This would normally go to the Campaigns forum, but I'm planning a large part of this campaign to be a learning experience (for you AND me!). As we go along, I'll revise this initial post to reflect changes. Keep a thick skin, but put your photometry out there for critique!
V838 Mon is a very peculiar star, having an outburst to 6th magnitude in 2002, a decline to 16th, and now over the past decade, a rise to about 13th magnitude. Theories as to what is happening are in the paper (highly recommended reading) that Brian Skiff posted recently on this forum:
Several things to note: the star continues in increase in brightness; the source is still extremely red; and there seems to be a long-period variation in brightness of about 300 days. Is this real, or is it a seasonal artifact? Are there short-term variations that have been missed by the daily monitoring? Are any of the colors peculiar?
I would like people to monitor this star, but with specific requirements. If you meet those requirements, I will comment on your results on this thread. Hopefully, we'll both learn something from the experience! I intend to monitor V838 Mon myself using a remote telescope, so you will see my results as well (and you can comment on them too!).
All observations must be in at least two of the common Johnson/Cousins filters. B&V would be great, though the star is quite faint at B; V&I are also good, as examples. I'd love U-band, but the target is about 20th magnitude there. If you have a U filter, this is a good test to see if there is a red leak! If you find a red leak in either U or B, let the forum know.
You need to observe with at least 3 sets of measures per visit. If you use B and V, for example, you should observe in the pattern BVBVBV. The submitted estimate per filter should be the mean and standard deviation of these sets - don't use the error that you get from your usual analysis software. Don't submit each set individually (except for those doing time series maybe). This means a bit of work for you, but will give a far better estimate of your uncertainty.
All observations should be transformed if possible. This is a very red star. If you don't transform, you will see significant offsets between your data and other observers. If you can't transform, go ahead and submit anyway - that will show others why transformation is important. If you don't know how to transform, the new CCD Manual describes the process, there are TG and TA applications on the website, and VPHOT or LesvePhot will help. People on the forums are always willing to help you learn the process!
No ensembles should be used. We'll do this the classical way, with one comparison star and one check star, and these will be the same for everyone.
Use the Extended Format properly. I want the comp and check stars to be identified with their AUIDs. Use instrumental magnitudes for CMAG and KMAG. Include airmass values. I don't care what analysis software you use, but if possible, put the name of the package in the NOTES field and perhaps your transformation coefficient as well. Include your aperture diameter in arcsec if possible, as there are a couple of faint stars to the east that might get included, depending on your seeing. These notes will be helpful hints to me if something seems off.
As an extra, there is a variable star that is currently listed as a comp star in the field. Sebastian looked at this star's light curve in GAIA, and says that it is an SPB variable with 0.03mag amplitude, with no period given. Let's monitor it as well and report data on both stars. That will also help in understanding the precision of your observations.
V838 Mon 07 04 04.82 -03 50 50.6 000-BBM-355 V=13.5
UCAC4 431-026668 07 04 03.55 -03 49 04.3 000-BBM-352 V-13.97 (original 139 comp star)
000-BBM-342 07:03:59.20 -03:51:45.0 current 123 comp star, fairly red
000-BBM-332 07:03:51.72 -03:51:28.0 current 143 comp star
This campaign will continue until early May, when the field gets too close to the Sun. Note: if you can do spectra, this is a very interesting target and I recommend monitoring it spectroscopically as well.
Hopefully we will get lots of observers for this carefully designed campaign. It should provide sufficient data to conduct a statistical evaluation of data quality.
I plan to analyze the target with a comp star ensemble as well (in addition to the single comp) since I'm interested in how this practice impacts the calculated magnitude. I plan to calculate both non-transformed and transformed magnitude(s) with the single comp and an ensemble of three nearby comps (2 more than currently identified).
Update: I plan to use comp star 127 and 130 as the two additional comps to avoid any saturation issues. May offset the image center slightly to help center all comps and target.
000-BBM-320 07:03:38.07 [105.90862274°]-03:50:09.6 [-3.83599997°]12712.706 (0.007) 0.535 (0.017)
000-BBM-323 07:03:42.49 [105.9270401°]-03:48:19.9 [-3.8055279299999993°]13013.042 (0.001) 0.290 (0.001)
It will be interesting to see your results! I specifically asked for a single comp for several reasons, a few of which are given below:
- there are many ways of making an ensemble, and I wanted to avoid critiquing the process
- using multiple comps also brings into play any calibration difference between the various stars. Using a single comp hard-codes that offset, so that you can concentrate on technique differences
- due to field-of-view differences, the stars in an ensemble for one observer may be different from another
- ensembles bring into play other possible systematic effects, such as improper flatfielding and vignetting removal
Used properly, ensemble photometry is very beneficial, especially in error/uncertainty calculations from a single image and for averaging other systematic effects like spatial or magnitude bias. For this particular project, I want something simple and homogeneous. Maybe a subsequent project (I anticipate this to be a continuing fun exercise) will cover an ensemble target.
I might mention that Peter Stetson has a modest set of photometry for this field here:
The file under this directory called V838_Mon.txt shows the sources of the CCD images he used. The photometry was updated as of 2022 Nov 26. This indicates that B,V,I are well-observed, but U and R are weak. For Arne's comp star BBM-342 = Stetson star S10, he shows: V = 12.326 +/- 0.004, B = 13.504 +/- 0.004, and I = 11.114 +/- 0.003. Thus B-V = 1.178 +/- 0.006 and V-I = 1.212 +/- 0.005. All this from 30 to 50 observations in each filter.
Now that Brian has mentioned the different mags from Stetson, which differ from AAVSO mags by a few hundredths, can you confirm that we should stick with the default AAVSO sequence mags (or not)!? Both VPhot and LesvePhotometry use those mags. IMHO, consistency is the most important issue for this campaign.
Please use the comp star magnitude values that are given in the Photometry Table of VSP. Please do not go elsewhere to find "better" values. One of the big reasons for going to the single comp star is so that everyone is on the same system, and if you start using other values, then I don't know if offsets are caused by that change, or by transformation, or what. Thanks!
I set up my procedure for analyzing V838 Mon (and UCAC4 431-026668) in accordance with your campaign criteria. Comp and check identified by AUID in report. (Single Comp 123. Check 143.). Targets identified by their names not AUID. BTW, the UCAC target initially shows as a Gaia target in VPhot. I changed the name manually in my saved sequence.
My first BVI set submitted to AID is not transformed. Future sets will be transformed. Could you use both transformed and non-transformed results?
The mean and standard deviation of the single reported magnitude were calculated by analyzing 3 images (as a time series) in each filter. I used TransformApplier (TA, semi-independent tool within VPhot) to facilitate this calculation. This tool allows one to 'aggregate' groups of transformed magnitudes (e.g., 3 magnitudes) to obtain a mean and standard deviation that represent this desired averaging process.
Your idea of submitting both transformed and untrransformed measures would be nice to show how transformation makes everyone's values move together, but I wonder how to identify them easily in the light curve generator. Maybe the way to do it is to submit the first few measures untransformed, and then subsequent ones transformed.
Arne et al:
Yes, 'filtering' and displaying is the trick! Might be worth talking to George Silvis?
Alternatively, I've been playing with VStar. This tool is readily available and will allow one to quite easily create 'filtering' algorithms to separate and view magnitudes that use a specific comp name (by AUID name ) to ensure that we are looking at campaign data and then that have been transformed or non-transformed to visually (or statistically) compare the two sets.
If we create the 'filtering' algorithm and share it, everyone can do it. Or you can do the separation and share the screenshot. I need a little help to finish this option but I think it will work? Any VStar experts who are participating in this campaign?
I'm not sure the AID will accept two measures of the same star from the same image: The second submission will knockout the first when the scan for duplicates is made.
A transformed record processed by VPhot will include the VMAGINS, the target instrumental mag. With this you can easily get the untransformed value:
VMAG(untransformed) = ( VMAGINS - CMAGINS ) + CREFMAG
My recollection of asking Matt Templeton if I should submit both transformed and untransformed measures was that it was verboten.
George et al:
Both transformed and untransformed mags for the same two identical images can be submitted to the AID. Both image mags are reported in the AID with their slightly different magnitudes. They are differentiated by the value of the Transform field in the report (yes or no). I just carefully tested this to prove it to myself.
So for this campaign, both transformed and untransformed reports can be submitted separately and displayed.
In general, I can see no reason for an observer to normally submit both mags (transformed is preferred) BUT it will be very helpful for a visual comparison and statistical analysis of the results of this campaign.
Ken notes that TA can average multiple observations. Richard Sabo has also written a nice utility for Windows:
Look for the Boxster zip file download at the bottom. His AAVSO Util application does the data set merging. So observers have no excuse for not taking 3-5 multifilter sets and averaging them before submission. :)
The chosen comp star has (B-V) = 1.152. V838 Mon is much redder than this, and the UCAC second target is much bluer. This was intentional - it will show the value of transformation. You can not always find a comp star that matches the target star color. That said, you may have difficulty with the differential brightness of the comp star at the Ic bandpass. We can think about this later.
As I mentioned, V838 Mon is a really nice test for red leak in your filters. Any blue filter (Johnson U,B; Sloan u',g') may have a "small" red leak. For normal stars, you don't notice it. For carbon stars and things like V838 Mon, ANY red leak is very noticeable. Usually it will be evidenced by: (a) the star appearing much brighter in the blue wavelengths than it should (brighter at U than at V, for example); (b) a fuzzy star profile compared to others in the field; (c) often double images when observing close to the horizon, where differential refraction splits the star into its red leak and its blue passband components.
Regarding Ken's double-submission of transformed and untransformed measures of the same image: I would prefer that you only submit one or the other, and not both. That clutters up the LCG. There is no LCG option that only plots transformed data, for example. If you want to see what effect transformation yields, submit untransformed values for a half-dozen nights, then submit transformed ones for a second half-dozen nights. After inspection (and maybe at the end of the campaign), you can go back and replace the untransformed values.
Enjoy! This star is near the Celestial Equator and so can be observed by everyone.
you wrote: "B&V would be great, though the star is quite faint at B; V&I are also good"
The photometry table of the VSP chart does not provide Ic magnitudes (nor Rc magnitudes). Which Ic magnitude for the compstar should be used?
To deal with the differential brightness of the comp star at the Ic bandpass, I intend to stack a number of exposures.
If there is also an Rc magnitude available I consider to observe it in 4 colors.
If you go to the photometry table from the Finder Chart from the AAVSO front page, you will only see the default V and B-V.
However if you back up to the VSP page from the Finder Chart Photometry page, near the bottom you can find check boxes where you can request other mags (e.g., R;I)
Try to find it before someone just gives you the answer. It shouldn't be too hard.
Hi Ken and Arne,
that is exactly what I did. For "Plot a finder chart or a table of field photometry?", I selected "Photometry".
At the bottom of the form, I selected the filters I wanted (in this case B, Rc and Ic).
Despite these selections, a chart is displayed rather than a table, with a link at the top of the page "photometry for this chart". This link provides only V and B-V magnitudes.
I have used VSP multiple times, and as far as I remember it always came up with a table instead of a chart.
Does Boxster also calculate error as you request with a standard deviation?
Is there any value in a time series or multiple sets/night such as hourly?
Arne et al:
My first set of images required some playing with the exposures to get reasonable SNR for the B and I images as you hinted at. I ended up using the following on my 12"CDK.
In my poor skies near Boston, for this southern target, I needed to carefully review the images and results, and remove some clearly 'bad' images.
As expected, the transformation of this red target certainly made some significant changes in the reported mags.
I got a nice set of 5 x BVI of V0838 Mon with exposures at 300, 60 and 90s
Star Filter Vraw TranMag diff VERR VERRt VERR2
"UCAC4 431-026668" B 14.279 14.180 -0.099 0.003 0.009 0.003
"UCAC4 431-026668" I 13.525 13.546 0.021 0.010 0.018 0.013
"UCAC4 431-026668" V 13.899 13.935 0.036 0.004 0.013 0.007
"V0838 Mon " V 13.096 13.027 -0.068 0.004 0.023 0.011
"V0838 Mon " B 15.706 15.893 0.187 0.006 0.017 0.006
"V0838 Mon " I 9.190 9.170 -0.020 0.004 0.032 0.012
Tbv= 1.1750, +/- 0.0090
Tb_bv= 0.1090, +/- 0.0090
Tv_vi= -0.0410, +/- 0.0070
Note the increase in the error with transformation from VERR to VERRt
This is because VPhot includes in the computation of error the error
the transform coefficient.
VERR2 is the transformed error with out applying the coefficient error.
Should VPhot not be applying the transform coefficient error?
This error growth might be one of the reasons users are shy about
transforming their data.
VPhot (TA) applies the transform equations to the transform group as a set of simultaneous equations.
Those equations are translated into a set of error equations applying the rules of errors accumulating in quadrature.
Special rules apply setting the comp instrumental error to zero (assuming it correlates to the target instrumental error)
and the documented error of the comp reference is also set to zero as it is to be ignored.
This set of equations of the error are processed handled like the magnitude computations, as a set of simultaneous equations.
The issue now is if the error of the transform coefficients should be ignored too...
Ignoring that error still does move the final error around, but only a little, based on the coefficients; it does not change the total error.
I'm happy to share gory details with any who want to see:
eg, for a Tb_bv equation:
(x is b, y is b an z is v)
mag: r= Xs = xs + (Xc - xc) + Tx_yz * ((Ys - Zs) - (Yc - Zc));
err: s = ((Ys - Zs) - (Yc - Zc));
if (s == 0) r = rXs = rxs;
else r = rXs = sqrt((pow(rxs, 2) + pow(rXc, 2) + pow(rxc, 2)) + pow(Tx_yz*s, 2)*(pow(rTx_yz / Tx_yz, 2) +
(pow(rYs, 2) + pow(rZs, 2) + pow(rYc, 2) + pow(rZc, 2)) / pow(s, 2)));
Arne et al:
I find the B images problematic because the target's B mag is so faint. I was having some difficulty avoiding slight trailing on the first B image in the series (BVRI). The other filter images have no problems. They are bright enough.
I decided to add a slightly shorter throw away exposure (1x150s) at the beginning of each BVRI set to make sure the gears caught up after the initial slew, followed by 2x300s B exposures, then the other filters.
So, I now run:
B 150s, B 300s, B 300s, V 180s, R 120s, I 100s
I repeat this set three additional times. I throw away all the B 150s images. I stack the pairs of B 300s images. I run the images in a time series (4 x BVRI) and save the AEFF Report. I import this report into TA and aggregate the four sets.
The stacked B images look good with a relatively obvious and well defined target star in B. SNR is still only 37. Extra effort but I feel much more confident in the B mag.
I'm still getting a much brighter B mag than all others. Is my coefficient correct? It is a very red star (B-V = 2.4+), and may exhibit a large diff from non-transformed mag?
I don't know enough about you observing system to comment but will offer a description of the equipment etc. that I observe with.
Perhaps you and others participating can reply with the equivalent information.
I observe portable, yes, every night, still and even at my age!
SkyWatcher Esprit 100 + Atik490EX + Astrodon Clear B V Rc and Ic filters vintage 2020. iOptron CEM40-EC with off axis guiding. Pixel scale is about 1.4"/pixel.
B***_27R1.25" Mounted Astrodon Johnson/Cousins Improved B edge-blackened
V*_27R1.25" Mounted Astrodon Johnson/Cousins V edge-blackened
Rc*_27R1.25" Mounted Astrodon Johnson/Cousins Rc edge-blackened
Ic*_27R1.25" Mounted Astrodon Johnson/Cousins Ic edge-blackened
I flat-field using DeepSkyStacker and I use Aperture Photometry Tool (APT) for photometry and excel to reduce the photometry and apply the transforms.
As mentioned previously, especially in B, there is significant contamination from field stars in or very near the central aperture and also in the sky ring for V838 Mon. At low SNRs these contaminators if in the apertures will skew the aperture-based photometry. APT provides a tool to reject pixels, called zapping, so they aren't used in the computations which is what I do as needed. None of this is automated which means I inspect every image I do photometry on...
The transform coeficients I am using are a simple average from 30 transform determination nights measured since 2020 when I got the Astrodon filters. 43% of the transform nights were obtained using SA110 with at least one of the red extension stars that Brian Skiff mentioned were available a few years ago. I usually use SA110-516 (R112) SA110-446 (R140) one or both that fits in the dynamic range I use for that field. B-V fits range from about +0.5 to +2.9. At the bottom is the forum link to Brian's SA-110 extension info. I have a hand marked AAVSO standard chart image that I can make available that shows the red extension stars if anyone wants it I can provide it.
There might be a slight 2nd order effect in the transforms but it is not at all clear if it is significant. Something for future analysis to get some formal statistical estimates of 2nd order significance and then apply that down-stream in regular processing.
Uploaded my first nights BVRcIc photometry. Fairly high air mass but both V838 Mon and 000-BBM-352 seem consistent with the others. I observe portable and didn't feel like waiting longer for the variable to get higher, V838 Mon having just cleared the tree line when I started! Fairly transparent night for east-coast. At least in my images there is a faint companion very close to V838 Mon in V and B especially. Also several had stars in the sky ring, notably V838 Mon. Aperture Photometry Tool has a zap function and I discarded the pixels as needed for stars in the sky ring. In future I'll adjust the exposures at bit, especially will make the B exposures longer. All the requested details should be in the db records.
Jim DeYoung (DEY)
Since I have been looking at a lot of GAIA3 spectra, I checked the one for V838 Mon. I estimate it to be about temperature type M2 from the TiO bands in the far-red part of the spectrum redward of H-alpha. The lack of the MgH band near 5200A indicates luminosity class at least 'giant', but there's no obvious discrimination for more luminous stars in the coarse spectrum. The spectrum is clearly of earlier type than the one for RS Per, normally M3.5I or M4I. Dunno the date for the GAIA spectrum, but presumably about 2016, the nominal epoch of the astrometry. The photometry seems to have been flat since then, so perhaps this is roughly the current type. Someone like Robin Leadbeater could probably do better!
Everyone still here and observing?
Weather has been poor on the east-coast of NA. I have a few more nights in the can but am behind in processing and uploading them.
I'm still acquiring data, but confess that I am behind in processing myself! I've been away from home quite often this winter. My plans are to present data from multiple AAVSOnet/personal telescopes soon, both transformed and untransformed, and then compare against everyone else. Be sure to measure the constant target too, as that one will be very important in the discussion.
V838 Mon will be observable for about 3 more months, so please keep monitoring the field!
Good luck when you try to understand the campaign data, especially in B filter, transformed vs non-transformed, for all observers. ;-)
Doesn't help that the target is so faint in B! Most of the problem??
We probably could use a few more observers and a better sampling rate by all. The weather here in VA has not been very photometric even on the best nights this winter. V838 Mon with the B filter is noisy due to SNR and faint companions in the apertures. The Gaia low-amplitude blue star (the 139 comp) is faint also. This is a tough task to get exposures that cover the entire range of colors and objects for sure. I'm using 300s B, 240s V, 120s Rc & 90s Ic. Had to lower the Ic exposure from 120s recently as the air mass has gotten lower since I started observing.
Unfortunately, something went bad in my mount.RA axis. Beyond my ability to fix after initial inspection and evaluation. Will be down for the next two weeks or so, until I get it back from iOptron.
The price has gone up $10, but you could order one of these and measure your filter transmission from 350 to 1020 nm..
It is the one I use.
The Amateur Telescope Makers of Boston (ATMoB) group were donated a very nice Planewave CDK17 telescope on a Paramount ME inside of an excellent commercial roll-off building. They purchased a QHY600 camera, and then borrowed some filters from me plus purchased a few. This resulted in a really interesting combination of filters:
B - Astrodon B***
V - Chroma square-topped
I - Astrodon Sloan i'
These are the filters I've used in monitoring V838 Mon with this facility. If you look at the AAVSO Light Curve Generator plot for V838 Mon, and then zoom in to just the past year, highlight observer HAXA. You will notice that this observer is Arne Henden, just like HQA is Arne Henden. The difference is that the HAXA measures are untransformed, and the HQA reductions of the same observations are transformed. The precision of the measurements is excellent - I did sets of 3 exposures per filter and combined them using either VPHOT or TA to single means with standard deviation. However, the accuracy of the untransformed and transformed datasets obviously differ.
The untransformed "Ic", really a Sloan i' filter but using the Ic value for the specified comp star, obviously differs from everyone else. This is an extreme case, but shows why other observers who also do not transform can have offsets. At the same time, look at the HQA reduction of the same dataset. Once you transform, the measures line up with the mean curve. Likewise, when you look at B, there is a smaller difference between the transformed and untransformed values. This is because Astrodon B*** is really close to the Johnson/Cousins system, and the transformation has less of an effect. However, the bottom line from this is: "Please Transform!"
I have quite a bit of earlier MAO observations that I'll upload over the next week. Then I'll highlight some data from TAGRA, a facility in Spain that uses Baader filters. This experiment has many options of data analysis, so I hope that you enjoy the process and get some insight along the way!
Also, the HAXA data will disappear at the end of the project.
My first question is what are you doing differently than some of us to get such smooth results from night to night? I'm guessing that my photometry has more night-to-night "stuff" because I observe portable. Observing portable introduces additional noise from polar alignment differences, psf changes, guiding differences, etc. from not being a permanent mounted telescope. I'm surprised at the amplitude of the noise of the individual image photometry that went into my N=3 averages, not at all what I expected, even some of the good SNR filter images seems more noisy that I expected. Maybe it is the Atik490EX which uses the Sony ICX814/5 CCD which is contributing? Here is a spec sheet on the CCD...
Jim DeYoung (DEY)
I find your light curves to be very 'smooth' AND the small amplitude sinusoidal curve in VRI to be very convincing! At the altitude we are working, I think the precisions are not unexpected. Is your reported error based on normal mean and standard deviation (from excel) of the three magnitudes for each filter image on a given night? Did you use Stds or Stdp? VPhot may use a slightly different algorithm.
The B data is certainly the poorest since the mags are so faint, but may also exhibit the expected largest variation? I find Arne's mags a bit surprising for that filter? He must have used a long exposure and had less problem with the nearby faint star. Probably used a small aperture?
I don't think they are as smooth as Arne's! As in many places in measurement science there are lots of places for error and bias to creep in, always difficult to isolate the sources. I'll have to scare up a standard error budget for CCD photometry... I'm sure there is one but I haven't looked at one in a long time. I transform in excel and use the simple mean and the old STDEV for this project. Btw, STDEV is now being replaced by STDEV.S by the excel wizards at MS. With N=3 I don't expect any differences, however. Now that I know MS is recommending STDEV.S over the olde and I mean (mean joke) olde STDEV I'll have to find out if there is a way to easily change them in all the excel SS I have. I haven't phased the very low amplitude "comp" star yet... that should be illuminating but olde Fourier may help with the smoothing!
The season for V838 Mon ended in April/May, and it appears again for northern hemisphere folks in September. I was able to get ~5 months of data from late October through the end of March. When I was observing this star during its nova phase, I pushed it at NOFS (+32 latitude) into the first few days in May. Pretty amazing, but awfully low in the sky!
I used two telescopes this past season. V838 Mon is relatively faint, especially at B, so I wanted reasonable collecting area. V838 Mon has some nearby companions, so I wanted decent seeing (at Ic, where it is still very bright, these faint companions are not important, and poor seeing and poor spatial resolution are ok). I did not want to take up AAVSOnet time, so I used the two telescopes outside of the network where I have observing time: the Mittelman-ATMoB Observatory, or MAO, in Massachusetts, which is a Planewave CDK17 plus Paramount; and the TAGRA telescope, a CDK20 plus L-500 in alt-az configuration in Spain. MAO uses a QHY600 camera with an Astrodon B(3rd generation), an Astrodon V (first generation), and an Astrodon Sloan i' (2nd generation). TAGRA uses a ZWO ASI-2600 camera with Baader B,V,I filters. So using these two systems let me compare a QHY with a ZWO camera; Astrodon vs. Baader; Sloan i' vs. Cousins Ic; alt-az vs. equatorial mounts; good dark location vs. coastal U.S, etc. Lots of fun! While the equipment used is not cheap, it is still commonly found in amateur hands. You can use less expensive setups and get good results, especially for brighter, uncrowded targets.
I'm in the process of finishing up the MAO analysis, with 21 nights remaining. I'm not spending all day on analysis, so it will take me until mid-week to finish MAO. It started first, and so it has data from October until early March. TAGRA started in mid-December on this field, and finished at the end of March. This gives plenty of overlap between the two systems. I'm actually a bit disappointed, because MAO had hardware problems in December, and I didn't immediately place V838 Mon on the TAGRA queue - so there is a two week gap without observations.
For most of the observing, I used a 21-exposure series. I got nine 120-second exposures at B; nine 60-second exposures at V; and three 30-second exposures at I/i'. After inspecting each image for problems, I stacked frames for each filter in sets of 3. The end result is a total exposure time of 360sec B, 180sec V, 30sec I per group, in the order BBBVVVIII. I performed photometry on each of the nine master images. I averaged each set of 3 values per filter to determine mean and standard deviation. Only the one comparison star was used for this experiment, and all data are transformed. VPHOT was used exclusively.
I recommend getting at least 3 measures per filter and submitting the average as the target magnitude and the standard deviation as the target error. This gives far more realistic uncertainties than trying to derive something from the CCD Equation, ensemble differences or Poisson statistics.
Like I've mentioned before, I also temporarily submitted some untransformed analysis of the MAO data under the obscode HQAX. This will get removed after all of my data for the season is analyzed. You can look at my data by using the Light Curve Generator, unclick the "All Observers" box, and clicking the smaller box between HQA and Henden.
More later in the week, when I'm finished with the MAO data. I'll also look through postings on this thread and see if there is anything I have not responded to. Sorry for the delay in completing this project/campaign; life always seems to get in the way!
Absolutely! Right now, I'm generating what I think is a good set of data. Then I plan on looking at all submitted data and asking questions/making comments. This is a learning exercise for everyone (including myself) as well as a scientific project.
I finished the MAO processing, so all of its data are now in the AID. The light curve is pretty good for V and Ic; less so at B. I'm going to investigate a couple of B-band data points later in the analysis. The early MAO data suffered from three problems: the images were mostly taken around airmass 2, and so had poorer seeing along with possible extinction issues; the drive was misbehaving, which resulted in the removal of several images due to trailing, and therefore fewer images were available for stacking; and because of the drive problems, the exposures were shortened, resulting in lower signal/noise. This really affected the B-band photometry. Normally, if this was a really important target, I'd revisit the B-band analysis and perhaps use psf-fitting where I can often get better results on faint targets, especially in crowded fields like this. But here, I want to use the same techniques and software as most AAVSO observers do.
There is an interesting result, even from just one telescope. V-band shows a nice sinusoidal variation over the season, as does I-band. However, the two light curves are out of phase, and the Ic amplitude is larger than the V amplitude. B doesn't show any obvious variation, though the poor quality of the measurements makes this difficult to see. These three light curves combined do not look like any astrophysical object that I've studied.
I compared some of the other observer's datasets, and we agree on zeropoint for the V curve, but differ on B and Ic, even if transformed. The precision also is better in my light curve than some others. There will be some interesting discussions shortly! I will ask some questions as we go along.
I want to include a fair chunk of the TAGRA data before I make any detailed comments. That data analysis begins this week.