AAVSO Alert Notice 563 announces an observing campaign on the bright (V~4.57) triple star b Per to cover the interval before, during, and after an eclipse predicted for mid-December 2016. Please see the notice for details and observing instructions.
Many thanks, and good observing,
Elizabeth Waagen, AAVSO HQ
I am eager to participate, but use a DSLR. Although I am sorely tempted to use a single comp star (000-BBG-572; with 000-BLL-386 as the check star) I fear that will result in an inability to produce results with the implied required accuracy of 0.01 mag or better. I've put in a request using CHET for sequences of some other bright comp stars, identifying HIP 19356, HIP 19727 and HIP 19986 - all of which have V magnitudes brighter than 7.0, and ask that any other DSLR photometrists express their support.
Alternatively, If someone can tell me that I can achieve the required accuracy using a single comp I am perfectly willing to shut up.
b Per is a very nice target for DSLR photometry, and if you follow the hints given in the Alert notice (defocusing, avoid saturation, potentially by stopping down the aperture), you should be fine. Using many comp stars in an enemble can help to minimize a systematic offset (e.g. your results are consistenty higher or lower than others by a certain constant), but for this eclipse project any observer-specific offset can also be removed afterwards. It's more important to pick a setup and then stick to it consistently during the campaign time.
It's important that many observers participate so that bad weather at some sites during the eclipse doesn't wreck the whole campaign.
Thanks, I will! Stopping down is not a problem for me because my DSLR is used with a 71 mm aperture (a WO Zenithstar with a focal reducer. I wrote the above post before discovering a way to access SEQPLOT, but now have a full (6 star) ensemble of comps of magnitude brighter than 7.0 V plus the 74 check star.
Just for yucks I am going to process the data four ways: Once using the 55 as the comp, once with the 55 and the 43, a third time with six comps, and finally using the 55 as the comp and the 43 as the check. Theoretically the third way should minimize the errors, but if it turns out that the fourth way is the more accurate, it will permit a much faster cadence as I will be able to cut my exposure time by a factor of around 5.
I am still very much a novice when it comes to photometry, and am still discovering new tools as well as new ways to make mistakes. Until very recently, though, I had the wisdom to keep quiet.
Thanks for the encouragement!
Sounds like a plan!
You will have to be careful about saturation, 77mm aperture is an aweful lot for a star this bright.
I'm not quite following what you wrote about exposure time, tho. Your exposure time selection will be constrained by b Per itself already, which is terribly bright at ca 4.5 mag. I know you are also doing WR 140 (like me), that one is at 6.8, so 2.5 mag fainter, or a factor of 10 (!!) in flux.
For b Per it might actually be easier to use a DSLR camera lens with an adjustable aperture stop.
I'm not sure yet what I will use, either a f=135 mm telephoto lens or an f=85mm one, stopped down to avoid saturation.
Don't be discouraged about writing here. This place is not (primarily) for bragging but for learning and discussing stuff :-).
Stephen, your plans look good! I like star '55' as the comp star - it's close to b Per, about the same color and brightness. The most important fact is that it is relatively close to the target in the sky, which is very important when some thin clouds intervent.
Bright stars are a problem to keep the noise down - paradoxical as it may seem. DSLR's and CCD's have the same problem. Things to keep in mind: be sure to avoid saturation (try for ~1/2 the full-well level for the imaging system. With a DSLR shoot with smaller ISO settings - around 400 ISO. This helps by forcing large integration times (at least 20 seconds) so the scintillations are distributed over all the seeing region. Finally, co-adding frames, or "box-car" averaging results to minimize the scatter is important. I can perform the last step when I collect all the data.
Don't worry too much about accuracy, but about scatter. When we compile every optical systems' data, we make a zero-point shift so that the out-of-eclipse data for each observer is consistent with the out-of-eclipse light curve.
Welcome aboard! Glad you are observing!
Regarding the various comments about using a telephoto lens: I started by trying that, but maintining focus is a nightmare. One is forced to do a focus exercise every night, which is a big waste of time. I bought the Zenith star 71 for the specific reason of avoiding that: I focus once per season and lock the tube in place. As the tube has graduation marks on it, I can check that the tube is undisturbed either visually or using a feeler gauge (left over from adjusting valves on my VW).
Having done the linearity testing that Mark Blackford describes in the DSLR manual, I use ISO 200 because it gives me maximum linear dynamic range. That would entail a 7 - 8 second exposure time for a star of b Per's brightness to keep the brightest object at around half the max ADU count. I was thinking about going to ISO 400 with an exposure time of 3 seconds but was reminded by the comment below that longer is better to counter the effects of scintillation. Of course my exposures are made in sets of at least 10 so as to obtain better statistics. I do not stack frames, but rather measure them individually before aggregating.
But all of this is moot because it has been cloudy, is raining, and the weather man promises more of the same over the next ten days. I hope enough people have clear skies so that this once-every-720-days event can be observed with precision!
From cloudy, light polluted, lousy seeing at best Virginia,
I agree about using a DLSR lens like a f=135 mm telephoto lens or an f=85mm one with a DLSR or astro CCD. Defocus the stars so that peak ADU is ~1/3 to 1/2 full well (or 20,000 to 30,000 ADU with a 65,000 max ADU CCD). Be careful when the lens cools at night as the focus changes more than a telescope per temp drop. I observe what direction the focus changes and then set the focus at a point where the star continues to defocus (lower ADUs) as the temp gets cooler (as opposed to saturating).
As this is my third round with b Per, I successfully used a SBIG ST-8300M with a 85 mm lens plus a 50 mm V filter inside the t-thread to DLSR lens adapter. I piggyback this camera on a polar-aligned telescope / mount setup and use the telescope as the autoguiding scope.
I have been observing b Per since December 8th. I think this one is my third campaign of photometry . At this moment (18:30 UT, December 11th) the star is still at his normal brightness (4.645 mag. and climbing).
With my current setup (ED 66/400, ST-8XME) I can reach easyly a precission of +/-0.005 mag. With a final improvement, +/- 0.003 mag. is possible.
By the way, I wonder if spectroscopy can be useful or it is a waste of time and energy. I could obtain low-resolution (R=700) spectra covering all the visible spectrum, or select a window of 700 A and do medium-resolution spectroscopy (R=5000).
You raise a very interesting point. I have been wondering whether DSLR photometry can even be applied to a system like this, because even if the individual stars were perfect black bodies, if they are at different temperatures the shape of the composite spectral curve - upon which the accurate transformation from Beyer G & B to Johnson V depends - is decidedly not that of a black body. Given the three objects are at different temperatures, the composite spectral shape ought to vary during the course of an eclipse, and I suspect that variation should be detectable spectroscopically. So to me it seems sensible to observe the event spectroscopically if one is equipped to do so.
We are very anxious to obtain radial velocity spectra during this campaign. Anatoly Miroshnichenko of the Three College Observatory in Greensboro, NC and his students at UNC-Greensboro are gathering excellent radial velocity data from several high resolution telescopes and spectrographs. They have been observing b Persei all season.
The most recent spectral studie of the b Persei system (Hill, 1976, Astron. Journal, 208, pp. 152-164) is ripe for updating with modern spectrographs.
Other new spectral studies of the b Per system, especially in and out of eclipse, will be more than welcome!
I am pleased to see good photometry of b Per coming into the AAVSO as seen by the aavso LCG (both version 1 and version 2) Good European coverage!. After three days and nights of excellent weather here in N. Carolina, we are clouded out for most of this week when the eclipse/transit is expected. We really need observers whenever you see a break in the clouds.
The eclipse could begin any day/night this week, so we really need large longitudinal coverage.
I'm onboard for this third "eclipse party" since last saturday evening. Tonight the sky is cooperative, but the weather prospects are not very good for France this week. Wait and see.
In case of utility, I can process separately RGB channels from the DSLR images and produce 3 different LC. Is this useful ?
I would say that if you have all three channels (either with filters or color imagers) by all means submit them to AAVSO. For the time being I will be concentrating only on the V channel for this eclipse.
Thanks to all for participating in this and previous b Persei campaigns!
After 2+ weeks of bad weather, we should finally start getting a few breaks in the clouds this week in Bloomington, IN, USA. My plan is to use the Atik 414ex + V filter on the guidescope (focal length of 280mm allows a descent fov on the small sensor) and maybe also a DSLR on the 80ed.
First window of clear skies: tonight between 7 and 9pm EST...
I think eclipse has already begun.
I started observing yesterday at 18:00 UT and b Per was. at 4.61 V mag. but at 21:00 (more or less) it began to fade steadily. I stopped observing at 2:00 UT and b Per was at 4.67 mag. and fadding . Please confirm !
Sent from my mobile Phone. Sorry for mistakes and shortness .
Hi, I am sorry for the false alarm. At this moment (December 13th, 18:25 UT), b Per is still at 4.64 mag., I will continue intense monitoring but forecast weather is not optimistic for the next days, specially for the week-end.
We expect the eclipse of b Persei to begin possibly as a series of small dips in brightness (as partial eclipse) within the next couple of days. Dec. 15 UT (tonight) is the most highly predicted beginning of the eclipse. The previous light curve for this phase of the eclipse in January 2015 was incomplete - a partial eclipse could have occured a day or two earlier than we first confirmed eclipse during a gap in our data.
Keep on shooting, through and between the clouds, and if you see a dip in brightness - Shout it out on this forum. The weather is poor for many, that's why we have many observers poised to collect! I observed about half an hour last night and I posted the noisy data that show no eclipse then. Please post any data if the data are good enough to show whether or not the eclipse has begun!
Just got some hole in the clouds for about 20 minutes, we were in thick haze for several days, it thinned out tonight for a while about 21:00 UT and now clouds are back.
40 x 10 sec images from which 5 points are extracted:
7737.36458 4.596 (0.004)
7737.36585 4.590 (0.005)
7737.36674 4.603 (0.003)
7737.36770 4.609 (0.004)
7737.36866 4.600 (0.005)
Seems the eclipse is not on.
Clear Skies !
Hi, all I know that 12"F4 is not exactly the best tool try take ccd photometry as bright star as b Per 4.57mag. B Per is a nice target, with nearby bright 5.456mag comp star - so I tried. I have no other equipment (12"F4 Newton, Atik314L+ and Jhonson-V filter) FOW is around 25' so b Per and 55 comp star fits just in same field. Fortunately 9.2mag star TYC 3336-2059-1 fit in field also, and this can be used as ref star. War plan was: defokus - a lot, and keep exposure time rapid enough to obtain max 30 000...35 000ADU. This meant 25" FWHM and 1 (or 0.9) sec exposure time. So i took 1sec exposure + 5sec delay.. This mean in six hours run i get 3600 pics. For light curve i average 10 measurement to one observation, to reduce noise from stars twinkle. The outcome, I was so pleased that I took the liberty to publish results. The following ten days, here is a full cloud. This was my b Per contribution. Clear skies!
Dear KTU, As you can see from the AAVSO light curve generator, your data with the 12 inch f/4 telescope with many short exposures fits the rest of observers' data very well. We can see the ellipsoidal variation. Since you have a protocol very well developed that works, I strongly recommend you use the same protocol for the the rest of this campaign. It would be interesting if for future campaigns on this star that you try placing an aperture mask (you could make one out of cardboard and experiment with optimum aperture. Then you could focus sharper and possibly get better results from small aperture, less de-focus, and 10-20 second integrations. I've never tried this with a similar-sized telescope for bright stars, but am interested in learning from the outcome, or if anyone on this list has tried a technique that I am suggesting. But wait until the current campaign is finished!
Thanks for your contributions!
Just had a quick look at the last images taken tonight (from ~2am to ~5am UTC Dec 15th). A quick stack of 10 images taken shortly after 5am UT and processed with iris show a magnitude of 4.655. Check star (000-BLL-386 / 55 on chart X17050EB) was measured at 5.493 (->error of +0.037). So, no sign of the eclipse yet...
Next clear night probably wont be until Sunday for me. I really wish I had a backyard to let the telescope do its job all night.
Will post the full data to the database after I take the time to do a correct processing of the whole thing.
Greetings, Many thanks to all the persistent observers who have been keeping us up-to-date with observations and reports of no eclipse yet. Data with no eclipse are just as important as data showing eclipses. The object is to pinpoint the times of eclipse - difficult with the constantly rotating A-B pair. I observed with an excellent clear break in the sky in N. Carolina (southeast USA) until 02:30 am UT this morning. Get a beautiful portion of the ellipsoidal curve through the peak brightness and no eclipse. Will post the photometry soon.
Keep on looking through the gaps in the clouds the next several nights, and through the clouds if they're thin enough. Thanks all for the good work!
b Per about 4.640 between 18:15 and 21:20 UT , somewhat dim for the low phase of the oscillation but stable. Apparently no eclipse yet. Sky somewhat hazy but ok, continue to observe.
mag. 4.67 at 20:10 UT, stable. Very few images, at this moment the sky is cloudy and there will be bad weather at least for the next 48 h. I have the hope that Saturday I could resume the observations.
I have uploaded all my observations up to yesterday. Sorry for delay...
From N. Carolina, USA. I, too, see the start of ingress. 4.67 and declining. Our weather here improved to cirrus only. IT LOOKS LIKE WE HAVE INGRESS! More to come as I have almost the whole night ahead.
It's almost 0400 hrs UT Dec 16, 2016 and the 4th optical eclipse in the b Persei system has been detected. Normal peak-to-peak out of eclipse variation of b persei cycles from 4.6 to 4.54 on my system. Thhe brightness has decreased steadily from 4.64 to 4.7 since 0.4 hrs UT.
My weather in N carolina has cleared as the decreasing noise in tthe graph indicates Ill continue observing until about 0900 ut, but am going to zzzz now.
Congrats to Paul and Roger for noticing a dip in brightness earlier!
mag. 4.685 about at 0:00 UT (I don't remember the exact hour). I am going to work now, I will upload the very few observations this afternoon. I also believe that eclipse is ongoing.
Clear skies, but here, i Spain, I believe this night, tururut... :-(
Observation started about 4.680 in a difficult sky, through thin clouds. Yesterday ended about 4.685. Not sure we got an eclipse ! But those values are somewhat dimmer than my usual low point of the oscillations, unclear to me.
Remarkably I had mostly good weather last night between Dec. 15, 2330 UT and Dec. 16, 1000 UT (5:00 am local time). The eclipse definitely started. The brightness dipped steadily to a minimum of ~4.7 (~ 0.1 mag fainter than the lowest ellipsoidal out-of-eclipse curve) midway through the run. The brightness then slowly increased to about 4.675.during the last 5 hours of the run. It may be a partial eclipse (expected at first contact) or a full eclipse of the third star being fully or partially blocked by the inner two stars. When the third star spends some time behind the larger A star of the pair, we should still see the ellipsoidal variation due to the different orientation mainly from the A star (spec A7). Time will tell. My weather looks clouded out tonight (a critical moment in the eclipse). Good luck to Roger and any others for getting some data tonight. I wish you all clear skies!
I am delaying uploading the light curve because I have some issues with flat fields.
I finally took the time to correctly process my data taken on Dec 13th, 14th and 15th. Sorry for the delay. I'm still a beginner and I don't have a very efficient pipeline setup for the data analysis.
The data from the 13th is too noisy to be of any value (with exposure times of 4 seconds it's not a big surprise). I was a bit less stupid on Dec 14th and 15th and I stepped down the scope (covered the end of the tube with a piece of aluminum and poked a small hole in it) to get exposure times up to 12 seconds and an ADU of ~40,000 on the target star (the Atik 414ex is 16 bit, so I was at ~2/3 of the dynamic). This data is much less noisy and seems to be in agreement with what other more experimented observed have reported (although I seem to systematically underestimate the magnitude of my target star).
On Dec 15th, I had some clouds and I was about to call it a night after a first series of images when a new break in the clouds gave me an opportunity to take a few more images. Unfortunately, that led to a slight change in my fov and I lost the check star I had in the first series (HIP 20370). I used HIP 20234 as the check star in the second set of images from Dec 15th. It's a variable star, so not a good thing, but the only star with good SNR and B-V not too far from the target I had in my images.
The good news, is that these two series of images taken on Dec 15th (from Bloomington, IN) might show the beginning of the eclipse at sometime around December 16 00:23:05.0 UT. Considering how noisy my data is and the fact that I'm not very experienced, I don't have a very high level of confidence in this conclusion. However, I not that my timing is in agreement with the graph posted here by CDK (Thu, 12/15/2016 - 23:12) if I read it correctly. I'm really bumped that the clouds completely took over after 8pm last night, but it was still worth spending 1 hour in the cold out in the city public park to collect this data! I'm keeping all my raw images in case it can be of any use.
I've attached a graph showing the result (bins of 3 images).
We wont have clear skies until Sunday night here in Indiana. I hope that you guys will have better luck with the weather !
Hi all, Several observers braved the cold, snow, and fought back against the clouds and obtained observations data for the 4th optically-observed eclipse of b Persei in about 4 years! As folks can see from the AAVSO light curve generator, if you show about a 10 day interval, we see a major dip and re-emergence on JD 2457739 (Dec. 16, UT), and it looks like we observed a second ingress/egress on JD ...740 (Dec. 17, UT) - both from the light curve generator and from forum reports. There are also active observers in the UT + several hours time zones (Asia time zones) whom we wish good luck with the weather.
Keep on shooting through every cloud break you get! Even after the eclipse is finished in order to calibrate our offsets. The data look good! Many thanks!
today the sky is clear and forecast is good for this night. I am already observing: mag. 4,690 at 18:00 UT (Dec. 17th). Hope it is not too late to see something exciting!
Now, at 20:30 UT, b Per has climbed to mag 4,645. So we are ending an eclipse...
I observed yesterday for a couple of hours around UTC~17:00 (plot attached, magnitude in sdss g), and it seems indeed that we missed the eclipse entirely due to the stormy weather in Tel Aviv last week.
Hi all, Thanks to observers in Europe and good weather, observers have seen the emergence of the third dip in brightness as the alternating A and B inner stars of b Persei interact with the passing C star. The Light Curve Generator shows traces from three separate observers egressing out of an eclipse. Thanks for everyone's persistence! It looks like we'll be able to tell with more confidence whether the C star is now passing the far side or the near side of the system. We are also anticipating modern spectroscopy data this year from Anatoly in North Carolina.
Please make sure that all observers observe observe out-of-eclipse both before and after the main eclipses. If anyone missed observing before the eclipse, then we'll make do with after eclipse. Keep on looking for eclipses at least one more day. There might be another eclipse or partial eclipse to come.
I only got about 15 minutes of very noisy, cloudy data last night.
Hello all b Persei observers,
I have now finished processing the numerous time-series results of the latest b Persei eclipse campaign of mid-December 2016. See the attached light curve. We had excellent multiple-observer coverage during the eclipse, although we are still "blacked-out" from Asian, Pacific, and western North America participation. I have processed the data to place the out-of-eclipse data to match a common ellipsoidal light curve. I only included observers data which consisted of time-series runs that also included out-of eclipse observations. Some erratic points from some observers which were inconsistent over several runs were also not included. Please let me know if i have omitted any runs that cover both eclipse and non-eclipse, or even runs of out-of-eclipse data that missed the eclipse as I want to include all the relevant data.
We have made several conclusions about the 4 eclipses that AAVSO has observed since February 2013. We conclude that the eclipses observed in February 2013, January 2015, and December 2016 are all secondary eclipses in which the C star of the b Persei system passed on the far side of the close-in AB binary pair. The average interval between these three eclipses is 704.4 days (2 days longer than the 702 day long-period cycle observed by Hill, et. al. (1976, ApJ 208, 152). Hence our eclipses appeared to occur "late". These three eclipses show approximately the same depth - especially the successive dips within the January 2015 and December 2016 eclipses. Equal depths are expected when the third star is totally eclipsed by the inner stars in turn. The March 2016 eclipse is significantly deeper than the other three eclipses, which is supported by the predicted dip for a type F star to block part of the larger, brighter A-type star. Finally there is evidence in the spectroscopy data from Hill (1976) of a discrepant set of data that don't fit the rest of the long-period spectroscopy. This discrepancy "could have been caused by a partial eclipse of the bright A star by the C star". This event is almost exactly an integer number of 704 day cycles earlier than the March 2016 AAVSO eclipse. Recent radial velocity spectroscopic measurements by A. Miroshnichenko during Oct 2016 to Feb 2017 indicate the average systemic radial velocity of the A-star is just beginning to recede indicating the AB pair has been in the foreground of the AB-C orbit. These radial velocity measurements will be continuing for future eclipses.
The next eclipse (a primary eclipse) is predicted in mid-February 2018.
Special thanks to all the observers in the latest campaign, to observers who contributed in any of the 4 campaigns in the past 4 years, and especially observers who participated in all four campaigns! I don't think there are many 3-star eclipsing systems known!
Don Collins (firstname.lastname@example.org)