Skip to main content

TYC 3973-2396-1 Transit LC

51 posts / 0 new
Last post
Roger Pieri
TYC 3973-2396-1 Transit LC

The discussion about this transit of TYC 3973-2396-1 has started under the " EE Cep Still Fading, Capture TOM " topic of the " Campaigns & Observations Reports " forum. I propose to continue here under a more specific title.

TYC 3973-2396-1 was just one of the stars of the sequence prepared for the EE Cep campaign referred as "117".
On Sep 3rd that star revealed variable and the light curve I got was a lot looking a transit one. It has been withdrawn from the sequence. Several of us started searching past data, images and observing. Now we have a good idea of the ephemeris of that binary transit.

Yesterday I finished to process a large stock (~600) of past images of EE Cep and was lucky enough to find one more transit. It's very interesting as it covers the ingress that was missing. It essentially confirmes the former estimate, with just a small shift of the mid transit and no significant period change.
The period has been adjusted to minimize the RMS error of the fitting. It's sensitive to 0.0001 JD.

The result is attached.

Clear Skies !

Roger (PROC)

CQJ
More Data

Hi Roger,

I recently added my last 5 data points in each of 3 colors for the Sept 18 eclipse.  These help fill in the mid-eclipse data.  I am now starting to submit some new data from the eclipse last night (Sept 22).

John Centala

Roger Pieri
Sep 22 ingress

Hi John,

Very interesting, that sharp ingress combined with my past one will provide an accurate reference for the period. I will do that tomorrow (here it's late in the night and the sky gets hazy...)

I just uploaded 158 of my observation sessions, each session result is the mean and SD of the mean of 10 images, 32 sec ISO 400 / EOSM and 8"F4 scope. Color correction made by VSF.

Clear Skies !

Roger

CQJ
Final Data Added

Roger, Sebastian,

I just added my last data points from the Sept 22 eclipse.  My mostly manual data process is too slow for so many images.  Each data point is for the average of 5 consecutive exposures of 1 minute each.  The eclipse was ideally timed for this location.  Note that there is a gap in the egress data.  My telescope lens dewed up on the surface inside the tube, and it took about 40 minutes to get it dry.  The telescope is a 130 mm F 6.4 Astrophysics, and is very nice, but not optimum for multicolor photometry.  I notice that there is a slight drop in light just before and just after the eclipse.  Is this traditional, or is there something wrong here?

John Centala

Roger Pieri
Superb !

I downloaded your data, it's a very nice result John. I will integrate it in the fitting. I am surprised Sebastian found a significantly different period, I checked if the JD instead HJD made a difference but it's just the same as I expected for those two months only of observation. Your new data will tell us.

I have also seen the drop before the eclipse in my data on 6899 but judged it was some scatter ! Now it looks real. The only explanation I would see is there is some ring, shell, there ? We have to continue the observation, sure.

Here I should have a full eclipse in the night of 29/9, if the weather permits... We have had a nice indian summer in past weeks but I fear it's the end now, clouds coming.

Clear Skies !

Roger

WGR
WGR's picture
Uploaded more data

Hello

I have moved my EE Cep center of image, to allow imaging TYC 3972-2396-1 (aka 117).  I have put 3 nights of data on the web page.  None of the nights were eclipse nights.  At Sebastians suggestion, I will be imaging every clear night looking for a secondary eclipse.  

 

Gary

WGR
WGR's picture
Min Temp of Secondary Star

I did a little simulation this morning.  Using the measured value for B-V= .25 mags and eclipse depth of 0.127 mags, and the fact that no secondary eclipse has been detected, I was able to calculate/estimate that the secondary star must be hotter than 10,000 K.  I did this by generating my own spread sheet to simulate the light curve, and ran several cases, and that's what came out.  Pretty interesting.  Anyone have a similar calculation or a paper with an extimate of the secondary color/temp?  If the Temp of the secondary is much less that this, we should be able to detect it in the light curve.  I do see something at phase = 0.5 of about .012 mags, but its sortof within the error bars. 

 

Gary

Roger Pieri
Secondary ?

Hi Gary,

I have observed several times at the supposed secondary time of the 3.867 d period and saw just nothing. But there is another hypothesis.

After my first observation of the eclipse I noted in my alert to the EE Cep team that the LC bottom looked well "flat", then we got several LC either from new observations or extracting it from past observation. At that point new bottoms looked much less flat, significantly "rounded" with a visible minimum.
The recent and excellent observation from John (CQJ) of a full eclipse, at my surprise, looked having a very flat bottom too ! We both observed the next one that looks "rounded". At same time I got an advise from Thom (GTN) considering the periode could be twice the 3.867 one: 7.734 d.
That would mean we are seing both primary and secondary of similar amplitude, mixed ! (~0.12 mag). That supposes both stars should have a similar surface luminance, the "flat" bottom eclipse being the secondary and the "rounded" one the primary.
This hypothesis needs more observations of the "main" eclipses to be confirmed. The related difference of both shapes of eclipses is of the magnitude of the scatter, even if it repeated in a couple of observations it's not surely real. It also needs advise from specialists (I am not !).

Anyhow I strongly recommend to continue to observe the "main" eclipse(s) about 6903.4055+3.867 x E.

Attached is both shapes of the eclipses extracted from observations of John and I. Sorry for the upside down mags ( my graph tool is not an astronomer one ! I have to make some change to its code... ).

Clear Skies !
Roger (PROC)

WGR
WGR's picture
Secondary Eclipse

Hello Roger

Interesting theory--both eclipses visible but at a period of 7.72 days.  I agree that the eclipses need more observation to substantiate the flat bottom vs the rounded bottom.  I am not an expert in Binaries, this is my first Rodeo, although lots of other ccd observations.  Can you think of a scenario where one would be flat and the other not?  I cannot make a simple model like that.

I did run my simulator for this case, primary and secondary eclipse, with a .127 mag (Flux = .052) depth for both.  So far the only case that works is that the stars are of the same color, B-V = .25 implies a Temp of 7794 K, and based on the primary star with a luminosity normalized to 1, with an area of 1, the secondary has an area of .052 and a luminoscity of .052;  So the secondary is 20 times smaller than the Primary.  I am calling the primary the star that is most stationary, ie has the most mass. 

BTW:  When the secondary star is colder than the primary star, the Primary Eclipse becomes the Secondary, and vice versa.  For all cases with the secondary star hotter than the primary, the primary eclipse is when the secondary is behind the primary  (Normal case I believe).

This suggests to me that I should change my cadence.  I have been doing BVRI, perhaps I should just do V until we resolve the flat bottom question.  This will also allow me to push the SNR a bit more, if I have only one filter to worry about.  I am going to shoot for 45K adu's.  Most of My observing is about 30K.  It looks like it might be clear tonight.

Gary

Roger Pieri
Flat vs rounded

Hi Gary,

It's also my first experience with binaries even if I observe exoplanet transits from time to time. The explanation for flat vs rounded could be the following:

- When the small star "transits" behind the large one there is no variation of luminosity during the event resulting in flat bottom LC
This is what I call "secondary eclipse" as it's the usage with exoplanet, but maybe the tradition is different with binaries ?...
- When the small star "transits" in front of the large one, the total luminosity varies in function of the surface luminance distribution of the large star. The center being more luminous than the limbe results in the "rounded" bottom.

The fact both bottoms have about the same depth could be explained by a similar surface luminance of both stars. (Here I say "luminance" as a physicist, cd/m². The luminosity of astronomers, mag, is then function of the surfaces involved)

Maybe binaries specialists would have some other explanation ? Anyhow something very interesting to explore making more observations !
Regretfuly here in western Europe the autumn weather risks to seriousely limits my observation opportunities !

Clear Skies !
Roger

WGR
WGR's picture
Flat Bottoms and limb darkening

Ah, yes Roger

That makes perfect sense.  Limb Darkening as astronomers call it as the star passes in front, and complete obscuration when its in back.  That would fit.  Its clear here tonight, sadly, no primary or secondary eclipses, but we take data anyway.  Going to do V only.  Dry run for when we get the clear skies and the eclipse action.

 

Gary

Tonisee
Gary, you could defocus the

Gary, you could defocus the stars just a bit to collect enough photons, to get SNR>1000 for both target and main comp stars. While the target is not an extremely bright one, scincillation is most probably not going to be a problem because of exposure times of several tens of seconds. But if you are concerned about possible "gray area of CCD nonlinearities", defocusing helps to get enough signal while it's level is still inside the CCD's linear ADU region. In that sense, getting really neat binary lightcurves is exactly the same business as detecting negligible exoplanet transits...

Tõnis

WGR
WGR's picture
Defocus

Thats another good idea.  I have done that for exoplanets myself.  It would be helpful here to avoid the scintillation and boost the number of photons without gettting into the grey non linear region.

 

Gary

Mark Blackford
Mark Blackford's picture
Flat vs Rounded

Hi Roger,

I've been following this thread with interest. Congratulations on finding this interesting binary. Rounded minima can be due to the orbit not being perfectly aligned with the Earth (incliniation less than 90 degrees) leading to partial eclipses. If one star is smaller than the other it is possible for one eclipse to be flat bottomed (i.e. total) while the other is round bottomed (i.e. partial). Cheers,

Mark

WGR
WGR's picture
Flat vs Rounded & combining BVI

Hello Mark

Sounds like you have some experience in this.  Thanks for following and posting. 

I can visualize what you are saying for the case that the orbit of the secondary is quite eliptical, and the long axis of the ellipse is more or less along our line of sight.  This means that the secondary could occult or be eclipsed by the big primary when its orbit is close to the primary star.  When the secondary is far from the primary star, and with the proper angle of inclination, the encounter could be partial, or not at all.  This could give conbinations of flat and rounded bottoms.  Is this the case you were thinking of?  If the ellipse precesses around the primary star, the round bottom and flat bottom could even change places. 

John, great idea.  Can you post your phase curve?

Gary

CQJ
Combined BVI

Gary,

Attached is the plot in pdf form, I hope.  This is my first try at an attachment here.

John Centala

WGR
WGR's picture
Combined BVI

Hello John

That's a pretty convincing plot that it was flat bottomed for that one.  It also shows the "horns" just before and just after eclipse.  Nice work.  

I appologize for not mentioning this before, and I don't think it will change the result, but technically these magnitudes should have been converted to fluxes, before adding/averaging.  You can do this by taking exp(mag/2.5)/exposure.  This assumes that all the apertures of pt were the same.  This will give you a represenative flux.  I don't think it changes the result, but it would be picked up by a referee if it were attempted to be included in a paper.  I suspect this is unique enough that someone will write this up into a paper, with contributors like yourself, as co-authors.

Gary

Mark Blackford
Mark Blackford's picture
Flat vs Round

Hi Gary,

I've been observing eclipsing binaries for a few years but have only very limited experience really in interpreting EB light curves. The scenario involving an elliptical orbit sounds reasonable but I don't know if it applies to TYC 3973-2396-1. If Roger's suggested 7.734d period is right then the primary and secondary eclipses look to be exactly 0.5 phase apart which suggests circular orbit or, much less likely, elliptical orbit with major axis pointed directly to earth.

My targets are usually much shorter period and often are tidally distorted so the light curve is continuously varying. Some of these show flat and rounded bottom minima, like AE Phe (see attached) where the primary at phase 0 is flat bottomed and the primary is rounded. In this case the explanation Roger gave on October 6th is correct.

TYC 3973-2396-1 looks to have steady brightness between eclipses and the long period suggests wide separation hence no or very little tidal distortion. Roger's explanation may still apply if the orbital inclination is such that the smaller star passes close to the top or bottom of the larger star when transiting so limb darkening accounts for the rounded bottom.

But the case for different eclipse shapes still needs to be confirmed with more high quality observations at high cadence during both eclipses. I look forward to seeing further progress from you guys north of the equator. Cheers,

Mark

Roger Pieri
Flat vs Round

Hi Mark, nice to read from you !

Sorry for my late answer, I was away those days. I am ok with your interpretation and the answer from Gary, this is another interesting interpretation. I am not sure we could determine our choice soon, it will certainly need much more observations.

Anyhow the combined BVI curve from John is impressive, and this with a depth of only 0.12 mag ! I will process my DSLR RGB the same way, I usually do it for exoplanet and the scatter is significantly improved.

But compared to nearby stars, I suspect this star shows a significantly larger out-of-eclipse scatter and particularly during the round eclipse. This activity could significantly limit our accuracy.

Here the autumn sky is usually very cloudy and I would probably have few opportunities to observe, even if I am well north of the equator !

Cheers,
Roger

CQJ
Another Plot

Roger, Mark, Gary, and all,

Attached is a new plot comparing the Sept 22 eclipse with the data I took on Sept 18, processed in the same way to average the B, V, and I data.  The Sept 18 data is shifted by 0.133 days (= 4 - 3.867) so that the two curves are superimposed, but with one in red and the other green.  The Sept 18 data is obviously noisier, but I still consider it to have a flat bottom rather than the curve posted earlier by Roger that showed clear rounding.

John Centala

CQJ
Shape of Eclipse Light Curve

Roger, Gary, and all,

I'm not too sure about the shape of the eclipse bottom; the data is not convincing yet.  Roger, I see that you have used only my V data in your light curve.  The noise on the V data can be reduced by combining it with the B and I data for the Sept 22 eclipse.  I averaged (B-0.23) and (I+0.29) with the V magnitudes.  I took the measurements in the order of V, B, and I, and processed these groups of three to get the color transformations.  So take the average of the first V, B-0.23, I+0.29, then the second V, B-0.23, I+0.29, etc, and you will get a smoother plot.  Ideally the time of each data point is that of the B data, since this is the midpoint.  The last V data point should be combined with the previous B and I data points, since it was processed that way.

John Centala

 

WBY
WBY's picture
B-V and CQJ Light Curves and sizes

Guys, 

I am a late contributor but have been following this discussion. Is there a contradiction between Gary's 10/4 post postulating stars of the same color (approximately the same temperature and John Centala's light curve? Don't pre and post eclipse "horns" (inappropriately named reflection effect) occur when you have two objects of quite different temperatures so that you see the increased surface temperature of the object being eclipsed due to heating by the hotter object? This is something, for example that you see with "hot Jupiter" exoplanets. I am not suggesting that this is an exoplanet phenomenon since the transit depth is much too deep at 120 millimags. If the two bodies are of very different temperatures but the eclipses are of essentially the same depth, doesn't this argue against a primary and secondary eclipse in favor of an EA type eclipsing star?  I am certainly no expert on binaries. 

More data will settle the issue, but the horns just before ingress and just after egress seem to be very pronounced and I thnk that indicates objects of very different temperatures. I suppose the geometry could be just right that a larger brighter object completely obscures a smaller dimmer one and the smaller dimmer one only partially obscures the larger brighter one so that the total flux difference of the primary and secondary eclipse are the same and we see the reflection before and after the complete eclipse of the secondary star and not before and after the primary. If sequential eclipses show the reflection effect then I think it would be definite that there is no secondary eclipse. 

Does this make sense or have I missed something?

Brad Walter, WBY

Roger Pieri
Horns

Hi Brad,

If I understand well what you call "reflection effect" (never seen that wording here in EU) is the increase of flux that could be seen before and after exoplanet secondary transit ?
Usually such increase evolves in between the quadratures and the transit. Given the geometry imposed by the depth and the period we observe this should extends for about a day or so. This is not the case of the horns that spend less than one hour.
Then I am not sure we always observed those horns. I also got it in one of my observation but not the others. I got data from the two last eclipses but I have not finished the processing that is somewhat complicated by the haze of the 6957 one. The fully covered 6961 is much better and shows no horn.

Is it what you mean ? any other possibility ?

Cheers,
Roger

Sebastian Otero
Sebastian Otero's picture
Improve ephemeris and some discrepant observations

Combining all the data available from seven AAVSO observers with NSVS data, I get the following accurate elements:

HJD 2456899.538 + 3.86721 d. x E

GFRB observations are 0.015 mag. fainter than the observations from the rest of the observers.
AAM observations are 0.035 mag. brighter.
All others are consistent with each other so their zero point has been kept.

Range of the eclipses is V= 11.65 - 11.76.

Eclipse duration is 7%.

Mostly all data agree but some points by AAM (Andrzej Arminski) are at maximum while others are at minimum during the same eclipses.
So it is not a problem with the period, something one might think at first sight.

Andrzej, can you check your images for the observations detailed below? (keep in mind that the JD has been converted to HJD and your data was shifted 0.035 mag. to the faint side).

HJD-2450000  V  orbital phase

6864.6352    11.709    0.9741504
6864.6377    11.693    0.9747995
6864.6403    11.701    0.9754485
6864.6428    11.720    0.9760950
6864.6471    11.736    0.9772148
6864.6502    11.728    0.9780319
6864.6687    11.664    0.9828056  inconsistent

6868.6235    11.670    0.0054443  inconsistent
6868.6257    11.659    0.0060313  inconsistent
6868.6280    11.664    0.0066183  inconsistent
6868.6354    11.657    0.0085397  inconsistent
6868.6458    11.681    0.0112265  inconsistent
6868.6565    11.768    0.0140013
6868.6588    11.762    0.0145883

I attach two plots. One showing the primary eclipse and the other showing the complete light curve.

The NSVS data have a lot of scatter (larger than normal) but are enough to improve the period by increasing the time baseline from some weeks to 15 years.
AAVSO data show that the eclipses are total.

There is no sign of a secondary eclipse.

After checking the above observations we can update the VSX entry.

Cheers,
Sebastian

christopher lloyd
Colour plots

Hi All,

I have just been through all the available data again and produced a whole load more plots, some of which will look very familiar now. The plots are warts and all, so no effort has been made to clean them up and they highlight that some of the data are very noisy.

So, the plots... BVRI_phase, BVRI_phase_Pri and BVRI_phase_Sec are all the data showing the whole phase diagram and the bits around primary and secondary minimum, well phase 0.4 - 0.6. There is no sign of the secondary at any phase.

I've also looked at the primary minima from alternate cycles in V_2PphasePri and there is no difference. To all intents and purposes the eclipses are identical.

On all the primary eclipse plots there is some rounding at the bottom but the eclipses are essentially flat so this has to be a total eclipse. The rounding is probably due to limb darkening as a smaller secondary transits the primary.

The colour data are a not as complete but I've looked at what we have in B-V_phase, V-R_phase, V-I_phase, with the V data in V_phase. The combined plot around primary minimum is in V_phasePri. There is no perceptible change in colour through the eclipse.

If we accept that the alternate eclipses are identical then the period is 3.867 days and not twice that. If the eclipses are total then we are not looking at two similar stars but two that are very different. The dominant light source in the system is a B8V star so if the secondary eclipse is 0.01 magnitudes then the secondary contributes 1% to the luminosity of the system. Anything less massive that a K0V star would do this. The other possibility is that the primary eclipse is caused by the occultation of something small and hot, perhaps a B subdwarf, but the boring K star is probably more likely. I’m doing a photometric solution on this so that should provide better limits on the nature of the secondary.

Regards,

Chris Lloyd

 

Roger Pieri
Data scatter

Hi Chris, all,

I think there is a problem with mixing data. When we look at the best possible mean mag of a star it's certainly ok to do so. But looking at small variations it's better to use data that are coherent, accurate, even if not precise. The mixing of number of observers data generates a lot of scatter that makes small variations unreadable. That "mixing" scatter is due to various reasons: comp stars not the same, filters not the same, vignetting, local conditions... We should remember the Nova Del 2013 experience !

That all data mixing plots have too much scatter to differentiate possible primary and secondary in the 7.734 hypothesis. When looking at coherent data there are several characters that are somewhat different and repeat in a couple of cycles. I fully agree that the 7.734 hypothesis is not for sure at all and the 3.867 solution is also possible, but I think it's far too early to decide.

There is also the possibility various flares, spots, and other behavior affect the shape of the bottom. All this could be solved only with more observations, more accurate if possible, regretfully here the season is not the best for !

Cheers,

Roger

bdg
bdg's picture
Spectrum of TYC 3973 2396

 

Here's a spectrum of this star taken last night. It looks like a normal spectral type A7V star.

Regards,

David Boyd

 

bdg
bdg's picture
Spectrum of TYC 3973 2396

 

The link in my previous message didn't work. Here's one that hopefully should work.

David

Roger Pieri
Spectrum & B-V

Thanks David,

this much more in agreement with a 0.27 B-V ! Simbad says it's a B8V which should be about -0.100, I suspect they have some problem here.
I compared it with a Pickles A7V and you are right, it's very similar. I only see some difference of continuum about 4600 A, there are also some differences after 6800 A but I suspect it's telluric ?

It would be very interesting to determine the radial velocity cycle of that star, but I wonder if we could access it with amateur means ?

Cheers,
Roger

Roger Pieri
Simulation

I just finished a simulation based on an A7V star, the result is attached. I checked my simulation with the Nebraska-Lincoln U simulator, we are perfectly ok. The simulation is based on masse and radius from the literature for an A7V star. Le limb darkening is such of the sun, the only I found in the literature.

In fact the system has a lot of constraint, ingress and egress slopes, depth, balance of depths between eclipses, lenght of the eclipse... In particular the orbit has to be tilted. I have only explored circular orbits.

From the various constraints the 3.867 period seems incompatible with the eclipse lenght. The 7.735 is just right. But I have not tested if an elliptical solution could work.

The orbital velocity of the large star is not well constrained, the order of magnitude should be a few km/s, probably difficult to catch at mag 11.6 .

Cheers,
Roger

GTN
GTN's picture
Spectrum and B-V

Hello David and Roger, et al -

 

 You have to be careful when inferring spectral types from B-V, and vice versa.  TYC 3973-2396-1 is highly reddened - Barbier et al. (1973, A&A, 27, 421)  find the B-V color excess to be 0.29.  The unreddened B-V therefore corresponds to a spectral type no later than A0.

 

David, did you determine the spectral type by comparison to the Pickles models?  If so, they must be reddened - or your spectrum must be dereddened - to compare to your spectrum with them.  The standard classification procedure is to take a bunch of spectra with the same gear you used to take the unknown and then to compare the spectra until you find the closest match.  Anything else is, at best, a guestimate, usually based on incomplete or missing information.  If you can get a higher SNR spectrum that includes the Ca II K line, it would tell you the spectral type right off .  At A7, the K line is nearly as strong as the hydrogen lines that flank it.  In late-B or early-A stars, that line is much weaker than the H lines.  Also check for He I lines.at 4471 Å.  If you see any lines from it and the spectrum is otherwise normal, the spectral type cannot be later than A0.

 

At what phase was your spectrum taken, David?  May I suggest you try to get spectra around quadrature?  With sufficiently high resolution and SNR, two sets of spectral lines will be visible; if they are, the 7-day period is the correct one.  You really must know the true period before modeling even begins to be meaningful and in this case spectroscopy could quickly eliminate one of the suspected periods.  Working backwards from a model in order to determine what the correct period is, is doing science backwards.  For example, any synthetic light curve based on a temperature for the primary star that assumes the spectral type is A7, will almost certainly fail to fit the actual light curve.  The problem is made even worse if B-V is used to estimate the spectral type.

Interesting results, guys!  I've had a fascination with late-B star binaries, especially eclipsing ones, since my undergraduate years in college.

 

Cheers,

Thom Gandet (GTN)

Roger Pieri
Spectrum / Synthetic LC

Hi Thom, all

Thanks for the advise Thom, I did check the spectrum against Pickles but completely missed the obvious K line difference ! (too much focusing on the continuum, as usual... )

For sure a spectroscopic observation of radial velocities would be the best way to solve our problem but I am not sure it's in the amateurs capability. The orbital velocity of a small secondary is of the order of magnitude of 100 km/s, if we observe the primary and secondary lines at one quadrature we would need a resolution of 5000 at the very least and probably more. That needs a LHIRES, not a LISA. But the secondary is dim, at best mag 14 plus the extra (x10) dispersion and the mix with primary, I think it's out of amateur telescope capability.
Another way is to compare the two quadrature shifts of the primary, it would require a good calibration, this would work only if the secondary has a high mass, at low mass it's probably not accessible to amateurs too.

The synthetic LC is probably more questionable for a B8V than with an A7V I agree. From the literature the spectral type/mass/radius relation seems to be much better defined for the A stars. Thom, Is it what you refer too ? Or any other problem ?
Otherwise that technique is commonly used for the exoplanet studies and we are not so far that situation. We know the secondary is small and the eclipse is total. The luminosity of the primary, by itself, has not a lot of impact on the simulation, overall it's more a question of geometry.

Gary recently reported new observations and there is something visible in the V LC on 6939.51, about at phase 0.34 of the 3.867 period, not much visible in the B LC. That supposes an elliptical orbit, any idea about it ?

Cheers,
Roger

GTN
GTN's picture
Radial velocities, periods, and Pickles

Hi Roger and all,

    Good questions!  I don't know much about exo-planets, but the orbital mechanics are the same for binary stars - straight-forward two-body problem, except that for exo-planets you know which object is the most massive of the two bodies.  Terminology is slightly different, though.  "Primary minimum" (an "eclipse", used to distinguish that stars are meant; planets "transit") is the time at which the cooler ("secondary") star eclipses the hotter ("primary") star.  The deeper minimum thus occurs when the fainter, cooler star is between you and the hotter, brighter star; how much deeper depends on the geometry, stellar radii, temperature difference.  I'm going to talk only about eclipsing binaries with orbital periods between about 1-15 days, circular orbits and a moderately massive primary (an A0V star).  I haven't had to teach this stuff in decades, so I hope it's helpful.

    The simplest case is one in which the two stars have the same temperature and luminosity and the orbit is circular and the inclination of the orbital plane is close to 90°.  Each eclipse will thus have the same depth, and, according to the mass-luminosty relation, the masses of the two stars are virtually identical to one another.  But to know that for certain, knowing the correct period is crucial - if there are two minima having the same depth, then it is virtually certain that the spectrum of each star can be seen at a high enough resolution.  If secondary eclipse is undetectably small, the temperatures of the two stars are very different and the spectrum of the secondary won't be seen (maybe in the IR, though).  There are eclipsing binaries whose periods and time of minimum make it uncertain whether the period is P or 2P!

     We want to know that masses and thus need radial velocities to complete the picture of an eclipsing system.  Kepler's third law can be used to compare what the radial velocities would be under a few different, but reasonable, assumptions.  (You don't need to know the orbit's eccentricity and its geometric orientation with respect to the earth to do that.)  However, reasonable masses for the stars need to be plugged in to the equation.  A B9V star's mass is roughly 3.5 solar masses (B8 is 3.8 and A0 is 2.9 solar masses; as you noticed, the mass-luminosity relationship is almost degenerate earlier than about A0!).  Solve for the semi-major axis will get you the orbit's circumference, divide that by the period to get the circular orbital velocity, then double that and out pops the maximum velocity difference between the two stars (at quadrature, in my example).  Knowing the resolution of your spectrograph will then allow to estimate if the spectral lines can be seen as double.  The Mg II 4481 line is generally one of the sharpest lines around A0, and it may show any double lines much more easily than the hydrogen lines will do.

     One thing I forgot to mention before is that any model spectrum (like Pickles's) should be "spun up" to the projected rotational velocity of the star.  Unfortunately, that hasn't been measured for the TYC star yet!  Fortunately, determining the projected rotational velocitiy and MK spectral class is best done at low resolution anyway.

Best regards,

Thom

bdg
bdg's picture
Spectrum of TYC 3973-2396-1

 

Hi Thom,

Thanks for your comments and advice, I'm on a learning curve with spectroscopy so all input is gratefully received. You are correct, I had not deredenned the spectrum. Using the value of E(B-V)=0.29 from the Barbier et al paper I get the spectrum attached. This now looks more like spectral type B8V as listed in that paper.

The resolution of this spectrum is about R=1000. The Ca II K line is not visible at this resolution whereas the H-epsilon and H-zeta (H8) lines on either side of it are clear. I do not see the He I line at 4471 but the Mg II line at 4481 is present which, as I understand it, points to a late B type.

Using the ephemeris 2456903.4055+3.867 x E which I found in a previous email, my spectrum was taken at phase 0.56 relative to the eclipse, so not at quadrature. However, with my resolution, I doubt I would see the spectral line split at quadrature. What do you think?

David

 

GTN
GTN's picture
TYC 3973-2396-1 spectrum

Hi David,

 

     I missed something when I looked at your spectrum the first time.  The Na D lines appear to be cleanly split, and their line centers are roughly 300 km/sec apart.  For the case when the two component stars have about the same temperature, a maximum velocity difference between them of about 100 km/sec could, in principle, allow distinguishing between the two periods.

 

Cheers,

Thom

 

WGR
WGR's picture
Eclipse tonight (Wed Oct 22, 2014 eve, JD 6953

I am at my Remote Observatory, all set for the eclipse tonight.  The weather is not cooperating.  Hope someone else can observe this tonight.  

 

Gary

Roger Pieri
Next eclipses

Good luck Gary !

For me it's too late in the morning, I will try the two next but not sure the weather would cooperate !

Clear Skies !
Roger

WGR
WGR's picture
Rained Out for Eclipse

It did not clear.  Got no data.  In fact, did not even put the camera on the telescope.  Weather radar pretty much was spot on.  I will plan to shoot for Sunday eve here.  We probably only need partial coverage to confirm the flat or round bottom characteristic.  I plan to keep trying.

Gary

CQJ
Clouds

Here in easten Iowa it was partly cloudy on the last 2 eclipse nights, so I decided against trying to get some rough intermittent data.  But it was perfectly clear for the 3 days and nights between these eclipses.

John Centala

WGR
WGR's picture
Weather

John

Apparently you never learned "Murphy's First Law of Weather".

Gary

WGR
WGR's picture
Sunday Night Eclipse, TYC 3973-2396-1

Hello

It cleared enough to observer on 26 October, Sunday night.  It seems I got the ingress right on schedule.  Some issues suspected.  I am using the 104 as a Ref/Comp star and the 119 as the check.  Any comments on those stars?  I am still going.  Hope to get egress also.

 

Gary

GTN
GTN's picture
Sunday Night Eclipse, TYC 3973-2396-1

Hi Gary et al,

 

     I haven't been able to take a look at any of the results on this star since a couple of weeks ago, but am eagerly following the results.  Would you mind posting your results on tonight's data?  Hope your skies stay good!

 

Cheers,

Thom

GTN
GTN's picture
Questionable comp stars?

Hi Gary,

 

     What are the IDs for the 104 and 119 comps?  HD no's. preferred.  Alternately, what is the chart ID?  I'll do a little digging in a couple of hours.

 

Cheers,

Thom

 

 

WGR
WGR's picture
Chart ID

Hello Thom

I am using chart ID 13591DQQ for all of this. I believe the AUID's are 000-BCQ-031 and 000-BCQ-040 for the 104 and 119 respectively.  

I am still observing, but will definitely post the results in the morining (Midnight now)  plan to go all night if weather holds out.

Gary

GTN
GTN's picture
Chart ID

Hi Gary,

 

Thanks for the info.  The chart plotter part of VSP is not generating a chart for 13591DQQ, but it gave me the photometry table.  Unfortunately, I can't find a way to use the AUIDs to find the stars in SIMBAD.  So I picked the two stars in the photometry table that I think are your two comp/check stars and searched the APASS database.  The errors in APASS are around ±0.25-0.50 mags for them, if I've got the right stars. Besides APASS, both stars have only Tycho data.  Sorry I couldn't be of more help.  Some comparison stars do turn out to be variable eventually!  Using comparison and check stars is a wise choice.

 

Hi Roger & John,

 

Thanks for posting more about your work on this star.  On the spectroscopy,Roger,  I will have to find some time in the next few days to read over your posts more carrefully.  John, you guys are getting great data, but observing a comp and a check star, which bracket the variable in brightness, will improve measurement errors.  Extrapolation is like being at the end of a tree branch with no place to jump to.

 

Cheers,

Thom

WGR
WGR's picture
Comp/Check Stars

Hello Thom

You can go to VSP and print out an "F" chart for EE Cep, and the 104 and the 119 stars are identified.  The PT chart can give you the RA and DEC.  I went to Simbad and searched by coordinate identifier, and both stars are there.  104 has 7 alias'.  GSC 03973-02150 and TYC 3973-2150-1 plus 5 others.  Only info is that it is a B8III.

 

119 only has TYC 3973-1261-1

 

Hope this helps.

GTN
GTN's picture
Comp/Check Stars

Hi Gary,

 

    Thanks for the tip about the EE Cep chart.  I was looking up the wrong stars on SIMBAD.  A shockingly high percentage of late-B type stars, especially giants, are variable at the 0.02-0.03 magnitude level and have very strange looking spectra seen at high resolution.

Cheers,

Thom

Roger Pieri
New observation,

HI all,

Here it was possible to observe the eclipse last night but in an hazy sky, then the scatter looks high. I have to reprocess the images before reporting, probably tonight. I had to stop before the end of the egress due to the star being too low in the haze.
I use an ensemble of 6 comp stars including 104, 112, 113, 119 plus two others. They have been stable since July during the EE Cep campaign. 104 is maybe a little less stable, I have to check.

Cheers,
Roger

CQJ
Comp Star

Gary,

It is once again partly cloudy tonight, but last night was perfectly clear, dusk to dawn, with K'v = 0.17, a record low for me.  I have used the 104 star as my only comp star, and the 113 as the check star.  I have not seen any problems with either of these.  But I think I would get better data if I used both as comp stars, and took the average of the 2 results.  Why use a check star if it neither the comp nor check stars have been varying?  I have not been measuring the 119 star.

John Centala

WGR
WGR's picture
Ingress/Egress Sunday Night (JD6957)

Hello

I was wrong about getting the Ingress last night.  It did not clear at twilight and I started an hour or so after that. I did get the Egress in full and about 3 hours post Egress.  I had some clouds from time to time also.  

I also have TA giving me an error of too many groups.  I have emailed George to see if he can up the limit.  I had 390 images in B and 390 in V last night.  I will post the data as soon as I get a new version of TA.  

I also got about 2 hours of a pretty flat bottom last night on this eclipse.  Plot attached.  No editing.  Scatters is clouds rolling thru.

Gary

WGR
WGR's picture
Eclipse of JD 6957

Hello

Resolved the issues of a large data set being transformed in TA with large coef.  George modified TA to version 2.17 to fix it.  I have posted the data from JD 6957 on the web page.  It was a poor night, some thin clouds moving thru, and it shows up in the scatter. 

However, after phasing up data from the last 60 days at 3.877 day period, I can see a flat bottom for 2.3 hours, mostly from the data from 6957.  It also shows that the eclipse phase up with all observers is about 7.9 hours long, with approximately a 2.3 hour dimming followed by a 3.8 hour flat bottom, and that followed by a 3.2 hour brightening.  Total eclipse depth appears to be about 0.15 magnitudes.  I have attached a portion of the phased diagram below.  I also attached the complete phase curve for the past 60 days.

Keep observing, we need more data to clearly define this object.  I plan to observe for another month at least.  Welcome all.   Its nicely placed in the evening sky, near the meridian at twilight. 

 

Gary

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