Recommendations for project to measure speed of light

[Kwak]

I would like to try to measure the speed of light, for a high school science fair project. I have no experience with lasers.

The experiment I envision would shoot a small, square-wave-modulated laser through a beam splitter. One portion of the light would be sent directly to a photo detector. The other beam would travel a longer path (probably 5 to 20 meters,) and be directed back onto the same photo detector. The response of the photo detector would be observed with an oscilloscope. If all goes well, I imagine a waveform showing two overlapping pulses, which could be independently removed by blocking each beam. Measuring the overlap would allow calculating the speed of light.

My budget for this is very low. I am looking for recommendations for:

1 - A modulated laser source. Are cheap laser pointers modulated? Do I need to get a laser diode, and drive it with my own circuit?

2 - Would the laser require eye protection? I'm interested it the lowest power that would do the job, mostly for budget reasons.

3 - A photo detector. I am guessing I need a photodiode with a pretty fast response time.

Any advice would be welcome, especially specific part numbers, and sources for purchase.

Did I mention that my budget is low?

Thanks for reading this far,

Kwak

 

[lasersbee]

Have you done a Google search...:thinking:

I seem to vaguely remember a project you could build to test
the speed of light....


BTW...

to the Forum...


Jerry

 

[nikokapo]

I don't know if you can actually measure it like that, climbak probably knows about it. Though you can try to calculate it easily.

Take Ampere's law for magnetism:

Link: http://upload.wikimedia.org/math/2/6/e/26e845d2bf8d061986e114bc51ed8254.png

Thenk Coulomb's law:

Link: http://upload.wikimedia.org/math/e/7/d/e7ded9530e66b14b024639d072d1d40f.png

If you get all the data needed (can be calculated theoretically too), you can see if you can calculate the constants Epsilon 0 (electric constant) and Mu 0 (magnetic constant).

Then to calculate the speed of light:

v = 1/ sqrt(mu 0 * epsilon 0)

Try it with their values (find them with google or at Wolfram|Alpha) and see the value you get. It should be very close to "c"

 

[charliebruce]

If I'm correct, in one microsecond, light would travel around 300m. Unless you have a very accurate oscilloscope and long distance, I don't think you'd be able to get an accurate value. You should be able to drive a low-power laser diode using a standard signal generator and resistor, however. A 5mW module such as the ones Dave was selling should be enough, presuming the divergence is low enough for the distance you want. The photodiode from a laser sled should be sufficient, but make sure you have the option to calibrate the 2 inputs to compensate for any differences in the photodiodes. If you're doing it outdoors, then you'll also need to take into account interference and may need a filter. This should be done with a red diode - green isn't fully stable under modulation, violet may also be suitable but more complicated and harder to focus, and IR may present further hazard to eyes, and would be hard to align.

 

[Benm]

I think you could probably pull off such an experiment using a fast scope... 100 MHz and over models being commonly available at the moment. The main question would be if you can obtain a photodiode capable of detecting signals this fast. I'm sure it can be done with the ones used in fibre optic communication systems, but i doubt the average diode intended towards remote controls and such would be fast enough.

The way you want to go about it is nice i think, something different from the usual interference based experiments.

On the laser end it wouldn't be very problematic at all, you just need a driver that provides a really fast rising edge. Since detectors are sensitive enough, you can afford some overshoot on the activation current, making it all the more feasible.

 

[Kwak]

I think you could probably pull off such an experiment using a fast scope... 100 MHz and over models being commonly available at the moment. The main question would be if you can obtain a photodiode capable of detecting signals this fast. I'm sure it can be done with the ones used in fibre optic communication systems, but i doubt the average diode intended towards remote controls and such would be fast enough.

The way you want to go about it is nice i think, something different from the usual interference based experiments.

On the laser end it wouldn't be very problematic at all, you just need a driver that provides a really fast rising edge. Since detectors are sensitive enough, you can afford some overshoot on the activation current, making it all the more feasible.

Well, I've learned some since my original post, and I'm making some progress.

When I used the term 'modulated', I now know I meant 'pulsed'. I found a driver circuit I understand, build it, and it looks like it will work for me. I should be able to switch it pretty fast with TTL: laserbias

Soon I will turn my attention to the detector.

I stumbled over a page by a guy who's done the whole experiment: Speed of Light with Nanosecond Pulsed 650 nm Diode Laser

Looks pretty sweet, but his detector is beyond my budget. Great results with a 100MHz scope, just like you said. (The scope I'm using is the baby brother, TDS 210, 60MHz. Should be close enough, I don't need perfect waveform reproduction.)

Thanks for the reply, Benm, and everyone else. I'll post again as I make more progress.

-Kwak

 

[Benm]

That link looks like a good approach - probably reproducible as-is. There are several ways of getting a short pulse out of a laser diode, but the principle is always the same, so you could work with that 'avalanche' driver.

I'm not sure about your scope though, as it is a digital model. With analogue ones seeing signals much faster then what they are specced for isnt difficult (shape and amplitude will be way off, but you see there is -somehting- present).

Then again, if you can manage to set up a path difference of 10 meters or so, the resulting shift would be in the order of 30 ns (30 MHz) so it should work out with any 60 MHz scope.

 

[Bluefan]

I tried this too, I've got too litte time to give it a better try, but I can post what I've got so far. I used a laser diode for a ultra cheap bullet style red laser pointer. These only have a resistor, no slow regulator. I took a 5V source and switched with a MOSFET driven by a pulse generator. The mosfet should switch in about 10ns, I can't remember the risetime of the pulse generator, but I guess a few ten ns. The photodiode had a risetime of 3.5ns and was also quite cheap, a few $ at most in a basic led style case. I used a very simple mirror to reflect teh beam back, so the light travelled twice the distance I had.

The total setup had a risetime of 100ns when the laser diode directly looked at the photodiode. The problem was that this cheap pointer had a high divergence, after 15m the beam was spread quite wide and my setup was not that solid, so I lost a lot of light, which ment a much slower risetime, making the ~50ns I needed to find undetectable. So, plans I have: build a stable setup with a lens to pick the reflected light up, correct the divergence of the laser diode with two lenses and take a larger distance, 50m should be enough. That would be ~330ns, which is perfectly measurable if both pulses have a similar risetime. This way the delay between both is better visible as both traces look alike.

I used a HP 54502A scope, 400MSa/S, 100mhz singe shot bandwidth and a BNC model 500 pulsegenerator. The scope triggered on the pulse generator. Saving the wavefrom on the scope ment I can measure without a beamsplitter and with only one photodiode, the signal reference is the pulsegenerator connected to the scope. Then it is easy to measure the wavefom of 0 distance and 50m distance, and the delay between those two.

 

[Benm]

I think the experiment would benefit from using 2 beam paths, just to illustrate how much longer it actually takes for light to make a run around the room. This will require smart triggering though, as currents travel slower in wires than light does though empty space

 

[Bluefan]

I used two different beam paths, I just measured them separately so I didn't need the beamsplitter and two photodiodes. I assumed the electronic delay between the signal and the optical output is the same, so I used the signal to trigger, and displayed both traces from both beampaths to compare them. One live from the long roundtrip, one saved on the scope from the laser diode directly to the photodiode.

 

[Benm]

Seems like a perfectly valid method to me. I think its nicer for a demonstration if you split a beam and visualize the time delay for both halves to arrive (on a single trace), though the result is identical.

What is nice to do: use a single detector and keep the electric trigger connection. You can then keep sending pulses (at say a kilohertz rate or so), and directly see changes in the delay when altering the beam path.

You can even do neat tricks like setting up an aquarium in the long path, and noting that light goes slower through water. It'd probably have to be a very large one to make a notable difference, but you can use a long tube with (very flat) panels of plexiglas on either end... ideally it would work both filled with water and without, and not require any adjustment of mirrors etc between the two.