edit 20260906
edit 20260907
This is everyone’s favorite target, with over 500K measures in the AID. Someday it will go into outburst! Are you all ready? The Sequence Team has actually generated a sequence that goes all the way to first magnitude. You can retrieve finding charts and photometry tables for various fields of view and brightness at:
that I highly recommend observers to check out. Are you prepared for the outburst? What exposure times are you going to use when the star is at 7th/5th/3rd/1st magnitude? Do some trial runs to get ready. You may even want to set up a wide-field imaging system specifically for this event. Think about it now, and not after the outburst is detected!
First and foremost, there was concerted effort in 2024 to improve the T CrB sequence. Some stars were dropped, others added, and in all cases, the sequence star photometry was carefully checked to be as accurate as possible. This latter step modified several magnitudes, especially Ic. PLEASE USE THE NEW SEQUENCE. It is the one shown in the photometry tables in that prior link. If you are submitting data using older chart information, you need to update. This includes any sequence you may have saved in VPHOT. The date of the change was July 2, 2024, or the equivalent JD 2460493.5. A Best Practice is to check your sequences about once/month. Unfortunately, the AAVSO gives no way to identify when something has changed. The sequence database is updated, but no notification is given and you have to compare with older data yourself. We should tell HQ to make this a priority! Most of the time, changes occur when a sequence star is found to be variable, or when improved Rc/Ic photometry becomes available.
So, let’s look at the light curve:
this is the entire light curve, including visual observations and fainter-thans. Since these posts are CCD/CMOS Best Practices, I’ll reduce this to BVRI:
I’ve also left in those yellow measures that I’ll talk about in a minute. Note that I’ve included data from JD 2460400 to current, and so the sequence change occurred near that first vertical bar. I don’t see any great change in the data quality, which I guess is good!
We could look at any of the individual filters, so I’m going to pick V and Ic. First V (with the yellow measures):
Note that the yellow measures don’t really fit V - but they don’t fit anything else either. These are submissions labeled “NA”. Best Practice: pick a filter type when submitting your observations! Really, the AAVSO should prohibit such submissions (and probably does with Webobs 2.0). So, here is the V curve without yellow:
It actually looks pretty good (with caveats). First, there are several faint datasets. Most researchers would just ignore them since there are simultaneous observations closer to the main curve. That main curve shows the 113.8day sinusoidal variation, which is the half-orbital-period signature of the ellipsoidal secondary star. You can tell it is ellipsoidal variation due to the amplitude and the fact that it is nearly the same amplitude in all passbands (geometry, not stellar temperature).
I said “pretty good”. If you bin/average the measures, you get a nice clean light curve. However, at any given time, the measures scatter by about 0.5mag peak-to-peak, or about 0.2mag RMS. This is far larger than the typical quoted uncertainties of each datapoint. The usual culprits of transformation, comp star selection, and classical/ensemble differences are probably to blame. You can do these adjustments for a well-studied star, but in most cases, you don’t have 500K observations to guide you. That is why it is important to submit your best photometry.
Now, look at Ic:
This curve is worse. There is a set of measures in the 2461180 to 2461280 range that are a magnitude too faint, but are labelled “transformed”. The observer has made some mistake, and has not gone back to notice this with the Light Curve Generator. There are also other fainter measures in the 2460700-2460850 range. Other than that, the main curve obviously shows monitoring by individual observers with an offset. Most of this offset will be due to sequence selection and transformation. To yield the most science, these differences should be examined and fixed.
For T CrB, what I worry most about are measures that are brighter than the mean curve. For example, there are a cluster around 2460938 that are a magnitude too bright. A researcher seeing those might assume that T CrB was going into outburst.
Finally, to show you another example of why multifilter observations are important, here is the U/SU lightcurve:
Notice that it doesn’t look ANYTHING like the other bandpasses. Of course, there is a lot of scatter. Much of this could be reduced by submitting averaged data instead of individual measurements. Still, you don’t see the ellipsoidal variation. This is because you are far enough into the blue that the majority of the flux is coming from the white dwarf and not the red donor star. Some of the short-period variation is probably due to flickering. I wish the photometric quality was better, so that we could really believe the structure. Note that U and SU are pretty similar, though the Sloan u’ bandpass is narrower and bluer than the Johnson U and therefore picks up even more of the white dwarf emission. U is a difficult band for many reasons, but I think SU is better defined and useful even at lower elevations.
One way to decide whether the scatter is real or due to poor photometry is to look at the check star. If an observer has quite a bit of submitted data, such as WGR for the SU data, you can use the check star’s scatter as a surrogate for the target. For the target, the differential photometry formula is
T = (t-c) + cstd
where T is the target standardized magnitude, t is the target instrumental magnitude, c is the comp star instrumental magnitude, and cstd is the comp star catalog (standard) magnitude. The same formula holds for the check star:
K = (k-c) + cstd
where K is the standardized magnitude for the check star and k is the instrumental magnitude. If you plot the SU standardized magnitude for the target and the SU standardized magnitude for the check, you can see if the target’s scatter exceeds the uncertainty values for the check. WGR did this for me and the target’s scatter looks mostly real. It is a Best Practice to include the check star values in your report. For classical differential photometry where you just have a single comp, the AAVSO requests that you enter the instrumental magnitudes for the comp and check. Then you can use the formula above. For ensemble photometry, the AAVSO requests that you enter the check star’s standardized magnitude, which is calculated with the same ensemble that you used for the target. You can download the check star values for submitted data by using the Webobs download tool.
T CrB is actually observable year-round from many northern hemisphere sites. It is low on the horizon near conjunction, but still doable. That is why the light curve is continuous. However, low-altitude photometry is tricky. You really don’t know the airmass very well. There is differential airmass and refraction across your CCD field of view that has to be taken into account. There may be distant unnoticed clouds that are low on the horizon. Seeing degrades near the horizon. So take extreme care if you go below airmass of, say, 3.0. I strongly suggest that you take multiple sets of observations and form average and standard deviation before submitting low-altitude observations to the AAVSO.
T CrB is like many other variables studied by the AAVSO. If you have a relatively wide field of view, you will find several other variables that will be simultaneously imaged. Don’t just blindly submit these measures. Just like the target star, each serendipity measure needs to be inspected. Some of the extra stars will be brighter than T CrB, and so you need to worry about saturation. Some will be on the edges of your field, and may suffer residual vignetting or bloating of the image due to optical aberrations like coma. Such issues will often make the extra star measures brighter or fainter than reality, and you should consider avoiding such submissions. It is a Best Practice to treat serendipity targets exactly the same way you treat the primary target: inspect their light curve.