ref: sivon.txt  	gross words=1817 June 1998

Review of the nu-View spectrometer    by Maurice Gavin

Introduction:  Modern detectors like the CCD camera and sophisticated computer software have the potential to reawakened the field of amateur spectroscopy that largely died a century ago.  The Nu-View spectrometer is thus a welcome addition to the amateur's armoury whilst also having a wider applications in school and college laboratories.   

Modern approach:   Normally a spectroscope (in an astronomical application) is attached direct to the focal plane of a Cassegrain telescope and unless the telescope is fairly massive, can easily upset the balance.    A more common approach nowadays is to feed light to the spectroscope via a fibre-optic cable (FOC) and isolate the spectroscope from the telescope so no weight is imposed on the telescope itself.  The Nu-View takes this approach and can be placed on photographic tripod (the base is threaded for a standard 1/4-inch camera tripod bush) or a convenient shelf adjacent to the telescope. 

The spectroscope is of conventional design and is well engineered.  It contains no electronics and requires no electrical power.   The housing is a substantial 16cm x 10cm x 7cm box machined from solid aluminium and black anodised throughout.   The 3mm thick plate lid is secured with four corner thumb-screws.   The optics comprise a collimating or entrance lens to render a parallel beam of light onto a grating which in turn reflects the dispersed light into an exit lens that brings the spectrum to focus just beyond the instrument's exit port.   A plane first surface mirror between collimator and grating folds the optics into its small enclosure.

Fibre-optic feed:  The 1m long FOC comprises a bundle of 25 separate silica glass fibres - each 50 micron in diameter and protected by an outer plastic sheathing. Threaded rings lock the cable onto nipples on the telescope's 1-1/4 inch eyepiece adapter and spectroscope respectively.  The telescope focuses the stellar target onto the input end of the cable where the fibres are tightly bunched to form a disk 375 micron diameter (about 0.4mm).  This is a much larger area than a traditional slit placed at the focal plane and this greatly eases the problems of target acquisition and guidance during long exposures.   The spectroscopic slit, essential to the design, is formed from the fibres themselves at the output end where they are teased into a single line 50 micron wide by 1.6mm long.   The FOC tips are glass encased and relatively robust with no moving parts,  nor subject to metal corrosion.

The supplied FOC has good spectral transmission characteristics for the visible and near IR region ie about 380nm to 1000nm and this can be recorded by a typical CCD given sufficient exposures at the extremes.  The sharp cut-off below 380nm and near UV region is due to a combination of factors, including glass within FOC, lenses in spectroscope, window and encapsulation of the CCD chip and typically poor CCD UV sensitivity.   Astronomical UV work is also limited by the earth's atmosphere especially at low altitudes.   Approximate one third of the spectrum detectable with a CCD on this spectroscope is in the near IR.   The maker's estimate the gross light loss through the cable is about 50%.  As most of this loss occurs on the input head and first few centimetre of the fibre, longer cables have essentially the same light loss.  Greater light losses may occur with spectroscopes using a conventional slit.

Optics:   The dispersing element creating the spectrum is a 'blazed' 1200 line/mm replica reflective grating approximately 25mm square supported on a 6mm thick optical glass plate. The grating base is epoxied to a rotating turntable - adjustable via a micrometer screw outside the casing.  The remaining optics are remarkably simple but effective - two 50mm focal length coated f/2 two-element achromats serving as collimator (entrance lens) and camera (exit lens).   Each lens is held in aluminium mounting block via an Allen keyed screw and a similar but larger screw locking the mounting block to the instrument base.   60mm or 75mm focal length exit lenses are optional and shift the focus progressively further from the exit port to satisfy most applications.  A set of Allen keys are included in the kit and the lenses are easily interchanged by the user. 

Strange but true...   Because each fibre-optic thread cannot be truly parallel throughout its entire length (manufacturing tolerances or bending) it behaves in a curious way. The f/ratio of a lightbeam passing through the FOC is changed and made 'faster' and this new beam must be accommodated without vignetting by the spectroscope's collimator lens.  For example an f/10 telescope beam fed into the FOC inlet end, exits at the virtual slit at f/5 whilst an f/4 inlet beam reach the maximum numeric aperture of  f/2.2 on outlet.  

For economic reasons the 50 micron wide (0.05mm) FOC slit is relatively large for a small spectroscope and this can adversely effects spectral resolution.  This is not aided by the collimator (entrance) lens and the imaging (exit) lens being the same focal length so that the slit is imaged onto the detector full size ie a ratio of 1:1.   To compensate the spectroscope is supplied as standard with 1200 lines/mm reflective grating to give a higher spectral dispersion and thus longer spectrum.  Reducing the effective slit width as projected onto the detector increases the resolution of the spectroscope. Traditionally this is done by making the collimator of longer focal length than the camera lens.   This is not possible in the Nu-View as supplied.  The net results is the spectrum as viewed or recorded via the Nu-View spectrometer appear very soft and of poor resolution.  In reality the captured image is greatly oversampled for the Nu-View does not give an instant results.  These must be extracted via the supplied software later.   

Common targets:   Spectra can be observed without connecting the spectroscope to a telescope.  and it can be enlightening to just point the input end of the FOC at some common light sources around the home and neighbourhood and view the results with an eyepiece in the spectroscope.   Bright emission lines can be seen in the spectra of fluorescent lamps, tv or computer monitors and mercury and sodium street lights.  Tungsten lamps show just a colourful band of colour called the continuum.  Daylight through a window will show a similar continuum but crossed with numerous dark lines originating in the sun's atmosphere in reflected daylight.

Because of the relatively high spectral dispersion of the nu-View only a small portion of the whole spectrum can be viewed at one time.  Turning the micrometer screw on the side of the spectroscope brings other parts of the spectrum into view.   The visual spectrum extends from about 390nm in violet to about 700nm in far red and a common focus holds throughout the range. The CCD extends the range into near IR but refocus of the exit lens (moving the camera outwards) is necessary.

Viewing and recording the spectrum
The spectrum may be viewed with a regular push-fit eyepiece applied to the 1-1/4 inch diameter exit port or via suitable adapters into most SLR and CCD cameras.  Sivo supply as standard a 50mm focal length exit lens.   

Camera focus is very critical and somewhat crudely implemented by physically sliding the camera back and forth in the exit port via the 1-1/4 inch adapter.  Whilst focusing the camera it must also be rotated so the spectrum run parallel to the CCD's longer side.   A locking nut secures the camera when both are achieved.   A supplied dial gauge helps to reregister the camera focus when subsequently removed and replaced.   However the dial gauge will only interface with the large front plate (cooling fin of cameras like the ST-6) and with other cameras is ineffective .  

The supplied exit lens can be removed altogether and the image recorded with a lens attached to the camera (see photo adjacent) applied to the exit port.   This permits precise and easy focusing via the lens's helical mount.  However this arrangement causes some vignetting as ideally the imaging (exit) lens should be as close to the grating as possible.   The gap between lens and spectrometer can be covered with a black cloth to stop extraneous light spoiling results.   Precise alignment of camera and spectroscope can become more difficult when they are physically separated as described. 

At the telescope:   Setting up the spectroscope and coupling it the telescope is best done in daylight or under good lighting. To avoid entanglement of the FOC the manual advises removal of the eyepiece adapter before slewing the telescope.   The finder is accurately aligned with the FOC target area - the tip of the lunar crescent makes a good reference.  Now aiming at a bright star the telescope focus is adjusted whilst viewing the downloaded images.  Typically a star when sharply focused will occupy only part of the 0.4mm diameter FOC target area and the downloaded spectrum will appear as a series of random horizontal streaks - there is no correlation between the fibre ends at input and exit.   

For a brighter target it is prudent to slightly defocus the telescope image (but no larger than 0.4mm) so the spectrum is fully illuminated but of lower intensity so the image is not saturated.  Using my 30cm Meade LX200 the longest exposure attempted was 20 minutes (on Beta Lyrae H-alpha region) and the bright emission lines were clearly visible.  To record other portions of the spectrum the micrometer is adjusted and the procedure repeated.  An evening's output will be fair low as stars targeted get fainter.

Extracting results:   The 1.6mm high slit projected onto a CCD may occupy more than 100 rows of pixels providing a large sampling area where (after dark and flatfield calibration) remaining defective pixels make a very small contribution that may otherwise create false spectral 'lines'.  

Pixels are binned in columns (under software control) to extract the complete signal and product an intensity plot of the spectrum.  Because the spectrum may be a 100 pixels 'high' data transfer into a spreadsheet for plotting is advised.   In this way 'binning' is not limited by the CCD's A/D converter to 4096 (12-bit) or 65,000 (16-bit) values but may have a much greater dynamic range especially in 'highlights'. MS-DOS Software for conversion to FITS and spreadsheet readable format is supplied.  

Conclusions:   The $1625 price tag [June 1998] may come as a surprise for a small box with two simple lenses, two mirrors (one a grating) and a fibre-optic cable.  There are no electronics or moving parts except a simple devise for rotating the grating to bring various parts of the spectrum into view.   However in a modern 'disposable' age the spectrometer's simplicity is its major asset and with care it will give a lifetimes service.   This equates to a few cents a day which is a bargain for a precision instrument able to reveal many 'wonders' to the enquiring user.    
===============end of text=================