[Laser] Prism tuner

Tim Toast toasty256 at yahoo.com
Mon Dec 5 06:06:07 EST 2011


All valid points Dave. 

It looks like filters are winning this contest. I hadn't thought of using a dichroic filter. A red one amounts to a low pass filter kind of. 

And tunable over a small range by tilting if needed etc. I'm not sure of the cost for those vs. a bandpass type filter, they may be 

similar.

One place to get red and green dichroic mirrors, for cheap, is inside the older LCD projection TVs. These are the last of the old-style 

projection TV sets before they decided flat-screens were the way to go. There are also an assortment of good quality mirrors and
lenses insde these sets - also a very bright gas discharge tube and a special "fish eye" projection lens that has a nice flat focal field

- a gold mine of exotic parts.


It looks like, if using a prism for filtering in a communication setup, you might as well go for a full fledged spectrograph complete with 

slit, collimator, prism and condenser. 

On the optical efficiency of a homemade spectrograph verses a bandpass filter - i would guess the filter wins again although being able 

to tune a wide range would make up for it depending on your application. With LEDs, a slit doesn't need to be extremely narrow and 

can let through more or even all the light from the source. A spectrograph needs to have a bandwidth at least as wide as the LED 

spectrum - about 10 to 20 nm wide i think. Also for best efficiency, the spectrograph needs to use a similar F ratio collimator as the 

main collector lens, but it probably doesnt matter as much since we are dealing with point sources most of the time.

For fixed wavelength optical comms, filters should be fine as long as you can find them in the correct bandwidths and frequencies 

and at a resonable cost. And as you say, they aren't needed at night

Here is a Littrow - grating spectrograph ive been thinking about building. 
http://www.aladal.net/toast/litspec.jpg 
http://www.aladal.net/toast/litspecs.jpg
I have all the parts for it except a proper adjustable slit. To tune it you rotate the grating a small amount. The nice thing about a 
Littrow type spectrograph is - you only need one good quality lens. This lens works as both collimator and condenser. The offset
slit and CCD arrangement (with the slight off-axis lens effects) causes some distortion in the focused spectrum but is generally 
smaller than the individual pixels on the CCD and so doesn't really matter. A good spectroscope based on a CCD/webcam would 
make a valuable piece of equipment on any experimentors optical bench.

Here is a good page on homebiulding spectrography equipment:http://astrosurf.com/buil/


-toast



> ----------------------------------------------------------------------
> 
> Message: 1
> Date: Thu, 01 Dec 2011 15:05:55 -0500
> From: Dave <wa4qal at ix.netcom.com>
> Subject: Re: [Laser] Prism Tuner
> To: laser at mailman.qth.net
> Message-ID: <4ED7DE23.40108 at ix.netcom.com>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
> 
>>  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.
> 
> 
> ------------------------------
> 
> _______________________________________________
> Laser mailing list
> Laser at mailman.qth.net
> http://mailman.qth.net/mailman/listinfo/laser
> 
> 
> End of Laser Digest, Vol 86, Issue 2
> ************************************
> 


More information about the Laser mailing list