0.316
0.299
0.302
0.300
0.311
0.313
0.307
7.360E-03
Tv((BV)
0.039
0.030
0.031
0.028
0.036
0.041
0.034
5.269E-03
Tvr
1.085
1.100
1.099
1.103
1.083
1.085
1.093
9.072E-03
Tv(VR)
0.071
0.055
0.057
0.052
0.066
0.074
0.063
9.094E-03
Tr(VR)
-0.007
-0.035
-0.033
-0.042
-0.011
-0.004
-0.022
1.649E-02
Tri
0.996
0.991
0.990
0.982
1.001
1.007
0.995
8.826E-03
Tr(RI)
-0.007
-0.040
-0.038
-0.048
-0.011
-0.003
-0.025
1.962E-02
Ti(RI)
-0.003
-0.031
-0.028
-0.030
-0.013
-0.010
-0.019
1.199E-02
Tvi
1.042
1.047
1.046
1.044
1.043
1.047
1.045
2.137E-03
Tv(VI)
0.039
0.030
0.031
0.028
0.036
0.040
0.034
5.020E-03
Ti(VI)
-0.002
-0.015
-0.013
-0.014
-0.006
-0.005
-0.009
5.492E-03
My Tr(RI) at 1.962E-02 seems to have the greatest night-to-night standard deviation while Tvi has the lease at 2.137E-03 standard deviation; i.e Tr(RI) has 9.2x the deviation of Tvi. R and V filter also suffer from the greatest backround amounts from my urban sky. I wonder if these would have smaller deviations when shot from my darksky site?
I don't think I'll do further work on this until Gordon's version of PTGP is released running with Arne's latest NGC 7790 standards and able to accomidate both AIP4WIN and Maxim DL extended photometry formats.
Between these 6 nights of data, I probably made 1,200 (200 per date) keystrokes in Notepad to get the data and semi-colons spaced correctly after pre-formatting it in MS-Excel in an acceptable form for PTGPV4.0 processing.
James
Here is a spreadsheet with the numerical values of the transformation coefficients and graphs for each TC for the first 3 dates (30Aug14, 06Sep14, & 07Sep14). Large download because I cut and pasted PTGP's image graphs of each TC into this spreadsheet.
http://www.astroimage.info/images/files/TRANS%2822Stars%2930Aug%20&%206,7,10,11,12Sep14-Formated.xls
James
Thanks for the spreadsheet and plots, James! The one coefficient that is missing is Tr(VI). I recommend this color index for determining Rc if you are doing VRI photometry. It means using three filters to transform, but the coefficient is better determined.
Tb still bothers me. The KAF-3200me chip is the most sensitive chip of the Kodak/Trusense line, and has good blue response. Your coefficient says that for a (B-V) of 1, you need to make your instrumental magnitude 0.3mag fainter in order to get the true system. That is, the instrumental magnitude is too bright for red stars, which is common if the filter+sensor is too "red". That happens with the non-blue-enhanced CCDs like the old KAF-0400.
With BSM South, which is currently using an ST10-XME, Astrodon filters, and a refractor, we're getting Tb(BV) around -0.05. Do you know the vintage/generation of your Astrodon B filter? They've made three versions; the latest has a ** beside the name.
It is probably ok; I just look at things from many angles when trying to understand the calibration process.
Arne
To: Arne,
RE:"Do you know the vintage/generation of your Astrodon B filter? " Yes, here is what the label on the filter says:
Astrodon Photometrics J/C B***
I assume that the J/C means Johnson/Cousins. I purchased it last year after my old Schuler blue glass filter became crystallized. The filter looks blue when white light passes through it and has a yellow appearance when viewed from the side from backscattered light; like most dichromatics, it produced the opposite color when illuminated on the surface.
RE:"That is, the instrumental magnitude is too bright for red stars, which is common if the filter+sensor is too "red". " Don't know if this means anything, but my backround ADU in red is worse than through any other filter. R filter backround is as high as 2252 ADU while the V filter is a little better at 1930 ADU; 180 sec 1x1 bin exposures with ST-10XME. Both my R & I filters are old schueler glass filters. Might this also be the cause? My B & V are modern Astrodon dichromatics.
Could the other reason be that I'm using identical exposures per BVRI filter? I had the vary the exposures for taking XZ Cet because of saturation issues (B&V=45sec, R=30sec, & I=60sec, all 1x1 bin).
James
I know that this is probably a question resulting from ignorance and bordering on heresy, but why are the new APASS values for the NGC 7790 standard field preferred over Arne's values from October 14, 1999 (attached)?
I have compared the old Henden sequence data to the new APASS values. The differences overall are not statistically significant and the uncertainties of almost all APASS measurements are much larger than the uncertainties published for the 1999 sequence (or derived from those uncertainties using the normal propagation of uncorrelated error formula for addition and subtraction).
I did a comparison, star by star, for all of the VSP standard field stars and comparison stars (which are not included in the VSP standard field for reasons discussed earlier in this thread) corresponding to stars in the 1999 sequence for the standard field. 29 of the 31 stars of the 1999 field have counterparts in the new data. The comparison is attached.
I also have attached a file which includes calculations of single color uncertainties derived from the V and color index uncertainties published in the 1999 data.
Can someone explain what I am overlooking that makes the new APASS sequence the better choice? Are there systematic errors in the old data that are eliminated in the new? Except perhaps for the V filter this isn't apparent in the data. However, even the differences between old and new V data do not pass the normal 95% confidence interval test for rejecting the null hypothesis.
Arne, you are probably most able to provide insight since you are the primary author of both sets of data. I'm Not trying to be a pain, but If left to my own devices I would probably select the 1999 data over the APASS data except for your admonition to use the new data. I think understanding the reasons for this would be very enlightening
Brad Walter, WBY
Attached is a revised spreadsheet with the 53 stars in the last one but with the embedded finder chart updated with the additional stars. The additional stars are NOFS stars that were included on a comp star sequence chart 13644BVP and sequence table 13618HDU printed on October 12, 2104 and, therefore, not included on the standard field chart. The spreadsheet includes comparisons to the 1999 Henden NOFS sequence values. I also added a chart showing the distribution of B-V values. These stars contain a reasonable spread of B-V values.
I fit a straight line to the U-B vs. B-V chart only to highlight that the expected slope reversal (dip) in this curve is exhibited at around B-V = 0.5.
Arne, just as a matter of curiosity, the minimum of this dip is shifted toward the red about 0.3 to 0.4 B-V from the curves one frequently sees published in textbooks, including Henden and Kaitchuck. Is this primarily because these stars are in a cluster rather than main sequence field stars?
Also attached is a full page version of the "click list" finder chart. Both the full page version and the one imbedded in the spreadsheet have been updated so the sequential "click numbers" are black rather than red to avoid printing problems and to allow them to be seen under red light. Thanks to Gary for pointing that out to me. Everything showed up as black in the PDF and I didn't realize the numbers were red until he mentioned it.
Hi Brad,
If you look at the photometry table, the source ID is 10, which is NOFS, for all of these stars. They are not APASS calibrations. The NOFS table that I originally generated on the binaries site used 4 nights of 1999 data; the VSP table uses many more nights. The quoted errors are the errors of a single observation, not of the mean, so they appear larger than they really are (I do this to be conservative).
The more nights that you include in the determination of mean values, the better the result, assuming random distribution. Four nights was a minimum in my mind, so I continued to take data, and to use both short and long exposure times to better cover the magnitude range. I think the VSP values are better and more homogeneous than the 1999 data and recommend their use.
Arne
I don't know haow I made that mistake. APASS is 29.
Thanks for the explanation. It makes perfect sense.
Brad Walter, WBY