Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Speed of Light

  1. Nov 5, 2007 #1


    User Avatar

    I have been searching the net for measurements and how they are done of the speed of light. For some reason, I don't find much on them. I see sites promoting various theories whether sped up or down or is not constant. Claiming prehistorical differences in the speed of light is unverifiable I presume.

    I am looking for negative proofs of the variability of light speed in a vacuum, i.e. it is constant.

    SR states that speed of light is the same in all frames from all sources. Is that a definition or verified fact? If it is a definition (rather than fact) then of course the Lorentz transformations have to be made. If it is fact, the same is true.

    Have any measurements actually been done on the "instanteous" speed of light as opposed to simply dividing the distance by time to arrive at the speed? I see claims of faster than light group velocity. Is that really any different from simply picking an underlying lower amplitude signal out of a higher frequency signal superimposed upon it?

    Also, light is deflected by stars due to gravity "wells." Have there been exercises done that assign some mass to a photon that would cause the same deflection as in gtr and then prove it doesn't weigh that much?

    I see it argued that the law of causality requires light to be constant. I don't see that. "Causality" is broken all the time. Just go a baseball game. We see a ball being hit before we hear it. Of course there is an unlying principle of the speed of sound being less than that of light. Arguments are put out that variability of the speed of light could allow a situation where you could arrive before you left. Seeing something from a distant point after seeing the same object nearby proves nothing, since seeing is responding not to reality but light which travels at a finite speed.

    I see a site stating that the periods of Io's eclipses varies (from an Earth observation) according to whether or not we are approaching or receding from it, thus claiming the speed of light changes (as received, not emitted). Given that the distances are changing, the times should change as well, so does not prove anything since the math is missing.
  2. jcsd
  3. Nov 5, 2007 #2


    User Avatar

    Staff: Mentor

    A google for "measure speed light" turns up just under four million hits. Here is the second: http://math.ucr.edu/home/baez/physics/Relativity/SpeedOfLight/measure_c.html
    That would be the Michelson Morley experiment.
    I don't know of any way to measure the "instantaneous" speed of light. Why does that matter?
    The deflection due to gravity doesn't depend on the mass of the passing object. That's why g is constant (9.8m/s).
    Could you cite the site? Timing the eclipses of Jupiter's moons was one of the first ways the speed of light was measured (as you can see in the link I posted). Either you misunderstood your source or the writer of it misunderstood the experiment.
    Last edited: Nov 5, 2007
  4. Nov 5, 2007 #3
    gtw: Be careful how much you say about something you know so little about, since you come off as a bit of a crackpot. If you have one specific question, by all means ask it here. If you have an entire field of physics you don't understand, go loan a university textbook.

    Stellar aberration is not redshift-dependent: this exemplifies a measurement of the "instantaneous" speed of light that was emitted from sources of different velocities.

    Gravitational deflection is independent of (relativistic) mass.

    Apparent variations in the period of Io were used to calculate the speed of light (with evolving accuracy); quite likely you just completely misunderstood many of these sources (which you have conveniently omitted citing).

    [Russ, that was quick! Hey.. MM only measures local anisotropy..]
    Last edited: Nov 5, 2007
  5. Nov 5, 2007 #4
    The speed of light is one of the most accurately determined constants known. Its constancy independent of the motion of the source or observer, though counter-intuitive, has been verified for decades .. more than a century even. This constant is so good that we no longer measure using metres defined elsewhere. Instead, we now use it to define the length of a metre.

    The speed is exactly 299,792,458 metres per second, and any future experiment to greater accuracy will adjust the length of a metre, not update the speed of light.

    Do have a look at http://en.wikipedia.org/wiki/Speed_of_light[/PLAIN] [Broken]

    There are many, many devices that have to allow this truth for them to function (start with modern GPS navigation). This behavior of light is known. It is not up for debate.

    The (awkward) consequences for explaining it is why I read these forums. We are better at quantifying it, describing it, getting the maths to model it, etc. than we are at coming to terms with it. The whole reason we use the word "relativity" is because, like throwing a ball from the back of a moving truck, we expect speeds to be sensibly counted including the speed of the source and observer, and light just does not do that! It gives us problems when we consider two objects coming towards us at somewhat more than half light speed, and whether the anticipated passing happens when we expected, and why an observer on one does not have a mass infinite, even when contemplating the other.

    I have more delicious counter-intuitives, but have a ways to go before I tax anyone with them. :smile:
    Last edited by a moderator: May 3, 2017
  6. Nov 5, 2007 #5


    User Avatar

    Didn't that just disprove any ether or global frame of reference? The source and receiver were fixed relative to one another so that wouldn't apply would it?

    I know that some of these questions come off rather basic, but I am unable to find experiments trying to find a variability in light speeds. I assume that any made would have a negative result. I still haven't found a controlled test. There has to be a simple one that would eliminate all the conflicting chatter.

    I understand any questions along these lines might seem "crackpot" given the large number of sites out there purporting one thing or another with varying amounts of detail and format.

    That's a good point.

    You wouldn't want to know. I pulled it up and it mentioned the experiment. It had its own agenda to push.
  7. Nov 5, 2007 #6
    may have misinterpretated you here but causality is a consequence of the speed of light being constant not the other waya round and the speed fo light being constant is not obtained from laws or theories it is a well established experimental fact. Causality in relativity only means no information can travel faster than the speed of light not that it ahs to travel at the speed of light so don't see where the baseball analogy comes in as useful.

    This site has taken a correct experimental result and twisted the results to a false answer. the difference in lengths of eclispes when the earth is approaching io and when we are retreating is due to the difference in travel time between the start of the eclips and the end of the eclipse due to the shorter(or longer) distance between us and io. This infact with the correct maths proves the speed of light is constant and gives a rather accurate value to it.

    As for a proof of the constancy of teh speed of light as the guy above said, the michelson - morley experiment and this has been replicated to higher degrees of accuracy since.
    Last edited: Nov 5, 2007
  8. Nov 5, 2007 #7


    User Avatar

    I just read the Michelson-Morley experiment in some detail. It seems to me that the mirrors could be considered to be light emitters in sequence. Since the reflected light is opposite to the generated light why couldn't this simply be a null experiment in the absence of ether?

    I just thought something else. Light emitted perpendicularly from a moving object. Does it move in the direction of emission or is there a component along the direction of movement? I would think it should move in the direction of emission or would it stay in the same inertial frame?

    So finally I found something that is a good independent (no mirrors, no change in directions) test of the absolute speed of light. I had to go down a number of links to finally find it. That would seem to refute any other ideas. Plus the experiment measuring when pion decay into photons (all showing similar speeds).

    So finally I can shut down this line inquiry.

    Thanks for your patience. Most of my questions have been answered.
  9. Nov 6, 2007 #8
    This is what they got, they got no fringes. The interferometer measures differences in teh speed of light along perpendicular pathlengths, you will get a different speed of light moving along the same direction as the ether as one moving perpendicular to the ether, or any sin x of an angle somewhere in between. BEcause they got no interference patterns the onyl answer is that teh speed was the same along the two interferometer arms, so tehre was no ether. It wasn't to emasure the speed of light, only the speed of the ether relative to the earth.

    Light from a moving source in my example is going to be a spherical source emitting equally in all directions is changed along the z and y directions even if the source is moving only along the x axis. The light is effectively coned in both directions along the x axis(the axis of motion) ie both positive and negative.

    you can do the maths for each particular incident direction of the photon in the sources rest frame using the standard lorentz boost equations. Take the derivatives of them and then divide your dy or dz(parallel axis components which remain unchanged) by dt(the derivative of your time dilated time t which depends on the component of velocity in the x direction) so the resultant equation dz/dt must also depend on the speed along the x direction.

    Sorry can't remember the equation at the moment
    Last edited: Nov 6, 2007
  10. Nov 6, 2007 #9


    User Avatar

    I was looking at the Twin Paradox where the stationary Twin is older than the moving one upon reuniting. However, if they are both moving in opposite directions,(out and then back) the relative speeds would be the same as in the normal example. I would imagine they would be the same age when they came back together. So, the relative aging couldn't be due to the velocities but the accerlation. Is that right?

    Or one could put them on opposite orbits, so their acceleration would be constant and opposite around the circle. So the angular velocities would stay constant. (I never see things put in these terms.) The point of these exercises is to better understand where the effect occurs.

    I already commented on the Michelson-Morley experiment below (to me it seems like mirrors can be treated as emitters in the opposite direction which cancel out any test of c+u speeds with c-u speeds at the receiver, thus proving nothing).

    I looked into the Thorndyke experiment but couldn't get enough detail to analyze it. Seemed to have the same issue though. That is if treated mirrors conceptually as emitters is valid. The photons are absorbed and reemitted I would suppose.


    I was trying to get at a way to eliminate the single source from the determination of light speed. It seemed that the speed derivations required outside assumptions of distance and speed that accounted for by assuming a constant speed, thereby nullifying the test. The binary star solution got rid of that by having two sources at approximately the same distance with two speeds showing light arriving at a constant speed. I believe armed with the constant speed in all frames, the standard formulas just fall out handling two frames by simple algebra with no other choice.

    For the perpendicularly moving light, I was thinking more of a directed laser flash. I found through a proper search query "light perpendicular" etc. and found (hopefully not some other non-reviewed theory):

    So that question is hopefully answered.

    If causality is being defined in terms of light, then ok: it is self-defined that way. I was treating the idea of causality independently.

    The point I was making about causality was that people can "break" it or think that they do until they have an unlying reason why it isn't broken. There is an idea about lining people and having each shout HI to the next as soon as they see the previous one doing it. This is supposed to show that a sound "signal" can be propogated at faster than sound. It is a bogus example, since they are not responding to sound, but light (sight). This was put forward as an analogy to light group velocity in wikipedia but I have seen it before.

    I read thatt group velocity is supposed to show faster than light information movement with the lead package at c. I am guessing but it really just seems like a lower frequency curve supporting higher frequency changes (like a temperature graph over a year with daily fluctuations on top of the seasonal changes.) My point is that it just seems like something that falls out of the math. (Unless the changes propogate forward within the group.)

    WRT to Io. Correct. The site did not provide any math. Seemed to me that the additional distances would account for his conclusions. The binary star tests and the pion tests are the best test I've seen for eliminating additional variables when calculating constancy of speed.
  11. Nov 6, 2007 #10


    User Avatar

    Staff: Mentor

    Then do yourself and us a favor and don't repeat such nonsense here. It appears you know you are doing it. We're getting tired of this game.
    Last edited: Nov 6, 2007
  12. Nov 6, 2007 #11
    No it is not the accelaration, yes if they both followed exactly the same path in opposite directions and both turned and returned their clocks would show the same time still. However this will be different to a third person(triplet?) that was left behind on earth.
    The easiest way, i think, is to imagine a third observer moving at any speed constantly the whole time. In the classical twin paradox with one leaving turning arounda nd returning the third observer will see the twin who is left behind always moving at a constant velocity whereas the twin who turns around changes relative speed. So the turning around twin must be the one who's time is not "proper time".

    This is more complicated, and i'd never thought of this so thanks for it. I'm gonna think on it more but my first thoughts are that of course no age difference will occur, the third observer(O) this time positioned at the centre of their orbits would see them performing the exact same motion around him so they msut have the same time dilation with respect to O's clock.
    The other thing that needs thought of here is that they are actually in inertial frames as defined by SR because gravity is only a pseudoforce so SR concepts cannot be applied easily to this example.

    The mirrors can be considere emitters only if you also remember the photons they emit will be exactly in phase at the time of emission(impossible if they were in honesty the original emitterrs of the photons) If there was an ether and our apparatus was moving through it in the x direction of our x-y arms then there will still be a path difference with the x path being (c+u)t where u is the speed of the ether along the x axis and the y being perpendicular to the ether motion will be just ct this will cause fringes on the detector, the only way to get no fringes is if there is no path difference, so there must be no lumineferous ether. Michelson morley experiment never set out to find the speed of light or even show that it is constant.

    In SR causality is always defined in terms of the speed of light

    My personal favourite if the travel of information appearing to be > c is : consider a wave hitting a shore line at some angle to the normal you line the shore with detectors with synchronised clocks that record the time the wave reaches them. If the wave is almost parralel to the shoreline then the detectors will have very short differences in time and when you take into account the distance between divided by delta t then it looks like the information of the wave breaking was possibly transmitted down the line of detectors faster than c but what you have to remember here is that the information is travelling in the direction of propagation of the wave, ie almsot the normal to the shoreline, not the direction of the wavefront of the wave ie almost the direction of the line of detectors.

    WRT group velocity, yeh in certain mediums you can get patterns in light and sound waves that seem to propagate at speeds greater than the speed of light in a vacuum or even with the right setup you can have a group velocity in the opposite direction to the propagation of the photons, this is all an illusion no information is passed at the group velocity, each photon is still moving at c so all is well.
  13. Nov 6, 2007 #12


    User Avatar

    I didn't repeat it, that's why I didn't cite it originally. Actually, when I started looking into this, I had no idea so much "nonsense" was out there. Sometimes when you are searching for something using keywords, this is what you get. Some are presented with lots of math purporting this or that looking pretty professional.

    Finally I found the binary star example which matched a question I had asking if a test with light sources on a rotating system being observed by a stationary receiver.

    When the science is presented in college, the major tried and discarded ideas (and why) are rarely mentioned. Rather than simply excepting it, I was trying to intuit (hard to do of course) my way through some of the problems.

    Anyway, thanks for your time. I found what I think is the incontroverable test.
  14. Nov 6, 2007 #13


    User Avatar

    Thanks. Someone treating all of this as being a crank. I read sr only works for inertia states. Perhaps it just has nothing to say for this situation. (Can it be restated using angular velocity and put into the sr equations?) It seems always that the clocks need to be adjusted at the turnaround spots to correct the math. For example (Wikipedia: Twins paradox):

    I was trying to eliminate that, since I suppose nothing can be said during periods of constant velocity (or so it seems the presentations of the Twin Paradox imply). There is constant relative acceleration away and returning to the source in this way (no starting or stopping either to restate the problem).

    Of course, having a third observer just degenerates back into the classical case. I believe something going through an orbit will undergo time dilation (relative to a fixed location) even though I suppose there are no constant inertial states, witness the 1963 mu Meson experiment. Nothing I am able to solve now. I just wondering if other people knew how these other formulations of standard problems were dealt with.

    Several people have brought this up as a test of the constancy of speed of light. I know it was only a test of the ether. My point was: wouldn't distance along the x arm be

    2ct = (c + u)t + (c - u)t,

    because of the return from the mirror? If so then the effects of the ether would cancel out. ct = distance between emitter/receiver and mirror.

    I am assuming that the distance of the receivers and emitters are the same distance from the mirrors. All the discussions and diagrams show the emitters and receivers being the same distance from the mirrors. I took this as fundamental to the experiment. If the distances were different then the interference might be seen I would think and my point moot. This is my reasoning anyway.

    Thanks. That's all.
  15. Nov 6, 2007 #14


    User Avatar

    Staff: Mentor

    If the ether flowed along the X axis, it would cancel out for motion along the X axis. But it would make the Y distance longer.

    The usual analogy is a boat traveling across a river. If the river's flow is the Y direction and the boat points directly across the river (in the X direction), it still moves in the Y direction. Thus the total distance traveled is greater than CT.
    Last edited: Nov 6, 2007
  16. Nov 6, 2007 #15


    User Avatar

    Good point. I have never seen that mentioned. Simply the extension in the x direction.
  17. Nov 12, 2007 #16
    You can look up the Fresnel experiments on measuring the speed of light in a moving medium. The experimental results agree with the relativistic prediction. Light in a medium travels at less than c, so you can expect the speed of the conducting medium to have an effect as measured in an inertial lab frame. Yet the velocities do not simply add the way Newton would have predicted.
    Last edited: Nov 12, 2007
  18. Nov 12, 2007 #17


    User Avatar

    Staff: Mentor

    Actually this is Fizeau's experiment, not Fresnel's. Fresnel predicted the value of the "drag coefficient" that this experiment should produce.

    Better versions of this experiment have been done since Fizeau's time (1850s, I think). There are references in the Usenet Physics FAQ.
  19. Nov 12, 2007 #18


    User Avatar

    Thanks, that is an interesting effect. I'll look it up. Perhaps the "speed" of the light was slow enough that it was below relativistic speeds but still showed "relativistic" properties rather than Newtonian.

    This is all for my own edification. The notion of time dilation is easy to accept and hard to understand as to why and how.
  20. Nov 17, 2007 #19
    Perhaps a little colourful illustrative scenario will help you some. It was , I think, due to Archibald Wheeler, and I can only offer a half-remembered, paraphrased version..

    This is about locating an event (such as switching on a light), in space and in time, and further, fixing the interval, between two events, such as a fast-moving chap switching a lamp on and off. Fixing the position of the events is a matter of distances in three dimensions, and clock times. The argument is that when the values are large enough, we will find discrepancies in what happens compared to what strictly Newtonian mechanics would predict.

    Imagine a mythical community in a walled city with a main gate, and somewhere within, a well. The sage surveyor and his assistant at times re-measure and refine their record of well position relative to the gate, which they do by keeping records of distances North, and East of the well relative to the gate. Imagine further that because of centuries of religious and cultural tradition, all measures North are in miles, and all measures East are in yards. No matter, they keep recording, this being the only way they have.

    Then assistant, being radical heretical, decides to to figure all distances in yards, and notices that provided one apples a constant to turn the miles into yards, the numbers he sees allow him to note [tex] Interval^2 = Northings^2 + Eastings^2[/tex]
    Its not even too tough to figure the height of a pole at the well with a version of

    Applied to events, in analogous fashion, if we are going to describe a spacetime interval, we need to have the fourth dimension (time?) expressed in the same units as the distance. So apply a constant to the seconds to convert them to metres. Then (crudely) a 4-dimensional version of Pythagoras becomes viable. The 'constant' is the speed of light. We do notice, however that the sign of the time part is unexpected.
    For our own cultural reasons, we still cannot routinely express time in metres.

    Its an old story, and nowhere fully provides what is needed. Make of it what you will, it helped me expand my view of the world at the time
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook