Skip to main content

Uses for U filter?

10 posts / 0 new
Last post
Meeka
Meeka's picture
Uses for U filter?

I have a U filter that isn't getting much use - I usually take UBVRI of any target I'm doing unless I'm looking for something more specific, but 90% of the time the U image doesn't have any comp stars in VPhot so I eliminate it from subsequent observations of that target. I don't really feel comfortable requesting a sequence for a filter/target combo unless I had some project in mind, which I don't at this point.

I'm wondering if there is anything in particular the U filter might be useful for? It's the Astrodon U filter on an SBIG ST8-XME if that's relevant.

sric
sric's picture
U filter

I just added a Astrodon U filter to my AT-10RTC with a SBIG STT-3200 camera.  Searching through the CV target list it seems that stars in the ZAND variable type have U comps.  If someone can suggest other categories with U comps it would be appreciated.

I just calculated transforms with M11 using Gordon Myers TG program.  They look appropriate except for Tu_ub which is 2.132+/-0.018.  This doesn't look right and I wonder if the Astrodon U filter covers the same bandwidth as the U data in the Variable Star Plotter table.

Comments appreciated.

Richard Sabo 

HQA
HQA's picture
U filter

Hi Richard (and Meeka!),

The tu_ub coefficient is about right for the typical front-illuminated CCD.  The KAF-3200me sensor has good QE in the U-band, but not nearly as good as a backside-illuminated sensor.  Also, there are several glass interfaces that may come into play.  I don't know if SBIG removes the cover glass on the sensor, or what kind of glass they are using for the entrance window into the camera.  All of these come into play when dealing with the "UV".  These issues usually impact the blue end of the U bandpass, pushing the U response to the red, which is why you get the big coefficient.

What you will typically find is that exposures are very long, and so you will probably be observing only stars that are much brighter than you are accustomed to.  I'd go 3-5 magnitudes brighter in order to get good signal/noise in a reasonable exposure time.

That said, few amateurs do U-band photometry, and so all measures are valuable.  Start with blue objects; the Z And variables you mention are good candidates.  Many of the comp stars that are brighter than V=9 have U-band photometry from photoelectric observers; look at the GCPD via VizieR if there is no U-band measure in VSP for a particular comp star.  APASS DR11 will include its new Sloan u-band photometry of the brighter stars, which can be transformed for the Johnson U.  SRO has a U-band filter in its filter wheel, and so you can put in an AAVSOnet proposal to have it calibrate any specific field at U.

Good luck!  It should be fun.

Arne

sric
sric's picture
U Filter

Thanks, Arne.

I was unaware of the GCPD source.

Richard

tcalderw
tcalderw's picture
tutorial?

Does there exist some kind of tutorial that addresses the complexities of UB data reduction?  I have not, so far, been able to find one. 

I was considering UB photometry of VV Cep, which has been requested, but it appears that the corrections one would apply to U band (eg: to handle Balmer jump) are based upon the spectral type of the target.  Since VV Cep is a blue/red pair, with the blue star dominating in U but the red star dominating elsewhere, it seems impossible to make the needed corrections.

Tom

 

Stan Walker
U Filter

Hi Tom,

Good to see someone planning UBV photometry. I had a look at the field and there are some blue stars around but Guide 9 gave no idea of the eclipse depths which would indicate the relative contributions of each star. I looked at Wikipedia which had some information differing from the GCVS but a rather poor B-V colour indicating about two or three times as much reddening as in the eta Carina region. But obviosly the blue star, if its class is B0-2V as stated there, is a substantial contributor of the total light.

There must be dozens of open clusters with UBV values in catalogured from the 1970s and 80s. Do the ubv measures in the normal manner and plot b-v against B-V as normal and then u-b against U-B. But if you plot U-B against B-V you will see that there are two zero points in U-B, one at B-V = 0.0 and the other near B-V at 0.40 depending upon MK classes and other things. Our first standardisation of the Auckland system was in 1970 using IC 4665 I think which is a long way north of Auckland's S37! I presume you're using PEP and not CCD.

The question is: do you derive three different coefficients in the u-b correction or merely correct u-b against B-V? I'd try to determine a slope in the region from A0 to B0 spectral class using u-b against U-B and use this for everything. This means that your comparisons would also be in this region - I see a few there and there would be a long transformation for the red star in V, B-V and U-B but if U or U-B measures are required then that's where your emphasis should be.

Perhaps you should contact the people requesting the measures but if they're spectroscopists they probably will not be of much help.

You know my email address so contact me if you wish. We spent thiousands of hours doing UBV photometry in the 1970s to 1990s and I'll probably return to this again shortly using an Atik camera with a Sony detedtor chip which has good UV response. One good thing about these blue stars is that the signal in B and U is very high.

Regards,
Stan

HQA
HQA's picture
VV Cep

Hi Tom,

These "composite spectra" stars are often difficult to transform, because their colors do not necessarily conform to any astrophysical single object.  If it was a hot B star accompanied with a cool dwarf star, then the blue part of the spectra would be dominated by the blue star and you could get reliable UBV photometry of it.  In the case of VV Cep, it is a luminous red supergiant with a smaller blue star, so the spectrum tends to be dominated, or at least heavily influenced, by the red star, nearly everywhere in the visible spectrum.  So transforming U-band data may or may not work, and will most likely fail as the eclipse begins anyway.  About the best you can do is to submit untransformed data if you are going to do wide-band photometry.  Note that most amateur spectrographs do not cover the U bandpass, so anything you contribute will be unique and important.

The other thing to keep in mind is that U filters are particularly difficult to make.  The biggest problem is red leak, which for measuring a blue star, is not often important.  However, for measuring a red star, where the star can be 10,000 times or more brighter at 900nm than it is at 350nm, how you handle red leak becomes critical.  Astrodon's current U filters do pretty well, but I don't know if they've been thoroughly tested for extreme cases.  You might try measuring known red stars, like R Leo or S Ori, and see whether the U-band photometry makes sense.  That might give you confidence in your VV Cep measures.  The other option is simultaneous U-band photometry between a photoelectric (preferably photomultiplier) system and your CCD system, to make sure that you get the same results.

These composite systems are always best to observe with a well-calibrated spectrograph, or with narrow-band filters like Halpha.  That said, wide-band results will be important for many reasons, such as the determination of eclipse width and depth, and will be a good complement to those performing spectroscopy.

Arne

hambsch
hambsch's picture
U filter

Hi,

I also have recently installed an U filter on request of professional astronomers I work with, hence there is interest in U band photometry. I have not used it much though but will start imaging standard fields for getting transformation coefficients.

Josch

HQA
HQA's picture
U filter transformation coefficient

Richard indicated that he was getting Tu_ub = 2.132.  I mistakenly replied that this was not uncommon.  At the time, I was thinking of Tub - the color coefficient.  Having a filter coefficient as large as 2.132, when it should ideally be around 0.000, is far from normal.

I think the best approach at this point is to post the raw photometric data that you fed into TG so that we can see if there is some obvious problem.

Sorry for my misunderstanding!

Arne

HQA
HQA's picture
Richard's U transform

Just to keep people informed, Richard's problem was that his "U" exposure was actually taken through an open slot in his filter wheel.  He will post new coefficients when he gets another opportunity to image one of the standard clusters.  Finding the issue was immensely helped by VPHOT's ability to share images between users, so that I could actually see the images that Richard was using.  It is nice to have such tools available to AAVSO members!

Arne

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