Connect with us

Speed of light circuits

Discussion in 'Electronic Design' started by Ancient_Hacker, Mar 8, 2007.

Scroll to continue with content
  1. Let's say I want to measure the speed of light using a LED pulser and
    a photomultiplier tube.

    A quick web search doesnt turn up anything relevant.

    I suppose I need an avalanche driver into a fast LED.

    And a fast amplifier on the photomultiplier tube.

    And a good sampling head on the 7S11 plugin. Or I'll have to fix the
    bad power supply on the 7104.

    Now I could guess some ballpark numbers, like a few hundred pF
    discharging through a fast transistor. And maybe a MMIC stage or
    three to amplify the phototube current.

    But of course itprobably takes plenty of optimizing all the little
    details to get the best risetimes.

    Anybody been there, done that, and have any proven circuits?


  2. Phil Hobbs

    Phil Hobbs Guest

    You'll do a great deal better using CW and a phase measurement. Not as
    much fun, though.


    Phil Hobbs

  3. The LED pulse should be no slower than the response of the
    photomultiplier. say 2ns risetime 5ns FWHM.

    I've used a slight modification of the circuit presented
    in "A fast timing light pulser for scintillation detectors"
    Nuclear Instruments and Methods in Physics Research Section A:
    Accelerators, Spectrometers, Detectors and Associated Equipment, Volume
    241, Issues 2-3, 1 December 1985, Pages 612-613
    J. S. Kapustinsky et al.

    It works with some blue LEDs (you've got to match the photocathode)
    Not really. If you use a high gain tube. The pulser will kick
    about 10^7 photons per flash. You can run the photomultiplier
    straight into the scope.
    No idea. Sampling scopes plus photomultipliers are normally
    a bad idea. You've got 20ns transit time on the photomultiplier
    plus ~3ns per metre on the optical pathlength - total delay of ~100ns
    for a reasonable size experiment. Just use an oscilloscope with
    We did this to calibrate the Antares neutrino telescope
  4. John  Larkin

    John Larkin Guest

    Spark gaps can be interesting. Charge up a low-inductance cap until it
    fires a gap. You'll get kilowatts of light with ns risetime. With some
    moderate optics and a pmt (fresnel or curved mirror for the
    transmitter, small telescope for the pmt) you can get echoes off


  5. Oh, that sounds so COOL! An excuse to play with high voltage!
    I'll have to try that as soon as I make it throught the Honey-Do list.
  6. Robert Baer

    Robert Baer Guest

    PMTs seemto be in the 7nSec region unless you want to $pend a lot
    more for (maybe) a X3 improvement...
    Now if you take that light source and bounce it off the moon
    (assuming you know the distance) or off a reflector on a rather distant
    mountain (again assuming you know the distance), then it might be a
    little more practical.
  7. Genome

    Genome Guest

    Some blokes did it a bit back with a vacuum cleaner and a couple of pint

    I'd suggest you are not really an Ancient_Hacker and more like a Lazy_Arse
    if you can't sort that one out using modern tecknowledgy.

  8. I was thinking of the Electron Tubes 9813 and similar (sim. Burle 8575)
    which are ~2ns.
    OEtech have them for USD 215

    I've got a whole load of these looking for a project...
    No, it is quite feasible in a largish room. I've done it.
    I even offered the technology to Ralph S******* as evidence
    that the speed of light was actually c - but he wasn't interested....
  9. Guest

    There is also the method of a moderate speed spinning mirror, a hene
    laser or even a good quality laser pointer, a larger mirror and a
    football field. Needs far less scope bandwidth and only a simple diode
    detector, but its a phase method and not so satisfying.
    You can get the commecial off the shelf kit for ~100$. Just add

    Steve Roberts
  10. colin

    colin Guest

    Are you going to measure the average roundtrip and bounce the light back or
    go for the realy hard part and try to measure it in a single direction ? if
    so how will you avoid the propgation delay of the cables between
    trasmitter/sensor/computer from negating your results ?

    If you use a continuos train of pulses you can use a hp5328b to average the
    difference between two pulses with sub picosecond resolution, with only
    moderatly fast tramsitter/sensors.

    Colin =^.^=
  11. John  Larkin

    John Larkin Guest

    Don't change the cables!

  12. jasen

    jasen Guest

    use two tubes one with a short optical path (possibly take from the side
    of the led) through an attenuator (if neccesary)

    ajdust the pulse frequency until the phases match.


  13. Use pulsed light and do time of flight.
    Rig some contraption with mirrors and corner cube reflectors.
    In one configuration the light takes a short path (10cm); in the other it
    goes to the end of the room and hits a corner cube and comes back (20m).
    You measure the distance with a tape measure to 1mm. You leave
    all the cables untouched and don't screw with the pulser or
    photomultiplier supplies between the two configurations. You measure
    the time delay to a few hundred picoseconds and you have the
    speed of light to less than a percent accuracy. Its easy.

  14. Rich Grise

    Rich Grise Guest

    Instead of a "larger mirror" on the far end, you should use a corner
    reflector/retroreflector. Aiming a flat mirror would be a nightmare.

  15. Iwo Mergler

    Iwo Mergler Guest

Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day