Let me guess ... for fringe locking? I'll bet you could get at least a few wavelengths of motion at low voltages from a piezo-ceramic speaker element. These can be found in just about any handheld electronic gadget that beeps, bleeps or whistles. Or Radio Shack probably has some, most likely in the form of piezo buzzers. You might have to do some hacking (removing the case and electronics) to get at the bare piezo element, usually a gold-metalized thin ceramic disc. DigiKey sells several kinds, including some bare elements from a company called CUI STACK. I've got a bare piezo element kicking around somewhere. When time permits, I'll try gluing a mirror to it and see what it does in an interferometer setup.
So I ripped a 1.25 inch dia. piezoceramic element out of a long defunct "key tracker" - a silly gadget that you clip to your keychain and when you clap or yodel or whatever for your lost keys it goes beep-beep - maybe. Glued the disc with epoxy blobs at 4 corners to an angle support, then glued a 1 inch front-surface mirror with another small epoxy blob on the other side of the disc at center.
Drive circuitry was a 9V battery with a 10K potentiometer across it, pot center tap and ground going to the piezo. I checked it out as an element of a Michaelson interferometer. Path lengths were about 1 inch, the incoming laser pointer beam being diverged by a 1 cm focal length microscope eyepiece lens. The bullseye fringe pattern was stable with the piezo mirror in place, and when I changed the drive voltage from 0 to 9V, I observed that the central spot went from bright to dark and bright again for 1 full wave of change. I was expecting more sensitivity - the piezo makes a slightly audible click when attaching the drive voltage so you would think the travel would be much more than 650 nm - maybe the piezo and mirror mountings are not optimal or I'm misunderstanding the sensitivity of my setup. Anyway - it does seem to work.
Remounted the piezo, this time gluing it to a ring so
it was glued aroung the entire perimeter. Now I get
3 waves of travel for a 9V drive voltage change.
I just now tried looking for movement in the spot from the reflected beam at 6 meters distance and was unable to see any at all over the 9V drive range. I marked the position of the central spot on a post-it note stuck to the wall at one drive extreme, then compared the other extreme. Spot size was about 2 mm. If the smallest movement that I could have detected was say 0.25 mm, this places an upper bound of about 2 millidegrees on the pointing stability. More accurate measurements would have to be done outdoors, I guess. I wonder if there is a more accurate interferometric way to measure angular change independant of linear motion?
"a silly gadget that you clip to your keychain and when you clap or yodel or whatever for your lost keys it goes beep-beep"
You might not like it, but I´ve been searching for something like that for a long time. Also to find out how something like this works. Could give me some details about it - f.e. where you bought it etc. - if you still have any, of course - it seems to be quite old.
If you do a search on Google for "sonic key finder" you'll find lots of sources, lowest price is around $3.95. Using the piezo transducer as a microphone, it listens for a whistle or a handclap pattern, then starts beeping. In my testing, it didn't work very well for keys buried under books or behind sofa cushions and I felt like a damn fool wandering around clapping my hands and whistling. There is a more expensive ($50) RF based device at http://www.sharperimage.com that might work better.
I still have not given up on the TEC and decided not to make a fringe locker but I do have it in the back of my mind.
I am hoping to make a scanning interferometer so I can directly measure the line width of some of the lasers I have. I would need 4 matched Piezos, a sine wave generator, a photo detector and an oscilloscope. I think it could all be put together for a few hundred dollars. Has anyone worked on this?
One of the specs important for fringe lockers is the angle change when you apply voltage. Most cheap piezos will change your angle and mis-align your spatial filter.
Christoph has been helping me design a ring laser for some time. Here is something he posted to alt.lasers about scanning interferometers:
Hi Colin and everyone else,
I have used several scanning Fabry-Perot Interferometers in the past,
and I must admit, I am getting a bit confused by the discussion.
The first question is: Why do you need such a high scanning range in
the first place? If you want to use a true Fabry-Perot, then it will
give you a peak after every half wavelength scanning path. I think
that's the important thing to remember, that the Fabry-Perot gives a
repetitive signal, as it scans.
It seems to be kind of hard to explain without figures. The problem
is, that most of the textbook explanations consider a thing
Fabry-Perot (like the one in http://hyperphysics.phy-astr.gsu.edu/hbase/phyopt/fabry.html ), where
the light comes in at an angle and produces fringes. For laser
testing, you would come in at right angle, have a big gap between
mirrors (I used between 50mm and 250mm ones) and then scan the
distance by a minute amount.
If you assume, that you have only one very fine laser line
(single-frequency operation), then you will get a strong transmission
(almost 100%) when the distance between the two is exactly a multiple
of lambda/2. If we take Colin’s 635nm laser and assume it works
single-frequency, then you get a peak every 317.5nm of scanning, or
three peaks, if you scan 952.5nm (or let's say a micrometer). I
actually just found good pictures, look at the first two of: http://physics.nist.gov/Divisions/Div842/Gp5/admc.html . The second
one looks like what you would get on the oscilloscope, if you have a
single-frequency laser and scan ~1 micron with a constant speed.
The important parameter of the Fabry-Perot is the "free spectral
range" (FSR). It is the spectral range it scans through from peak to
peak. It is given by:
FSR = c/2L
With:
L: Distance between the two mirrors
c: Speed of light (in the medium, in case of a monolithic Fabry-Perot,
but for Colin it would be just the standard speed of light in
air/vacuum)
The FSR is NOT the resolution; it's the spectral range the Fabry-Perot
covers in one scan. The above gives you the FSR in frequency (Hz, if
you use m/s for c and m for L). If you use 10cm mirror spacing, the
FSR is 3GHz (3*10^8 / 0.1). To get it in wavelength, you can convert
that to:
FSR (wavelength) = lambda^2 / 2 L
You have to be careful, to use the same units for lambda and L!
For Colin, that's 0.002016 nm.
Now, if the bandwidth of the laser is bigger then this, then it will
fill the whole scan and overlap, so that you cannot measure the true
width. If the laser has just two lines spaced by more then this, they
will look like two lines closer to each other.
However, it doesn't matter where the laser operates, the Fabry-Perot
will always see the line. It is really an instrument for relative
measurement (drifts, for example) and line widths.
The resolution is somewhat harder to calculate. As a rule of thump,
you will get a resolution about 1/100 of the FSR with 98% reflectivity
mirrors and careful alignment. Alignment is critical, and the beam has
to be very well collimated (at least for two flat mirrors). For the
10cm Fabry-Perot above, this would be 30MHz resolution. The
"calculator" at http://hyperphysics.phy-astr.gsu.edu/hbase/phyopt/fabry2.html#c2 gives
0.0000130 nm, which is 0.6% of the FSR, but that's the theoretical
limit.
Anyway, your single-frequency laser line width should be well below
this (at least for single-frequency DPSS lasers, I don't know for
single-frequency diodes). You can really only use it to confirm,
whether it is single-frequency, for how long, and whether it drifts.
But don't forget that an unprotected open-air Fabry-Perot drifts as
well. That’s why some commercial ones are in Invar-tubes and
then the whole thing in a thermally stable box. However, if you do
manage to couple your laser into a Fabry-Perot, and then slightly
change the setting on the temperature controller, you will see the
peaks running across the oscilloscope.
I have spent far too long on this now and have to get back to the lab.
Just some practical remarks:
- From the above, it should be clear, that you don't need a long
scanning range PZT. 1 micron is more then enough, and most PZTs will
do that at a moderate voltage.
- You could use two 98% reflectivity output-coupling mirrors. For
1064, they are stock items from CASIX, for example (see http://www.casix.com/optics/ndlaseroptics1.htm ). I don't know,
whether you could get 98% HeNe output-couplers.
- The big problem is coupling the laser into the Fabry-Perot. If you
use flat mirrors, you need a very well collimated beam at absolutely
right angle. I used to have both mirrors on mirror mounts and align
them alternating. And I used one collimating lens, which I moved up
and down between the laser and the Fabry-Perot. That works well. The
advantage of flat mirrors is, that you can vary the distance from
sub-mm to >10cm.
- The even bigger problem is, that the Fabry-Perot acts like a 100%
mirror when not on resonance. This will mess up your laser, unless you
use an isolator or a polarizer and a lambda/4 plate. I did use the
latter, which also worked very well. But it might be beyond your
budget.
- An alternative is a confocal Fabry-Perot, where two curved mirrors
are used with a spacing exactly equal to the radius of curvature. If
you then go in with the laser beam not at the optical axis, but
slightly offset, then you get a "figure of eight" in the Fabry-Perot
and the back-reflection will be at an angle. I just found a figure:
figure 2 in http://www.davidwoolsey.com/ucb/phys/111/fabry-perot/fabry-perot_inst_1.1.html
(this should give you a lot more to read anyway, I haven't read
anything myself, so I cannot comment on the quality)
Unfortunately, they don't draw the reflected beam, but it is easy to
see, that it will be at an angle. The disadvantages are that you have
a fixed length and that you have to get the spacing right to a very
small tolerance.
Because of the four passes instead of two, the FSR is
FSR = c / 4L
If you do not go in offset, or the spacing is wrong, you might end up
with
FSR = c / 2L again without noticing. However, if the back-reflected
beam is at an angle, and if you get a good signal, then you know you
got it right.
Enough!
Some more links I found for the enthusiast (again, I haven't read
anything, so I cannot comment):
alt.lasers has been very busy with homemade scanning fabry-perot interferometers. As I understand it the work is a result of the information provided by Christoph to my querries on the subject taken to a prototype by Sam. The mirrors are from dead HeNe tubes. The parts are simple and can be made with a drill press. Sam claims to have $3 invested into both devices shown here.
Here are some things from Sam's site:
The plan for the first one.
The photo of the first one.
The plan for the second one.
The photo of the second one.
Christoph's diagram of a confocal scanning fabry-perot interferometer.
The circuit to amplify the photodiode so it can be viewd on the ocilliscope.
