new SX spectrograph

[nytecam]

I've got on loan from Starlight Xpress one of their new spectrographs that's been some years in the planning ! Weather's been foul so only got a spectrum of sunlight via the OSC Lodestar with a camera lens up front as shown below.

The SX spectrograph is an interesting design eg a modified Rowland Circle spectrograph with concave toroidal 500 l/mm reflective grating forming the spectrum at the camera's flat focal plane. A full spectrum from UV - NIR [340nm-900nm] is 31mm long and, whilst contained on the top-end SX full-frame camera, most of the spectrum can be recorded in segments even with the modest SX Lodestar on the sliding rear carriage. A 10%/90% beamsplitter prism serves a built-in Lodestar auto-guider, Ar/Ne calibration lamp and 6-slit carousel to the grating.

No SX announcement as yet but hope it's of interest ;-)

Nytecam

[PeterW]

That's a good solar spectrum... Can you add an eyepiece,..? What height is the slit? Ie could you do drift scan solar imaging?

I am looking for a low cost device to see a nice high res solar spectrum.....

Looks interesting, wonder what the price will be.

Peter

[DaveS]

Very interested in this. Saw it at Astrofest and was very impressed, especially if it can split the sodium D lines. On a time I've considered making something using a concave holographic grating from Edmund Optics, but the cost and sensor / software stuff put me off. If it comes in under £2k I'll be very surprised.

[Gasman]

Interesting Maurice, does it have options for different gratings ?.

Thanks for posting.

Steve

[nytecam]

Thanks guys for your interest. I've made a short video

showing the general working of the instrument and sample stellar spectra from the provisional SX Handbook and one of mine shown above.

==================

Peter - don't think the spectrograph will be cheap by any means - it's primarily designed, in its amateur application, for faint stellar spectra and a spectral resolution to match. The smallest slit is 20micron [w] x 1mm [h].

The sun is a million times brighter and much higher resolution is thus possible like my Ebert spectroscope I built a decade ago @ http://home.freeuk.com/m.gavin/ebert.htmwhich, despite its simplicity, matched Questar QMAX by placing solar lines within the D1/2 sodium duo !

=========

Dave - pricewise you're probably not far off - daren't ask !

=========

Steve - no the grating is not interchangeable and precisely matched to the design.

Pic below shows light paths for autoguider/calibration lamp @ 10% and spectrum @ 90% via beam-splitter prism.

Hope that helps ;-)

[JeremyS]

That's a very helpful video, Maurice - many thanks for posting it.

Please keep us informed oh how you get on with the new spectrograph.

Go well!

Jeremy

[PeterW]

Lhires and L200 come in under 2k... Hoping for lot lower... Hey I don't even need a calibration lamp for the sun....

Cheers

Peter

[nytecam]

Lhires and L200 come in under 2k... Hoping for lot lower... Hey I don't even need a calibration lamp for the sun....

Cheers

Peter

what you need Peter for solar visual is the Lhires Lite @ 1300€ ! But you should be able to DIY it for a fraction of that if you've any skills :-)

[robin_astro]

Hi Maurice,

An interesting design. The reflective optics should give good performance at the uv and ir ends

The ~4A resolution with a 20um slit would potentially put it up against the LISA for example, though that has fast optics so particularly good on larger scopes. Do you know the focal ratio of the optics in the SX design ?

The choice of  beam splitter for guiding is interesting. (One was used with the L200 but abandoned in favour of the mirror slit system currently used in all current commercial amateur spectrographs AFAIK)  Beam splitters need careful and stable alignment to keep aligned and parfocal with the slit and with a 90/10 split, it might struggle to guide with fainter objects, say >mag 9-10 and say a typical 8-10 inch scope

Using it at 900nm is going to need an order filter of course. Is this incorporated in the design ?

Cheers

Robin

[nytecam]

Thanks for those comments Robin. I suppose any designer will set certain perameters like here ..

100% reflective optics.

Fixed and flat focal plane.

cover the full spectrum 340-900nm in their largest full frame camera SX-M35/6 in one hit.

Use their Lodestar auto-guider as standard

The 10% light to autoguider from the beam-splitter is a loss of ~2.5mags but suspect they are confident; the same loss to calibration lamp is inmatterial ! Haven't done the maths on focal ratios of to/from grating yet but Handbook says good for f/5 scope to max of f/3.5 at inlet.

The solar spectrum I posted via their smallest sensor [Lodestar] covered the visible spectrum. A larger sensor would extend that into NIR ~900nm and <340nm. Not sure what is recorded into UV except in the labs under UV illumination! No reference made to an order filter in NIR as yet.

The SX design [like the Ebert] causes astigmatism perpendicular to the dispersion which effectively widens the spectrum into 2D for instant viewing. In so doing the spectrum, for a given exposure, is to a degree weakened but SX suggest this avoids loss of data through saturation. Hopefully the fully signal can also be recovered via 'compression' to 1D in software.

Interesting times ;-)

Nytecam

[nytecam]

Been tanking down with rain all day and 'holed-up' - so play with SX spectrograph on fluorescent lamp in kitchen and feeble daylight through skylight in mono [inset] - note the ghost fluorescent in the daylight spectrum marked with ticks !

Note also the neutral colour from the Lodestar in the colourless near-infrared spectra to right beyond red ;-)

Nytecam

[robin_astro]

The SX design [like the Ebert] causes astigmatism perpendicular to the dispersion which effectively widens the spectrum into 2D for instant viewing. In so doing the spectrum, for a given exposure, is to a degree weakened but SX suggest this avoids loss of data through saturation. Hopefully the fully signal can also be recovered via 'compression' to 1D in software.

Hmm..  an interesting strategy by SX turning an aberration into an apparent advantage ! Saturated spectra  is a "problem" I would love to have  .

Yes  the full signal can still be measured by summing over all the rows so not a big issue but the extra rows will still add noise from the camera and the sky background so not ideal when trying to measure faint objects. The astigmatism also presumably means that it would not work so well as an imaging spectrograph ie to show structure in extended objects.

The other practical problem in the presence of astigmatism is how to best focus the star image on the slit for maximum throughput. Without astigmatism you can get close by focusing for the narrowest spectrum but this would not work here. Alternatively with a mirror slit you can focus for minumum overspill of the slit but that would not work with the beamsplitter guider design here  so I assume you need to rely on the guider image being exactly parfocal with the image at the slit.

Robin

[nytecam]

Robin - you never had a saturated spectrum ..wow ! It will depend on the energy curve on the target and it's almost inevitable if combined with the weakest section on a complete spectrum. If one is recording just part of the whole spectrum then its rarely an issue. Shame that detectors/sensors are not linear from UV to NIR!

You are correct that field spectrum without the slit (SX spectrograph has one 2mm? aperture for same) that astigmatism of the image will blur results.

We'll have to see how things pan out - everything is compremise.

Nytecam

[michael.h.f.wilkinson]

Very interesting read

[Macavity]

Good stuff. I intend to re-buy a simple P.H. diffraction grating.

My spot-size was always a bit large for quantitative measurement,

but it was ever interesting. And it seems natural to combine VA with

spectroscopy to cope with the (inherently) reduced intensity...

[robin_astro]

Robin - you never had a saturated spectrum ..wow ! It will depend on the energy curve on the target and it's almost inevitable if combined with the weakest section on a complete spectrum. If one is recording just part of the whole spectrum then its rarely an issue. Shame that detectors/sensors are not linear from UV to NIR!

Hi Maurice,

I dont see how introducing astigmatism really helps in that situation though. If the target is bright enough and with a high dynamic range, I just expose for the brightest part and take multiple exposures to get the total counts high enough in the fainter regions to give the SNR I need. The total exposure for a given counts will be  the same in both cases except that I will have included less sky backgound noise. For faint targets where saturation is not an issue, astigmatism just means more sky background and read noise.

Cheers

Robin

[NickK]

Thanks for those comments Robin. I suppose any designer will set certain perameters like here ..

100% reflective optics.

Fixed and flat focal plane.

cover the full spectrum 340-900nm in their largest full frame camera SX-M35/6 in one hit.

Use their Lodestar auto-guider as standard

The 10% light to autoguider from the beam-splitter is a loss of ~2.5mags but suspect they are confident; the same loss to calibration lamp is inmatterial ! Haven't done the maths on focal ratios of to/from grating yet but Handbook says good for f/5 scope to max of f/3.5 at inlet.

The solar spectrum I posted via their smallest sensor [Lodestar] covered the visible spectrum. A larger sensor would extend that into NIR ~900nm and <340nm. Not sure what is recorded into UV except in the labs under UV illumination! No reference made to an order filter in NIR as yet.

The SX design [like the Ebert] causes astigmatism perpendicular to the dispersion which effectively widens the spectrum into 2D for instant viewing. In so doing the spectrum, for a given exposure, is to a degree weakened but SX suggest this avoids loss of data through saturation. Hopefully the fully signal can also be recovered via 'compression' to 1D in software.

Interesting times ;-)

Nytecam

I saw this at Astrofest and had a good look and chat

You've missed saying it's got reference lighting built in for calibration.

If you set to dither on the mount - you can do a 3D spectral image. Then you can remove the sky glow by subtracting the skyglow spectra from the image spectra.

Also this technique means you also can select or adjust which bandwidths you want in the image, then render accordingly.

[nytecam]

I saw this at Astrofest and had a good look and chat

You've missed saying it's got reference lighting built in for calibration.

If you set to dither on the mount - you can do a 3D spectral image. Then you can remove the sky glow by subtracting the skyglow spectra from the image spectra. Also this technique means you also can select or adjust which bandwidths you want in the image, then render accordingly.

Thanks Nick - the calibration lamp reference is excluded from you extract and is in the pics and video. Your last para needs a diagram! In the SX spectrograms in the video you can spy LP emission lines which I understand Mt Wilson used in the past as a second reference spectrum:-)

Nytecam

[nytecam]

Good stuff. I intend to re-buy a simple P.H. diffraction grating.

My spot-size was always a bit large for quantitative measurement,

but it was ever interesting. And it seems natural to combine VA with

spectroscopy to cope with the (inherently) reduced intensity...

Hi Chris - the fact the SX spectrograph produces a live 2D and familiar spectrum has impact - to see a stellar spectrum on download is amazing especially if it's an emission object like a newly discovered nova and you're the first :-)

Nytecam

[robin_astro]

If you set to dither on the mount - you can do a 3D spectral image. Then you can remove the sky glow by subtracting the skyglow spectra from the image spectra.

Also this technique means you also can select or adjust which bandwidths you want in the image, then render accordingly.

Hi Nick,

Not sure what you mean by a 3D image but there is no need to dither perpendicular to the dispersion at acquisition to produce a nice looking spectrum , that just harms signal/noise. Once you have the spectrum, all  data reduction is normally done using software for example using ISIS, VSpec, RSpec, BASS etc. Sky background removal (typically using data from above and below the star in the slit) is a standard step in data reduction and all of these programs will produce wide spectra and render colours according to wavelength if you need a pretty spectrum. 

Or do you mean this kind of spectral imaging, scanning the object of interest ?

http://www.astrosurf.com/buil/scan/demo.htm

There could be problems using this spectrograph in this way though if it shows significant astigmatism and if the slit is only 1mm long as Maurice suggests, the field will be rather small.

Robin

[nytecam]

As a comparison for SX spectrograph I found some old solar spectrograms from my DIY project of 36yrs ago on Kodak 2475 film eg long before CCDs were available. My colourised Magnesium snippet below - the full spectrum to this scale could fill by backyard.

Of course the SX spectrograph isn't designed for this scale of work - my full colour solar spectrum above was a 'first light' project ;-0

Nytecam

[Gazabone]

I'm just starting down the road of spectroscopy and am using a star analyser 100 with Rspec. For the moment that set up is good enough to get me going and I'm enjoying this extra dimension to my astronomy. Seeing this makes me think that at some stage I may want something rather better!

[robin_astro]

I'm just starting down the road of spectroscopy

Talking about down the road  , are you coming to the BAA workshop on spectroscopy at the Norman Lockyer Observatory in Sidmouth  on 10th October?

http://britastro.org/vss/Spectroscopy%20workshop%20programme%20V2.pdf

There will be lots of information on how to do real science with spectroscopy using a wide range of spectrographs from the Star Analyser up.

Cheers

Robin

[Gasman]

Looking forward to it Robin . You likely to be doing any observing/spectro on the Friday night weather permitting?

cheers

Steve

[robin_astro]

Hi Steve,

Not sure if anything official has been organised yet but I suspect  quite a few us will be arriving on the Friday so we should definitely meet up at least. I understand the observatory is open to NLO members on Friday nights in any case so it should be possible to do some observing too :-)

Robin