Forum Replies Created
-
Posts
-
I measured these stars (and added a few more to better cover the spectral classes) on Saturday.
See the attached report.
I have also attached the reduced spectra as fits files if anyone wants to play with them. (Note the extension has been changed from fit to fits to be acceptable for upload to the forum. I also tried zipping them together but the forum does not allow zip files unfortunately)
The results are generally very encouraging with typically less than 2% error in relative flux from 4000-7300A. There are some larger systematic errors below 4000A which seem to be due to a small flux offset and will need further investigation. Extra precautions also need to be taken against scintillation when measuring first magnitude targets.
Cheers
Robin
Hi Steve,
You are effectively taking a spectrum of the flat lamp when you take a spectroscopic flat so you need a light source with a smooth spectrum with good output and no sharp features in the region you are interested in, which for low resolution spectra effectively means incandescent lamps only, running as hot as possible to give maximum uv, hence halogen are favoured. LEDs definitely do not meet these criteria. (At higher resolution some EL panels are ok over at some wavelengths but not from ~3600-7500A that we need) As far as I know Xenon torch bulbs (as opposed to Xenon car headlights) are incandescent filament lamps so should work ok so I dont understand why the flat using the torch looks essentially identical to your LED flat. I know it sounds silly but could you possibly have mixed up the files somehow? eg could ISIS be recalculating using the same raw files. I suggest checking back to the original raw flat images. Other than that, I have no idea what is going on !
Robin
Hi Steve,
Looking again they both look rather like typical blue white LED spectra to me eg
https://upload.wikimedia.org/wikipedia/commons/thumb/4/48/White_LED.png/350px-White_LED.png
Are you sure your torch is a halogen incandescent ?
Cheers
Robin
Hi Steve,
Those do indeed look strange and very similar considering they were taken using very different light sources which should have produced very different looking flats. Are these images raw or are they the master flat after processing in ISIS? If so do the raws show the same effect? Here is a flat from my setup last night (ATK428) using the halogen lamp in the calibration module. (reduced in size)
As far as I can recall flats taken with a halogen lamp plus white sheet over front of the scope as a diffuser before I had the calibration module produced a similar flat.
Robin
Hi Tony,
See https://britastro.org/comment/2420#comment-2420 for a comparison of two different setups
Dont worry too much about matching and focal reducers though. There is no need in general to match the focal ratio of the scope to the ALPY. You can use any focal ratio scope with the ALPY down to f4. It will not change the spectrum shape or resolution. The only reason for using a focal reducer is to better match the size of your star image to the size of the slit so you can collect more of the light and go fainter. This depends only on the telescope focal length (not focal ratio) and your seeing. If the star is smaller or comparable to the slit, then you are better off without one. Even if the star is bigger than the slit, if the star is a bright one, you may still decide not to use a reducer to avoid the risk of chromatic aberrations.
Cheers
Robin
Hi Tony,
If there is any starlight overspilling the slit then yes, any improvement in the seeing will reduce that and a higher percentage of the light will get through the slit into the spectrograph. (The spectrum resolution remains unchanged though as this is determined by the slit width. This is different from slitless systems like the Star Analyser where you always get all the light but poor seeing reduces the resolution)
Cheers
Robin
Hi Steve,
gnuplot seems to be working ok for me with ISIS 5.7.0. (see attached quick example)
Do you have auto scale unchecked on the single plot?
The Y scales cannot be set independently in the double plot so it just plots the spectra as they have been scaled in processing. Your two spectra appear to be scaled to1 near H alpha which is where they match. The hotter of the two stars (eps Aur) then rises faster towards the blue end as would be expected) You could try rescaling them at 5000A say using the “profile” “normalize” function when they will cross closer to the middle of the spectrum.
There are more sophisticated plotting tools around though. For example Tim Lester has recently posted one he is developing on the ARAS forum
http://www.spectro-aras.com/forum/viewtopic.php?f=8&t=1596
Robin
Eps Aur is not really a good choice as a reference star. It is a famous long period eclipsing binary but is not particularly typical even outside eclipse. Using Brian Skiff’s huge database of published spectral classifications
http://vizier.u-strasbg.fr/viz-bin/VizieR?-source=B/mk
we can see it has been classified as anything from F4i to A8i. (A hotter star would be better as it would have fewer lines)
It is also a pulsating variable star and shows varying H alpha in emission for reasons which are not exactly clear. The spectrum also has significant interstellar lines so the continuum shape may be affected by interstellar extinction. Main sequence (luminosity class v) stars are generally a better choice as reference stars as they are less likely to be variable and more likely to be typical of their spectral classification though this cannot be guaranteed unless (like the MILES stars) there is a reliable calibrated spectrum for it.
Cheers
Robin
Unless you have calibrated your flux (The Y axis) in physical units rather than just counts from the camera (which are meaningless to anyone else but you) you should always scale your spectrum in relative flux so the continuum equals 1 on average over some chosen wavelength range. ISIS does this automatically during processing.The wavelength range is your choice and can be set in ISIS under “settings” “profile domain for spectral scaling” This is also sometimes called Normalising eg in ISIS under “Profile” “Normalize” where you can also scale any spectrum in the same way.
“Rectification”, sometimes confusingly also called “Normalising” is a further step where the continuum is effectively removed to make a level spectrum where the continuum equals 1 everywhere. This obviously loses information (the shape of the continuum) so you would only do this step if you specifically needed to do so for a particular analysis. The non rectified spectrum is the one which you would publish in a database for example.
Calibrating the Y axis in absolute physical units rather than in relative flux is an advanced topic which is probably best left until you are comfortable relative flux calibration.
Cheers
Robin
Hi Steve,
I am not sure I follow what you have done.
To calculate the instrument response using a MILES star, you take a spectrum of the actual MILES star you have chosen, not one which happens to have the same spectral class. (The actual spectral class does not matter as long as it is a hot star, it is probably unreliable anyway!) That way you end up with a spectrum of the MILES star as measured with your equipment which you can directly compare with the one of the same star in the library, which was measured and calibrated by a professional. The ratio of the two is your instrument response.
You can then use this instrument response to correct a spectrum of any target measured using your setup, (even if you have no idea what the spectrum will look like) to produce a fully calibrated spectrum. (With one caveat – you should chose a MILES reference star at similar altitude to your target so the effect of atmospheric extinction is the same.)
Cheers
Robin
Hi Tony,
With your setup of 235mm aperture and f6.3, the star image size FWHM with 3 arcsec seeing will be 26 um so slightly larger than the slit width so there will be a small amount of the star overlapping the slit and it will not completely disappear unless your seeing is very good. You should see a significant drop in brightness though as the star crosses the slit. The trick is not to over expose the guide star image. Take a look at the short videos on this page of Christian Buil’s. It gives an idea of what you should see when correctly (fente3.wmv) and over exposed (fente4.wmv). (In French but Google translates well.)
http://www.astrosurf.com/aras/slit/method.htm
If you still do not see the star drop into the slit, your guide camera may not be precisely focused on the slit. This should be checked first before trying to focus the telescope.
Cheers
Robin
The key thing is that if your star image is already comparable/smaller than the fixed slit width, adding a focal reducer to reduce the focal length will not give you any more throughput or resolution even though the spectrograph might be able to accept a faster beam. You just end up adding more unwanted glass and risk chromatic aberrations.
As an example, taking Steve’s 120mm f7 scope, the 23um ALPY slit covers 5.6 arcsec of sky so in typical seeing the star is already smaller than the slit and reducing the focal length further will give no benefit. (If it was possible to reduce the slit width, you could take advantage of the extra potential resolution but the ALPY slit only has a single fixed slit width)
If we consider my setup though (An f10 C11 with 280mm aperture) The slit is 1.7 arc sec and my seeing is typically 3 arc sec so unless I use a reducer I lose a lot of light at the slit. At f5 the slit is 3.4 arcsec and a much better match to my seeing
Cheers
Robin
Since we are near Halloween, have a look at these pages concerning focal reducers on Christian Buil’s site.
http://www.astrosurf.com/buil/isis/lisa_ri/index_en.htm
http://www.astrosurf.com/buil/dispersion/atmo.htm
The first one describes precautions to be taken with focusing using a focal reducer with a LISA but the second one, looking at the effects of the interraction between chromatic aberrations and atmospheric dispersion is particularly frightening !
Robin
Hi Steve,
With a 120 mm aperture scope, depending on your seeing, it might not be worth reducing all the way to f 4/5. In fact it might even be counter productive as your star image might be so much smaller than the slit width that it “rattles around” in the slit making it difficult to guide on the slit as there will be little or no overspill. It could also give potential wavelength calibration issues as the star location rather than the slit itself will start to define the spectrum position. I suggest measuring some star FWHM under typical seeing conditions first and comparing this with the ALPY slit width (23um). The optimum is to have the slit ~equal to the FWHM
Re the ALPY focal ratio, I think the aperture of the ALPY optics will accept an f4 beam but like most optics, they work better stopped down a bit. I suspect you will get a sharper spectrum and potentially squeeze a bit more resolution out at f5 compared with f4, particularly at the ends of the spectrum. This is certainly true of the LHIRES.
Cheers
Robin
Hi David,
Which camera do you have? Some sensors benefit from deep cooling eg the Kodak KAF CCDs. Others, like the Sony CCD, less so. The ATIK cameras I have used only cool ~25C below ambient so I run at -10C all year round which gives negligible thermal noise and saves me worrying about the risk of non matching darks/bias. Incidentally if you do cool to very low temps, watch out for ghost residual images in following exposures. This can happen for example if you take a faint spectrum image after an overexposed calibration lamp image. I have never seen it myself at -10C but it was a problem than plagued a pro-am campaign I was involved in where they were using KAF CCD cooled to a low temp.
Cheers
Robin
Thanks Jeremy,
Three confirmations so far this year (bme,gfr and gxp) This latest spectrum was also a test of my camera upgrade from ATK314 to the more sensitive ATK428 which seems to be working well.
Robin
Hi Steve,
Have you been able to discuss this with the original author? Perhaps if you can reach an agreed view, an author’s amendment to the paper might be appropriate and acceptable to the editor, avoiding the risk a protracted discussion. Although not strictly peer review, ArXiv can indeed be rather good in this respect where the on line version can reference the conventionally published version and be subsequently revised with the original version still accessible. (and useful for us poor amateurs unable to access papers behind paywalls of course!) It still would not allow easy tracking back from the Journal version though unless pre published on arXiv and referenced in the Journal version. I see some journals actively encourage authors to mirror their papers on arXiv. I am not sure what JBAA view would be on this however.
Cheers
Robin
Hi Steve,
In practise critiques of papers in the way you describe are rarely published in any scientific journals. Where the findings of a paper are questioned it is usual to do so through the publication of another paper, offering further evidence which may lead to alternative conclusions. The scientific community is then left to judge the validity of the two bodies of work and carry out further investigations to clarify the situation.
If however you are not ready at this stage to offer alternative evidence by way of a formal paper and just want to test out your thoughts on specifically why “the process used in the paper to generate the estimates from the measured data” might not be sufficiently robust, even though the referees of the paper apparently thought they were, then I have found an informal discussion with ones peers as could be done here, can be a very effective in clarifying ones thinking.
Cheers
Robin
