Skip to main content

V0645 AUR: Early Onset of the Eclipse? And Recommendations about Best Observing Cadence?

15 posts / 0 new
Last post
V0645 AUR: Early Onset of the Eclipse? And Recommendations about Best Observing Cadence?

Hello! I've been trying to get several nightly observations each week of EA V0645 AUR.

    Last night's photometry showed a drop consistent with the eclipse, but it was 2.2 days early according to the ephemeris in VSX.

    During the last 6 months, I've only caught two magnitude drops consistent with eclipses as well as some irregular increases in magnitudes (most pronounced in I).

    1) What might have caused the drop in magnitude 2.2 days early that I saw last night, 04-04-2020?

    2) What is the best way to observe this variable? Should I obtain multiple images/magnitudes each night since the duration is listed as 0.44 days during each 10.8925 day period?

    3) Could the components themselves be variable given what appears to be a wavy baseline in between eclipses?

    4) How can I check the ephemeris for past eclipses through VSX? I'd like to check my observations on 12/20/2019 in relation to the ephemeris.

    Thank you and best regards.


BGW's picture
bad elements?

Hi Mike,  All the following assumes your last night's photometry is good.  Does the actual portion of hte light curve from last night fit the lightcurve from another eclipse night.  Magnitudes, rate of change?

I just went to the ASASSN data download page:

It plots a very scattered-looking graph phased light-curve graph, but their software used a period of 542+ days.  That often happens on smaller amplitude EBs.  Their automatically-determined (?) period can be quite wrong.  So I downloaded the data to a spreadsheet and made a phased plot using the VSX elements (epoch, period). Result:  also a scattered graph.  This suggests the elements (the period, actually) are/is wrong.  So one approach for you would be to do period analysis on the ASASSN data to come up with a new period and epoch.  An advantage of this would be it would based on recent data, so even if the period is of low precision, you'll be able to predict eclipses well enough to plan observations.

Another approach:  I went to the BRNO site, and it gives an O-C diagram that includes some recent observations by others that fit the VSX elements pretty well:

So that implies the elements are pretty good!

This leaves us with a quandary!  When I guesstimate a ToM from your report, of 245 8744.75, and enter it to that website as "user data", it indeed gives an O-C of -2.087 d, which agrees with what you said.

(Enter it as:   2458744.75 ccd p   )

BTW, you can use the same process with your Dec 2019 ToM, to see where it lands on the O-C. 

So, what I would do is work with this data as much as you can and also download the available from NSVS, ASAS-3, Hipparchos that Otero et al used in the 2006 IBVS 5681 paper to derive the elements in VSX.  Re do the period analysis yourself, try to reconcile it with the ASASSN data, is likley about 20 years later.  

A purely observational approach would be to observe every night you can, at the beginning and end of each night, to try to find another eclipse or two.  Then start "homing in"

I would also especially plan future observations that straddle last nights pseudo ToM, plus multiples of the VSX period, and at the offset of the secondary eclipse ( 0.789 x period ), and its complement ( [1-0.789] x period  ( I've seen that error in published data sometimes ).

Something funny is going on:  when you unravel it, you'll have added more knowledge, more than if the star had come in just when expected!

I hope this helps,     Gary Billings


P.S.  yes, wise to keep in mind variabilty of the components, but note the depth of the eclipse is supposed to be 0.5 mags, which is large compared some of the types of variability usually encountered in such cases.

Eric Dose
Eric Dose's picture
But "Early compared to what?"

But "Early compared to what?" is a question.

The VSX min ephemeris dates from 2003. Happily, a more recent observation from Feb 6 2020 (yours, I think) already confirms a more recent minimum that is also early compared to VSX's 2003 recorded minimum JD. The phase has shifted earlier by a day or two, or possibly later by 3-4 days if 50% out of phase (1'min vs 2'min).

So you'll want to recapture the phase, with observation JDs guided by the more recent min observation. You'll want to observe both minima to distinguish 1'min from 2'min, starting with good coverage around the JDs projected from the period (which probably isn't completely wrong) and from that most recent (Feb 6 2020) good min observation.

Updating the ephemeris is important work, and when you've succeeded you'll want to notify VSX. If you follow up again this year and once or twice next year, you might also be able to update the VSX period  as well (which clearly isn't exactly right for the most recent 18 years).

V0645 Period Questions

Thank you! I get good SNR between 45 seconds for V, 90 seconds for B, and 3 min for I - an unusual cadence since I-filter typically has the shortest exposures of the three filters for my system for other variables.

    V0645 AUR is nearing the end of this observing season - it goes below 30 degrees west at about 11:00 PM. With my system. Perhaps I should concentrate on V0645 AUR with this every 5 minute cadence for the first 2 hours of observing time for the next month in order to get better granularity for the data?

    That would give me the chance to catch 3 eclipses, even though with such a short window, there is a good chance of missing them.

    However, that would also help identify variability of the components (if any) other than eclipsing variability.

    Would this be too much data for LCG given its possible period(s) and may exceed the standard 50:1 observing cadence? Best regards.


BGW's picture

While just trying to confirm when eclipses occur, you don't need fine sampling.  As Eric said, 1/hr.  A primary min lasts 0.44 days, according to VSX.

If your goal is to get times of minimum, try to nail a minimum in one intense observing run, or two stitched together.  You'll get a more accurate time of min.  Then come back a year later.  

Once could talk over several beers about such things...  but EB times of min are usually done with one intense time series, capturing at least a couple of hours on either side of the min.  My cadence is determined by the expsoure time required to get a useful time of min.

I have misgivings about multi-colour time series of EBs.  I've seen a completely convincing case where the different ToMs reported in different colours were real (from Gil Lubke, in a light curve distorted by spots, so the ToM was displaced to varying degrees depending on its temperature and the filter band), but in general I see different ToMs in the different colours and I am suspicious that it is due to using multiple inferior times series, instead of one better one.

Of course, I am generalizing:  if someone wants to use the data for something other than times of min, then something different could well be appropriate.  E.g. very little observing is done to monitor long-term intrinsic variability of EBs.  For that I'd suggest tranformed photometry just at phases 0, 1, 0.25 and 0.75, once per year, but do that for as many stars as you can.  That would be more likely to turn up something useful.  Unless you have a particular reason to look for such variability in one specific star.  It is a big topic, and I'm being very brief here...

Others may disagree!


Eric Dose
Eric Dose's picture
Phase-capture cadence vs Lightcurve cadence

If you start by trying to recapture phase of this eclipser, you don't need rapid cadence at all. V0645 Aur's period is given as about 10.9 days, meaning that its 1' minimum probably lasts 3-5 hours. So for phase capture, if you image it once per hour--or per half hour if you'd be more confident--you definitely won't miss the minimum between your images. Since V0645 Aur's phase has become so uncertain, you'll need to capture both minima (or show that 2'min is absent) to be sure you're not off by 1/2 phase, i.e., mistaking 1'min for 2'min and vice versa.

Once you nail down the phase, then you can refine the minimum lightcurve with as fast a cadence as you please; if the minimum lasts 5 hours, then yes 8-12 images per hour would be reasonable. (Until September-October you'll probably have to splice together parts of the minimum lightcurve from different nights.)

Thank you! Mike

Thank you!


Sebastian Otero
Sebastian Otero's picture
Ephemeris was okay

Hello all,
prompted by these messages I went and checked the system's ephemeris, using the newly available ASAS-SN data in combination with ASAS-3 and NSVS observations.
Everything fits well. You observed the secondary eclipse which falls at phase 0.79 and not the primary eclipse predicted by the ephemeris.
I improved the period a little bit but it's not significant. I did find that there is apsidal motion in the system so I added an ephemeris for the secondary eclipse in the star remarks. New phase plot is also available through the VSX star page.

Sebastian (and his kidney stone :()

BGW's picture
well done

Well done, Sebastian.  I should have figured that out!

When I enter Mike's datapoint of 245 8837 in to BRNO O-C page:

(as user data "2458837.802 ccd p"), it plots at O-C of about -0.1 day, i.e. 2.4 hours away from the minimum.  That makes sense.

Now that Sebastian has pointed it out, if I enter Mike's recent datapoint as "2458944.61 ccd s", i.e. state that it is a secondary eclipse, then it plots at 0.07 d on the BRNO O-C diagram.  Again, this is fine for a "random" datapoint taken during a secondary eclipse.

So, problem solved!

Gary Billings



Thank you!


Didn't see any

I did a long stare last evening and V0645 Aur was not eclipsing as expected. However a couple others were very active:

TYC 1890-1262-1 and TYC 1890-384-1 were in my FOV of V0645 Aur.

BTW, both my focus and airmass got away from me on the last quarter of this run, so the last 8 or 10 data  points are not the best. I will try to get more on TYC 1890-384-1 during the next month to attempt a phase plot. Also, get a bunch more information about V0645 Aur.


Field of View?

Hello! Just curious. What set-up do you use? Your FOV is quite a bit larger than my 13'x20' field of 8" LX200 and ST402.


12" F8 SCT shortened to F5.63

12" F8 SCT shortened to F5.63.


Caught the Eclipse

Hello! I caught the eclipse last night as the ephemeris predicted. Best regards.



I also added some data points to the pile.

Also caught some of the NU UMa eclipse.That prediction may need tweeking.

Log in to post comments
AAVSO 49 Bay State Rd. Cambridge, MA 02138 617-354-0484