AAVSO Alert 655 "Anticipated secondary eclipse of b Persei by its third star in November 2018 " has a call for spectra but some essential details are lacking and some of the details there are seem rather strange:-
------------------------------------------------------------------------------------------------------------------------------------------
"Spectroscopy will also be extremely valuable, especially when the C star is expected to pass behind the A and B stars at different times throughout the eclipse during the three days centered on Nov. 21, 09:30 UT. We hope to see small changes in some parts of the spectra if the C star is totally or partially eclipsed. Please send any spectroscopy observations to Dr. Anatoly Miroshnichenko <a_mirosh@uncg.edu> . All spectroscopy data should be displayed with the following guidelines:
"The FITS header should include: coordinates of the observing place; the observer name, RA & DEC of the star, UT time of the observation (better mid-point), exposure time in seconds. The intensities need to be in ADU (not normalized to the continuum) and wavelength scale needs to be geocentric."
-----------------------------------------------------------------------------------------------------------------------------------------
Are these requirements as requested by Miroschnichenko or is this AAVSO's interpretation eg
What kind of small changes and in what parts of the spectrum? ie what resolution do I need? (If high resolution I need to know where to look) and what SNR is needed to see these small changes.
Are these data are to be used for RV measurements? if so what wavelength accuracy and precision is needed? The alert asks for geocentric spectra. This is very unusual. Does it actually mean topocentric?
Is flux calibration required ? (relative or absolute) and if so to what precision? (The Alert strangely asks for spectra in ADU. Does this mean raw spectra uncorrected for instrument response and atmospheric correction ?? This would be very odd. Or does it mean calibrated in relative flux but not rectified)
Are the RA Dec coordinates really needed in the header? This is a single object campaign. Are the coordinates expected to change during the campaign ??
Robin
Exactly the same questions from my side about fits header and format, as Robin asked (geocentric is also unknown for me what does it mean).
Be aware, most of our mount (eq6) is not precise enough to make RA/DEC for any use - e.g. verification or position determination (indeed, have lots of other means to find the object reliably).
I think it is too early to speak about "how to observe", we have lots of new data since 1976 without firm publication of 3rd component (I guess this is why 1976 mentioned), and I'm curious whether new data gathered by Miroshenko would confirm the F type of 3rd faint component, or detect it by any means. (As for me, my measurements of Ca II 3933A attributable exactly to 1 component - out of eclipse -, debating that the 3rd star was an F type.)
Let me remind, the paper "Hill et al., 1976, ApJ, 208,152" mentions 2 solutions, one is judged less likely. It was nice to confirm / debate using the newly gathered, much better qualty professional data.
- Peter
Robin and Peter, Thank you for your interest in observing spectra of b Persei during the transit of the third star in about 3 weeks. I have asked Dr. Miroshnichenko to answer your good questions. If we can obtain spectra when the third star is eclipsed by either the two close orbiting components, that will be extremely valuable to examine if there are any changes due to partial or total blockage of the third star.
Best,
Don (CDK)
The information and instructions in AAVSO Alert Notice 655 regarding spectroscopy are verbatim from Dr. Collins. I will relay to him and the other PIs, including Dr. Miroshnichenko, your questions and concerns.
Good observing,
Elizabeth
Hi Robin and Peter,
Dr. Miroshnichenko responded to your questions and comments with the information below. If you still have concerns, please post them here. Thanks! - Elizabeth
- from Robin Leadbeater:
Q: What kind of small changes and in what parts of the spectrum? ie what resolution do I need? (If high resolution I need to know where to look) and what SNR is needed to see these small changes.
A: I'm not sure we know what kind of changes to expect during the eclipse. Therefore, the higher the resolution and the SNR, the better. A resolution of over 10,000 is good for RV measurements. Lower-res. spectra will only show strong changes, which I do not expect.
Q: Are these data are to be used for RV measurements? if so what wavelength accuracy and precision is needed? The alert asks for geocentric spectra. This is very unusual. Does it actually mean topocentric?
A: RV measurements is one of the project goals. Another one is studying changes in spectral line parameters, such as equivalent widths and intensities. As far as the wavelength scale is concerned, we may have a terminological problem. To me geocentric scale is what you get from a comparison lamp wavelength solution. It is then transformed into a heliocentric scale using the star position and time of observation. Wavelength accuracy depends on the spectral resolution. For relatively high-resolution data provided by, for instance, the Eshel spectrograph from Shelyak, it is around 200-300 meters per second. I think an accuracy on the order of 1-2 km/s is still good for the project.
Q: Is flux calibration required ? (relative or absolute) and if so to what precision? (The Alert strangely asks for spectra in ADU. Does this mean raw spectra uncorrected for instrument response and atmospheric correction ?? This would be very odd. Or does it mean calibrated in relative flux but not rectified)
A: Flux calibration is welcome but not required. It is really hard to achieve using echelle spectra. If long-slit spectra are taken, a flux calibration standard spectrum needs to be taken and calibration procedure needs to be described. I personally take echelle spectra and do not do flux calibration. It can be done by contemporaneous photometry. The instrument response and atmospheric correction do not matter. Local continuum normalization does the job. The latter also applies to the ADU scale. When continuum normalization is done, the intensity is expressed in units of the local continuum.
Q: Are the RA Dec coordinates really needed in the header? This is a single object campaign. Are the coordinates expected to change during the campaign ??
A: RA & DEC are very helpful in the header, but they can be entered manually and not from the telescope communication, which may contain systematic errors. I'd prefer J2000 coordinates. Heliocentric correction calculation using precession is done in IRAF.
- from Peter Somogyi:
Q: Exactly the same questions from my side about fits header and format, as Robin asked (geocentric is also unknown for me what does it mean).
A: Geocentric or topocentric, I'd like to collect data products where calibration by a comparison lamp is done with conversion into heliocentric units. The latter procedure needs to be done in the same way for all the data to avoid accidental or systematic errors.
Q: Be aware, most of our mount (eq6) is not precise enough to make RA/DEC for any use - e.g. verification or position determination (indeed, have lots of other means to find the object reliably).
A: Again, RA & DEC need to be in J2000 to avoid mount communication problems.
Thanks for the feedback.
Although becoming more common. Amateurs equipped with high resolution echelle spectrographs are still pretty rare so most users working above say R~2000 are limited in wavelength coverage and at 1-2km/s RV precision eg with the LHIRES at maximum resolution, the coverage would only be typically 100-150A and changing wavelengths repeatedly during the night is not straightforward with this instrument, requiring focus adjustment and additional calibration frames. This is the case for me, hence the need at high resolution to target a specific line or group of lines. Additionally I would say to be able to compare RV between observers to 1-2km/s some measurements of RV standards would be advisable to quantify any inter-observer offsets.
There are a larger number of amateurs equiped to give full coverage at R ~500-1000, for example using LISA, ALPY, DADOS etc Would results at this resolution be useful ? (A common practise is to calibrate these spectra in relative flux using reference stars observed contemporaneously. (A few observers also convert to absolute flux, for example using simultaneous photmetric measurements.) The relative flux accuracy is typically of the order of 10-20%, though if you are not interested in the continuum this should not matter.
Cheers
Robin
Thank you for the answers, so having precise RV is not mandatory!
"A: I'm not sure we know what kind of changes to expect during the eclipse. Therefore, the higher the resolution and the SNR, the better. A resolution of over 10,000 is good for RV measurements. Lower-res. spectra will only show strong changes, which I do not expect. "
Since I do not expect with an LHires (R~10000 can do only at a narrow range) to provide a matching RV-quality with a fiber fed eShel, I will continue experimentation at medium resolution to catch any change - once after hit the lucky eclipsing period (weather-free time gambling).
- Peter