Following on from my post #318, I started wondering if there was an "optimum" setup for my scope, in terms of a focal reducer and its effect on the spectra. So I ran a few calcs, and found some interesting results (I think).
I decided to calculate the minimum band width I could resolve in a spectrum with the SA, for each FR with my scope. I found that regardless of which FR I used, the smallest bandwidth I could resolve with my scope is 6.25 Å.
At f/10, the maximum dispersion is 1.86 Å/pixel, but the resolution is 3.36 pixels; at f/6.3, the maximum dispersion is 2.963 Å/pixel, but the resolution is 2.1 pixels; and at f/3.3, the maximum dispersion is 5.7 Å/pixel, but the resolution is 1.1 pixels. These all multiply out to about 6.25 Å for my scope and camera.
Note: the above calculations ignore the effect of seeing and imperfections, too. So to do better than 6.25 Å I believe I'd have to introduce a slit to the system...
So the choice of which FR to use has to be from other factors:
the length of extension tubes between the SA and the CCD; and
the number of frames required to capture a given bandwidth.
For f/10, the maximum dispersion is achived when the CCD is 300mm from the SA. That's a long extension tube! And to capture , say, the whole visible spectrum of 4000Å would take about 5 or 6 frames!
For f/6.3, the max distance from SA to camera is a more reasonable 189mm, and the visible spectrum could be captured in about 3 frames.
For f/3.3 (if I could find a suitable FR for the job) , the max distance would be 99mm, and the visible spectra could be captured in 2 frames of the DMK 21AU04.AS.
So I'm a long way from pushing the limits of the SA!
It would be interesting to see if my scope has enough focus travel to handle the extreme distance from SA to CCD. I guess an experiment is the only way to tell.
For the time being (in the absence of an f/3.3 solution) the f/6.3 reducer still seems like the best setup for me ATM. Plenty of scope to explore yet!
Al,
I like your analysis!
Basically I think you're correct, but I'd add in a couple of "real life" variables....
The image on the SA will always be controlled by the seeing conditions, over here its got to be very good to get much below 4 sec star images, and that's for the faint stars; the brighter ones are "bloated"
Also, the larger the out of focus star image on the SA grating, the more lines are illuminated and the better the resolution. Add these to the dispersion and you may get different results???
Yeah, Ken... I did say I was ignoring seeing and imperfections. Seeing should be less of a problem at lower f/ ratio (the star image should be smaller) but the relative effect should be the same in terms of Å on the spectrum.
The defocussed star image at the SA is a full disc of 30mm diameter (that's my assumption for maximum dispersion) since if we go bigger than that then the SA becomes the field stop.
I have the f/6.3 reducer on my scope, and I get about 9 A/pixel with the SA in the filter wheel.
So decided to try my short barlow body to increase dispersion a bit. I could not get a thing on the screen, and got lost driving the motorfocus bacdkwards and forwards so I took it all out and put he diagonal in and my 10mm EP... can't achieve focus...
Oh well, thats a result worth knowing... but I got enough of an image to know I was at the right end of the focus adjustment, so I put the short barlow body and camera back in... adjust exposure to get an image and refocus...
No way! I still run out of focus. So the best I can do ATM is maybe 9 A/pixel with the f/6.3 FR. I might be able to stretch it a little bit by fitting the SA extension ring...
That being the case, an f/3.3 FR wouldn't reach focus at all...
I guess that's what you were saying before, Ken...
This is the first spectrum I captured tonight. I plugged NGC3372 into the scope and it slewed straight to this. I took one look at it and immediately thought I had Eta Carinae. But do I?
The continuum is a very different shape to the raw etas I've captured before - this one is much bluer / hotter... and there is a strong double line bluewards of Hb.
Can anyone verify if Eta Carinae is changing, or have I picked up another star?
Woooh......
That certainly looks odd! Whatever you got is a Be type star, emission in Ha and Hb; don't know what the peaks around 4640/4675 are....
I've taken the liberty of posting on the Spectro-L forum ( for the LiHiresIII guys) see what they know...
Can you get another spectra tonight... and check if it is Eta???
Woooh......
That certainly looks odd! Whatever you got is a Be type star, emission in Ha and Hb; don't know what the peaks around 4640/4675 are....
I've taken the liberty of posting on the Spectro-L forum ( for the LiHiresIII guys) see what they know...
Can you get another spectra tonight... and check if it is Eta???
That's cool. I'll try for another spectrum tonight, and maybe have a hunt around the area.
A thought that struck me this morning before i got out of bed... the blue wouldn't be the effect of scattered moonlight would it? I estimated the limiting magnitude in Crux to be 4.7 (due to moonlight), and Eta is not real bright. Whatever the star is, it is about the same brightness as Eta. I had the same settings as previous Etas wrt to exposure, etc.
I started processing another spectrum of Arcturus as well last night (before VSpec spat me out) and it also looked a little strange due to a lot of blue. I'm sure I had Arcturus this time, but maybe the full moon is a factor?
I have attached some odf the spectra I captured last night: Antares (x2), Arcturus, and the red star at the centre of the Jewel Casket. The spectrum of the red star at the centre of the Jewel Casket also had a strange blue response (like the eta car spectrum). I think that is the result of scattered blue light from the moon on a longer exposure. I have cancelled the blue out during corection for the camera response.
I have also attached a list of neighbours to eta carinae from CdC. There are a few possibilities with PEC spectra.
Al,
I'm guessing this is a situation where the use of an aperture stop in front of the SA would be an advantage...
What do you think would happen if you just put a mask over the front of the scope?
Al,
I'm guessing this is a situation where the use of an aperture stop in front of the SA would be an advantage...
What do you think would happen if you just put a mask over the front of the scope?
cheers
Mark
I dunno, Mark. A mask in front of the scope will reduce the aperture, so it will change the f/ ratio, reduce resolution and then the exposure time making it longer... a catch 22 I think.
I think maybe it's just something to be aware of when capturing spectra of the dimmer stars. And, it's another reason why a standard camera response curve is a problem. A standard response curve wouldn't compensate for an increase in scattered blue light from moonshine. I think I'm starting to realise why it's important to do a camera response curve for each session at least, and maybe each spectral class. In this case, actually, it's the exposure that's important.
Quote:
Originally Posted by Merlin66
Al,
I think you're getting the hang of this game!! Those spectra are great!
If and when I write "the" book I'll ask your permission to use them!!
Thanks Ken. I feel like I'm getting head around it, slowly at least... though sometimes I think I'm doing everything the same but VSpec obviously thinks otherwise and spits me across the room.
I've come to the conclusion that the filter wheel has got to go! I can get the SA 30 to 40mm further up the light train by not using the filter wheel. That gives me 30 to 40mm more back focus. Hopefully that will allow me to increase dispersion with the f/6.3 reducer with my short barlow body or possibly the star diagonal. I feel another experiment coming on.
I think I'll leave my SA in the barlow body if that works out OK. I don't use my 2x barlow much anyway now I have a Powermate. Will that make it a "SArlow"?
Not that I know what I'm doing but I've just had a browse through the finding list on NIST (http://physics.nist.gov/PhysRefData/...dinglist11.htm) and my guess at the origin of these two lines are Nd I, and Al II.
I just browsed the list for about 10Å each side of the peak wavelength and picked the elements with the strongest lines. Al II has two strong lines at 4663 and 4666, and Neodymium has two strong lines at 4634 and 4631.
The Al II seems possible, but the Nd? Seems a pretty heavy element (atomic number 60) so it doesn't seem so likely to me. But having a better guess requires a bit more knowledge than I have ATM.
While a lurker on IceInSpace I'm sorry to say I wasn't aware of this spectroscopy forum ... until alerted to the possibility by Ken's posting on Spectro-L. Melbourne's inner city clouds lifted long enough last night for me to grab a 100s low-res slitless grism spectra ... copy attached. While I'm no expert on this star it does look as 'normal' as can be said for anything regarding this far from normal star. While Eta car does have regular HeII (4686A) variability I don't believe it is anywhere near as strong as your spectra shows, and I don't see it in my spectra (recorded with Sony chip which has good blue response). So I would agree with you in questioning if you imaged Eta Car ... I strongly doubt it. But top marks for trying, and keep up the effort as it needs more monitoring
I forgot to mention that the first thing that made me doubt you had imaged Eta Car was the poor signal/noise ratio of your spectrum ... much more like that of a mag 9-10 star than Eta's current mag 4.7
I forgot to mention that the first thing that made me doubt you had imaged Eta Car was the poor signal/noise ratio of your spectrum ... much more like that of a mag 9-10 star than Eta's current mag 4.7
Bernard
Yeah, its a bit furry now that you mention it. I did notice that at the time but dismissed it as extra background noise due to moonlight, but could just as easily be a dimmer star. I was using the exact same exposure (0.6s) that I normally use on Eta, but I didn't notice the spectrum as particularly dim.
I think a 0.6sec exposure is way too short, unless you have a very large scope and a very bright target. I assumed you were using a C8+DSLR, is this correct and if so, is the DSLR a CMOS sensor? I'm using a C8 with a monochrome, binned, CCD and still need to stack 5 x20sec exposures to get a reasonable SNR.
Ooops, error ... my exposures were 5x 5sec (where the star was included) but I reduced it to 5x 4sec for just the spectrum.
As the attached raw image shows, the star and Ha are saturated but the areas of interest have a reasonable SNR. The stars are out of focus, which is normal when the spectrum is focused.
I think a 0.6sec exposure is way too short, unless you have a very large scope and a very bright target. I assumed you were using a C8+DSLR, is this correct and if so, is the DSLR a CMOS sensor? I'm using a C8 with a monochrome, binned, CCD and still need to stack 5 x20sec exposures to get a reasonable SNR.
Bernard
I used the C8, with an f/6.3 FR, the SA and my DMK 21. I find that with brightness =0, Gain = 1023, and exposure = 0.6s the eta spectrum uses most of the histogram without clipping and the homunculus is just starting to appear in the zero order image.
Quote:
Originally Posted by Glenpiper
Ooops, error ... my exposures were 5x 5sec (where the star was included) but I reduced it to 5x 4sec for just the spectrum.
As the attached raw image shows, the star and Ha are saturated but the areas of interest have a reasonable SNR. The stars are out of focus, which is normal when the spectrum is focused.
Bernard
Very interesting - thanks for posting Bernard. I haven't been exposing my spectra anywhere near as much as that!
I've attached a jpg copy of the spectrum as captured from the eta impostor the other night, and two previous eta spectra including the zero order.
Part of the reason I think I ballsed up the eta the other night was I was practicing taking spectra without the zero order. Maybe I'd have noticed the lack of homunculus in the zero order if I'd done it the old way.
Al.
Last edited by sheeny; 12-05-2009 at 05:15 PM.
Reason: typo
).
) , the max distance would be 99mm, and the visible spectra could be captured in 2 frames of the DMK 21AU04.AS.
I guess an experiment is the only way to tell.
so I took it all out and put he diagonal in and my 10mm EP... can't achieve focus...
What going on?
