Hi everyone,
A couple of weeks ago, I thought I found a good potential target for DSLR photometry, RS CNC. It has a decent sized amplitide, and the publications that I have been able to find on this star gives a B-V color of about 1.35, but I am finiding that maybe it isn't so good of a target for DSLR measurements.
I'm now starting to crunch through some measurements, and when I compare my calculated V magnitude with other measurements made at the same time, I noticed that my results differ by about 0.05. I usually don't worry too much if my results differ from other observers by about 0.02, but 0.05 is a bit larger than usual.
So I wanted to see if perhaps the color index I was taking from catalogs was wrong. To do that, I would measure the magnitudes of various stars in the same image field using the DSLR Blue and DSLR Green channels. I would then plot a graph of B-V[catalog] versus DSLR(blue)-DSLR(green). I'm attaching a spreadsheet of the comparison. It shows a nice leanear relationship for the color index, at least in the range of -0.1 to 1.6 for B-V.
After I calculated the color for RS CNC, I got a value of 2.13, significantly redder than 1.35. I find this a bit puzzling, because my measurements of other stars that have published colors as red as 1.5-1.6 did not show any such discrepancies. So I now wonder if RS CNC can get significantly redder than the catalogs indicate, especially when it is in the dimmer part of it's cycle?
I don't have hard data yet, but after trying some tests with Betelgeuse, I think with stars are redder than about 1.7 or so, I am getting enough IR light leaking through the green filter so that it makes the star appear brighter in the green channel. Which in turn causes my V calculations to be brighter then they actually are.
Let me know what you think,
-Mike Durkin (DMPA)
Using Seqplot I found that BSM photometry (41 nights) give B-V=1.667 for RS Cnc. (B-Verr=0.13)
You should ask Arne if you want to know more about the BSM data of RS Cnc. They were probably uploaded into the database Seqplot uses in the last few years.
We put RS Cnc on the BSM queue for an SRc/SRd program for Matt. It was observed at BVRI from March 2010 through January 2011, on about 50 nights. Looking at the data for that year, the color remained nearly constant at (B-V) = 1.7, and I doubt that it would be any different if you looked at it today.
It is an M-star, and so is much brighter in the red. (V-Ic) for example is about 4.1, which means that any filter that has a red leak will be compromised.
My suggestion to DSLR people (as well as unfiltered observers, or any observer with a non-standard filter) is to avoid red stars. There are lots of bright stars of other variability types that are fun to observe and neglected!
ARne
Hi Mike,
The typical "hot mirror" reflects NIR light and transmits visible; the normal use is for projection systems where you want to get rid of the heat. They typically transmit about 80% of the visible light, and get rid of about 90% of the NIR light. Basically, they would be ok with a loss of ~20% throughput, and they do cut down some of the possible red leak, but a factor of 10 decrease of the red light is really pretty minimal (usual blockers decrease the red light by 3-4 orders of magnitude). A better choice (but probably more expensive) is an IR blocker. However, either one is something to try, and see whether things improve!
Don't get too concerned over the 1.37 vs 1.7 vs 2.13 color index for RS Cnc; red stars are really hard to measure correctly.
Arne
Hi Arne, Mike,
In fact there is no need for extra IR filter with present "regular" DSLR. DSLR have both an interferential IR cut at 700 nm and a dye IR filter that reduces the response to deep red and further reject IR. The difference between the Johnson V filter and the G DSLR channel is that the DSLR red response is too low for us, not too strong ! To achieve a near perfect V response we could use an Hoya Y50 filter that will cut the extra blue sensitivity of the G channel, achieving a blue roll-off just right such of the V reference proposed by Mike Bessell. Then to achieve a near perfect red roll-off it's enough to compensate the missing red by just adding about 5% of the R channel output to the G output.
But, if I remember well, Mike has a 450D from which the IR filters have been removed by Hutech ? In this case it should be used with the external or clippable replacement IR filter. Regular IR cut are available for few ten $. Then there is probably no need for the addition of 5% R in this case.
But if I understand well Mike also uses the B channel to determine B-V. If the V from the present DSLR can be excellent, extracting B from the B channel is more difficult. The difference of centroid between both is large. This is probably the reason for the B-V discrepency seen by Mike. I also should denote that there is a large difference between the filters of old DSLR (before the 350D) and the present ones, present ones are much better for us.
Clear Skies !
Roger
I do have a modified Canon Rebel XT (350D) which makes it more sensitive to Hydrogen Alpha and the red end of the spectrum. This causes a red color shift which can be corrected by using a custom white balance.
Well, I think if you do photometry white balance correction won't help. Especially in the case of red variables. If there is H-alpha or other red emission that significantly effect the photometry. So I would not use modified DSLR cameras for photometry except in the case of CV's where the B-V color-index is around zero, but definitely avoid using them for photometry of red stars.
I think there is a confusion here. There are two kinds of modifications of the IR filtering of DSLR. This is linked to the fact the DSLR have a stack of two IR rejection filters. One is a classical dye (tinted glass) similar on red side to the Shott BG39 (used in Johnson filters) the second being an interferential filter that cuts the IR starting at 700 nm. Both combined in regular DSLR provide a very strong IR rejection, even stronger than the single IR filter used in the Johnson sets.
Then the modification of IR filter in DSLR can be the removal of all the stack or only the dye: two different ways.
If the stack is removed the sensitivity to the IR is high, like an unfiltered astro CCD. In this case the photometry would be very difficult as the RGB filters made from pigments deposited on pixels are transparent to IR.
But if only the dye is removed keeping the interferential cut operative, in this case (Mike's one) the rejection of IR remains strong thanks to that interferential filter (similar to what is used in more recent Johnson filters like Astrodon ! ) Only the sensitivity to the red is increased, but not a lot ! The resulting response curve of the G channel is then very similar to the red roll-off of the Johnson V pass-band. By the way I do not see why it would not be usable for red stars like a Johnson V ? The difference is in fact on the blue roll-off side, this could be corrected either with a transform, or my VSF technique, or an Hoya Y50 filter.
Color Balance: the color balance of the 350D doesn't affect the RAW image file output. The RGB response curves of the 350D are identical to such of the 450D. What I have published on the regular 450D (unmodified) is usable for the 350D. I recently made the response measurment of a very recent Canon "M" and I found it unchanged. The Canon filters system has apparently not changed since the 350D.
Clear Skies !
Roger (PROC)
Hi Mike,
As I have said before, TB and TG are not B and V. The filter bandpasses and central wavelengths are very different, and RGB filters are not necessarily IR blocked like Johnson filters are. Unless you carefully calibrate your system, you will get weird results like this.
I don't know of any fields that have bright red/blue standards. For telescopes, the usual fields are SA110 and SA111 along the equator, where there is lots of reddening, so that intrinsically-normal stars become very red objects. This means that they are stable, yet red - the ideal choice for standards. Arlo Landolt calibrated lots of these stars, but primarily faint ones. Bright stars in those regions tend to be nearby stars, with little reddening, and therefore blue.
However, what you need are some fields where there are stars significantly redder than the target star, so that you interpolate rather than extrapolate. Otherwise, I'd recommend that you avoid red stars.
Arne
Hi Mike,
We are setting up some calibration fields, using data from the Bright Star Monitor. These would possibly be useful for your setup, but won't be ready until mid-Summer (after the CCD School). I'll post here when they are available. With the existing data, I expect calibration stars to be in the range of 5-10mag, but only over the BSM 2.2x1.4deg field of view, so useful for some DSLR setups, but not others.
Arne