[Laser] Prism Tuner

[Dave]

> Date: Wed, 30 Nov 2011 13:18:42 -0800 (PST)
> From: Tim Toast<toasty256 at yahoo.com>
> Subject: Re: [Laser] Prism Tuner
> To: "laser at mailman.qth.net"<laser at mailman.qth.net>
> Message-ID:
> 	<1322687922.75700.YahooMailNeo at web37901.mail.mud.yahoo.com>
> Content-Type: text/plain; charset=iso-8859-1
>
> Hi Dave,
>
> That's true there will be some hard reflections from an uncoated prism. And i guess this applies to
> any optical parts that are uncoated. Stray reflections can cause ghosts and in general a lessening
> of contrast / higher noise etc. So you would want to start with a good quality anti-reflection coated
> prism at least.

Yes, although there are some problems:

http://en.wikipedia.org/wiki/Antireflection_coating

To produce a good antireflective coating, you need a material with a 
refractive index
matched to the material it is being deposited on.  However, due to 
material limitations,
it's not always possible to obtain such a match.

Note that broadband antireflective coatings are usually composed of many 
layers of
material.  However, for a single band antireflective coating, it may be 
possible to
achieve this with a single layer.  Thus, it may be cheaper to specify a 
single band
antireflective coating, based on the anticipated frequency of use.

> As you say, most normal spectrographs use collimated light but this is just a simple minimum parts
> setup that depends on the main lens to do all the focusing.

Yes.

> Another type of spectrograph that uses a single
> element is the "objective prism" type. It has the prism outside in front of the main lens and takes advantage
> of the fact that a very distant source (like a star) will have nearly parallel rays and can be considered
> collimated. One disadvantage with this though is you have to have a prism as large as the main lens.

That can be a significant problem if the objective is very large.

> Both of these simple spectroscopes are used in astronomy for getting a large number of spectrums
> all at once from a field of stars in the frame. Low dispersion, low resolution, whole field-image.

But, that's at cross purposes with a communications system.

> Someone on another group pointed out that this simple design (no slit) would "smear" or blurr the
> spectrum produced. In effect, increasing the bandwidth seen by the detector beyond what you
> would expect from a simple calculation based on total spectrum width vs. detector area width.
> At night this might not be much of a problem but in daylight there is much more scattered light and
> nearby objects in the field of view to overlap and interfere.

One of the issues, though, is that for a night time system, you probably 
don't need a
bandwidth filter, well, unless you're doing frequency multiplexing 
(which, I don't think
anyone is currently doing, at least in the amateur area).

> All in all i guess if you have to add lenses to collimate the beam or add a slit, then this probably makes
> a regular bandpass filter cheaper. This was mostly about which could do the same thing cheaper - a bandpass
> filter or a prism. And mostly during daytime use. From what i've read, they decided a regular bandpass
> filter would need to use collimated light and so would need two extra lenses - one on each side of the
> filter. But for reasons of the filter itself not having the same bandwidth for rays entering at a large
> angle compared to a collimated ray. This led to another tuning idea of tilting the filter to tune the bandpass
> a small amount.

Don't forget that the collimating lens should be a fairly small lens. 
The collector lens
will be quite large, but it will have focused the light down to a small 
point where it
encounters the collimating lens.  Therefore, the cost of including the 
collimating lens,
even given that is a antireflective coated lens, should be minimal.

Additionally, the performance gained by collimating the light very well 
may increase
the system efficiency to the point where the size of the collecting lens 
can be reduced,
perhaps resulting in an over all cost savings (Someone needs to do an 
engineering
analysis.).

> As you point out, the photodiode will have a different sensitivity depending on the wavelength. If you were
> using something like this for wavelength and amplitude measurements, you would want to note the
> reduced sensitivity to the blue end of the spectrum. But you would probably want to use a whole-spectrum
> sensitive diode anyway so the difference wouldn't be so great. I can imagine some regular diodes might
> not have any useful response before you even reach the green/blue parts of the spectrum.

I think most photodiodes will have some response all the way up the 
spectrum.  In
general, any photon energy larger than the band gap energy of the 
semiconductor
(or work function of a photocathode) will cause some response, although 
it very well
may not be as sensitive as at the peak sensitivity.  The problem usually 
comes from
lower energy/longer wavelength photons, which have energies below the 
band gap
energy (or work function).

But, such variations in sensitivity probably aren't overly important for 
a communications
system, as long as the signal to noise ratio is adequate for usability. 
  Such variations in
sensitivity would, as you note, be important for a measurement system.

> I was surprised to read, on the other group, that (in daylight) filters seem to help even when using the
> already narrowband LED-as-detectors setups. I guess even they have some noticeable response to out
> of band light.

Even narrorband LED-as-detector setups still have a substantial response 
in the non-peak frequencies (usually for energies above the peak).  They 
tend not to be very
responsive for lower energy/longer wavelengths.  Thus, it can be useful 
to suppress
higher energy/shorter wavelength photons for systems used in daylight.

> -toast

An alternative approach to a prism or diffraction gratings may be to use 
a thin film
coating designed to maximize reflections at a particular frequency:

http://en.wikipedia.org/wiki/Dichroic_filter

It may be possible to build a tunable dichroic thin film filter (either 
by varying
the reflection angle, or by varying the thickness of the thin film 
across the
filter).

Dave

P.S. It's been a couple of decades since I've done much optics work, so 
I'm a
bit rusty.