Physics Forums

Physics Forums (http://www.physicsforums.com/index.php)
-   General Physics (http://www.physicsforums.com/forumdisplay.php?f=111)
-   -   The SR Question of the Century (http://www.physicsforums.com/showthread.php?t=16196)

Martin Miller Mar13-04 10:52 AM

The SR Question of the Century
 
The SR Question of the Century
=================================
=================================
Why has no one simply used two clocks on a table
to measure light's one-way speed?
=================================
=================================

(Note: No, Roemer did not do it; he used slow
clock transfer, where one clock spanned the
Earth's orbit, and since moving clocks run
slow, his clocks were asynchronous.)

For those who may need it, here is a detailed
version of my question:

=================================
Given two new atomic clocks still in their
shipping crates, how would anyone use these
clocks to measure light's one-way speed in
a way that is proved to be correct?
=================================

Basically, I am speaking simply of the one-way
version of the Michelson-Morley experiment.

And if you happen to believe that the MMx
round-trip experiment somehow even _implied_
one-way invariance, then you need to prove
this via the critical case of a frame that
moves wrt the light source.

The given question is extremely important
because an incorrect measurement of light's
one-way speed means incorrect measurements
of all one-way speeds and all time spans
involving two or more clocks.

The given question is also extremely important
because special relativity (our current theory
of time measurement) is based solely upon
Einstein's light postulate, which claims light
speed invariance in the one-way case.

russ_watters Mar13-04 04:49 PM

How long must the table be and how accurate must the clocks be to satisfy an aether proponent?

Also, since GPS uses one-way transmissions, doesn't it qualify as a good test of the invariance of the speed of light?

outandbeyond2004 Mar13-04 10:51 PM

GPS good test? I don't know, so many effects are involved, I would hardly think it is "clean".

Here's a "Living Review" of the many effects involved in designing and maintaining the GPS:


relativity.livingreviews.gps


I think you have to assume one-way light travel goes at c, because of the issues involved in keeping the clocks synchronized as you separate them.

These issues are not important because in the solar system gravity is so weak and velocities so low, I guess.

Martin Miller Mar14-04 11:37 AM

The SR Question of the Century
 
'outandbeyond' is right; in the GPS case,
as in all current 2-clock cases, scientists
simply assume one-way invariance and isotropy.

Since my question pertained to theoretical physics,
it matters not how long the tables are, and the
clocks are really ideal clocks.

Let me put it to 'russ_watters' this way:

Given two ideal unstarted spatially-separated clocks,
show how they could be correctly related temporally -
bearing in mind that proof is required, and also
bearing in mind that no assumptions are allowed,
and further bearing in mind that you must show this
for a frame that moves wrt the light source if light
signals are involved.

I claim that so far no one has ever correctly related
any two clocks in any frame. (Thus, no one has ever
correctly measured any one-way speed, any momentum,
or any two-clock time span.) (If Einstein had known
how to correctly relate clocks, then he would have
had absolute simultaneity instead of his merely
relative simultaneity, and SR would never have been
created.)

Hurkyl Mar14-04 12:22 PM

Quote:

also bearing in mind that no assumptions are allowed
I can't even assume I can turn the clocks on? That they're at rest with respect to each other? That if I send light signals from one location at regular intervals that the destination receives the signals at regular intervals? That there's not a michevious devil between the two clocks that will intercept all communications and transmit fake data?


Quote:

If Einstein had known
how to correctly relate clocks, then he would have
had absolute simultaneity instead of his merely
relative simultaneity, and SR would never have been
created.
Sounds like you've already decided SR is wrong; I doubt any argument or experiment, no matter how sound, could convince you otherwise.

outandbeyond2004 Mar14-04 02:26 PM

Absolute simultaneity was assumed pre-Einstein's relativity theory. However, Einstein showed that this is something that has to be carefully defined in operational terms; say, partly by all observers agreeing that one and only one of them will be the arbitrator of which events are simultaneous and which not. He didn't acutally reject absolute simultaneity; just that it is something that has to be agreed on in operational terms beforehand, if it is to be used at all.

SR does not always need absolute simultaneity. GR theorists often define hypersurfaces in space-time, each of which has a unique time: i.e., defined absolute simultaneity. However, the definition can change from treatment to treatment.



I have been trying to think of a reason why it is necessary to specify that the clocks be unstarted. So far, i don't see that it is necessary.

wisp Mar14-04 02:29 PM

The SR Question of the Century
=================================================
=================================================
Why has no one simply used two clocks on a table
to measure light's one-way speed?
=================================================
=================================================
Good question Martin.

Don't get fobbed of with the GPS stuff or synchronization issues. Scientists are spending a fortune sending probes to Mars to try and find water, life, etc, surely this simple low cost test should be done.
I have no idea why nobody has done it, but I think the results will be surprising.
Two similar tests were done around the 1990s
1. Roland DeWitte used an electrical pulse to measure one-way speeds and found SR at fault (he is an amateur and so was ignored).
2. Krisher et al (JPL) did a test but the results were too noisy to be conclusive. However, they claimed it supported SR! (They are the professionals and so their results were accepted as good).
See http://www.kevin.harkess.btinternet....ay/oneway.html for more info and also go to "reasons Einstein was wrong"

outandbeyond2004 Mar14-04 04:13 PM

Perhaps this should be explained: Either two events are simultaneous as seen by an observer, or they are not. Einstein's point is that events seen as simultaneous by one observer are not necessarily seen as simultaneous by any other observer. Is this what Martin Miller meant by mere "relative simultaneity"? If so, won't he please explain, contra Einstein, why all observers can see two events as simultaneous?


Ordinary distance measurements require simultaneity. If one had a measuring stick, then we could just line up points on the stick with the endpoints of the interval whose distance we wish to measure. However, if a one-way light-speed measuring procedure requires a distant place, then we would have to transport a clock there, correct?


Could table-top measurements of the one-way speed be made with today's technology? I guess not. But Martin Miller has made ideal clocks (and by definition? ideal measuring rods as well) available to us. So, let me see if I can ...

Put a one-meter stick on the table top. Put one unstarted clock at one end and the other at the other end. Nah, won't do. We need to start the clocks and make sure they are well synchronized. Then put one clock at one end and the other clock at the other end. We can design read-outs so that any observer standing in front of the table and equidistant from the clocks can monitor them to ensure that they are synchronized throughout the measurement process.

Release a light pulse at time = sometime and get the time = sometime + traveltime when the pulse arrives at the other end. One-way speed is then 1 meter/traveltime.

Neat, huh? However, H.L. Mencken said something to the effect that every neat, obvious, simple solution is apt to be wrong.

outandbeyond2004 Mar14-04 04:18 PM

People are trying to see if light speed varies from place to place; from time to time; from direction to direction (anti-isotropy).

See this "Living Review" article on experimental tests of GR:

relativity.livingreviews.experiments

Martin Miller Mar15-04 01:56 PM

The SR Question of the Century
 
---

'outandbeyond2004' wrote:
"Put a one-meter stick on the table top. Put one unstarted clock at
one end and the other at the other end. Nah, won't do. We need to
start the clocks and make sure they are well synchronized. Then put
one clock at one end and the other clock at the other end. We can
design read-outs so that any observer standing in front of the table
and equidistant from the clocks can monitor them to ensure that they
are synchronized throughout the measurement process."

MM replies:
You did not tell us how the clocks are to be synchronized, and your
midpoint observer cannot verify correct synchronization unless he can
prove that he is not moving either toward or away from the light rays
from the clocks' readouts.

(To see that he may be moving wrt the light signals, simply add
a light source at each clock beside its readout, with the light
sources' frame moving wrt the clocks' frame, and let each source
emit light rays simultaneously with its nearby readout. This makes
it clear that only one frame out of an infinity of frames will
remain at rest relative to these sources, which in turn means that
only one frame out of an infinity of frames will remain at rest
in relation to the sources' light signals, which in turn means
that only one frame out of an infinity of them will remain at
rest in relation to the clock readouts' light signals. Bear in
mind that light's source independency guarantees us that the
light rays from a clock readout and a nearby light source will
travel together as if they were one ray.)
---

'Hurkyl' noted:
"Sounds like you've already decided SR is wrong; I doubt any argument
or experiment, no matter how sound, could convince you otherwise."

MM replies:
Try me.
---

'outandbeyond2004' noted:
"I have been trying to think of a reason why it is necessary to
specify that the clocks be unstarted. So far, i don't see that it
is necessary."

MM replies:
This was just an initial condition, not a final one. (I mentioned
it only to make sure that the experiment started from scratch;
the experimenter must of course start the clocks at some point,
but how and when are parts of the given problem.)

'outandbeyond2004' also noted:
"Perhaps this should be explained: Either two events are simultaneous as
seen by an observer, or they are not. Einstein's point is that events seen
as simultaneous by one observer are not necessarily seen as simultaneous by
any other observer. Is this what Martin Miller meant by mere "relative simultaneity"? If so, won't he please explain, contra Einstein, why all observers can see two events as simultaneous?"

MM replies:
Yes, that is exactly what I meant by "relative simultaneity," but I can
still see that a little more explanation is necessary here. The key
questions here are (a) Since events are observer-independent, why should
observers' mere _viewpoints_ of light rays from events be involved? and
(b) What is physical cause of Einstein's relative simultaneity?

To answer the last question first, the precise physical cause of the
relativity of simultaneity is simply different frame movements in
relation to the light rays from the events. No matter how two events
may actually occur, observers in different frames who view light rays
from the events _must_ see the rays arrive differently. For example,
Frame A's observer may move toward the light from Event 1 but away
from the light from Event 2, whereas Frame B's observer may move
away from the light from Event 1 but toward the light from Event 2.
(And it should be apparent that this answer also answers the first
question.)

Of course, simultaneity would not be relative if truly or absolutely
synchronous clocks were used to time events.

As I mentioned earlier, Einstein's clocks are not synchronous. In
fact, he explicitly admitted this in his book _Relativity_. Here is
how he put it:

quote from Einstein's _Relativity_:
" But an examination of this supposition would only be possible if
we already had at our disposal the means of measuring time."
[Ref.: http://www.bartleby.com/173/8.html]

In this brief sentence, Einstein said a lot. Since the context was
light's one-way, two-clock speed, he was admitting that he could
not correctly measure this speed. Also, he was admitting that he
could not correctly measure any other one-way speed because, as
he said, he did not possess the means of (correctly) measuring
time. Further, he was admitting that he could not correctly
measure any two-clock times. In short, he was admitting that the
clocks of special relativity are not truly synchronous.

We do _not_ need a theory whose clocks are asynchronous.
We _do_ need synchronous clocks.
---

'wisp' noted:
"Good question Martin."

MM replies:
Thanks. And as soon as its full meaning has become apparent, you will
see that it is the most important question re flat space-time physics
that can be asked.

See the following to see why no rotating clocks can yield anisotropy:
http://www.geocities.com/antirelativ..._Analysis.html

The only real test of Einstein's light postulate is a direct and
simple measurement of light's one-way speed between two same-frame
clocks which are correctly related temporally. (No rotating clocks;
no transported clocks; no asynchronous {Einsteinian} clocks.)
---

'outandbeyond2004' further noted:
"People _are_ trying to see if light speed varies from place to place;
from time to time; from direction to direction (anti-isotropy)."

"See this "Living Review" article on experimental tests of GR:"

"relativity.livingreviews.experiments"

MM replies:
I saw nothing there about correctly synchronizing clocks or correctly
measuring any one-way speeds.

Here is what one must do to prove that one's two-clock time measurements
are correct:

[1] One must prove that one's clocks are correctly related.
[2] One must prove that one's clocks are not slowed.
[3] One must prove that the rod connecting the clocks is
not contracted.

So far, no one has presented a proof of any of the above; therefore,
no one has presented a proof that even mere _relative_ speeds can
be correctly measured, such as the speed of a bug relative to a log!

outandbeyond2004 Mar15-04 02:19 PM

Martin Miller, have you studied the lunar laser ranging experiment? There should be time-varying components in the data making possible an analysis to determine the one-way speed of light. I admit a lack of familiarity with the state of art: I do not know if such an analysis is possible now.

Martin Miller Mar15-04 03:39 PM

The SR Question of the Century
 
To 'outandbeyond2004':
Since LLR (Lunar Laser Ranging) is a one-clock measurement of the
time taken by the light signal to travel to the moon and back, no
synchronization is involved.

To fully understand what itís all about, we need a proper and
simple example, such as the following one:

[clock]---------[rod]---------[mirror]
[source]~~>ray

Imagine a rod in space which has been ruler-measured to be
1 LY long. (Ignore the hardships involved if this were actually
done.) Picture a started-and-running atomic clock at one end of
the rod, and a mirror at the other end. Imagine a light source
that is moving relative to the clock-mirror frame. As this light
source meets the clock in passing, the former emits a light ray
toward the mirror. It is at this point that we must be careful;
i.e., we must ask ourselves What are the consequences of the
relative motion between the clock-mirror frame and the light
source frame? Clearly, if the mirror moves away from the light
source, then the light ray from the source should take longer to
reach the mirror than otherwise. Similarly, it is clear that if
the mirror moves toward the light source, then the light ray from
the source should take less time to reach the mirror than otherwise.
However, if the rod is physically contracted, and if the clock is
physically slowed, then these two physical distortions will yield
an incorrect result, namely, round-trip light speed invariance.
Of course, this incorrectness could not be revealed by a ruler
because the ruler would also be physically contracted. Therefore,
an LLR measurement will always seem to obtain very accurate
results of the ('instantaneous') distance between two objects, but
this ignores the fact that (a) no one has taken into account the
objects' movements in relation to the light signals, (b) no one
has proved that the clock is unslowed, and (c) no one has proved
that the rod is uncontracted.

In other words, the Michelson-Morley experiment did not correctly
measure light's round-trip speed, and the LLR cannot correctly
measure the distance to the moon. (Given undistorted clocks and
rods, we would find that light's round-trip, one-clock speed
varies.)

Here are the differences and similarities of the Michelson-Morley
(MMx) round-trip case and the one-way case:

Round-trip Case:
The MMx null result is a law of nature; however,
the MMx result was incorrect because Nature distorted
the instruments.

One-way Case:
In the one-way case, there can be no law of nature
because Nature cannot synchronize clocks; however, man
can synchronize clocks, and if he synchronizes them
_correctly_ (absolute synchronization), then he can obtain
a _correct_ result. (Disclaimer: Of course, he would have
to mathematically correct for clock slowing and rod
shrinkage, but we have the formulas, so this is easy!)

outandbeyond2004 Mar15-04 06:19 PM

Martin Miller, your response does not show that you know how useful time-varying data can be. For the benefit of other readers if not yours, let me give an example.

A laser based on earth is shooting photons at a mirror that is orbiting earth. Let's pretend we do not know exactly what the speed of the photons are, so we can calculate only roughly the round-trip distance to and from the mirror. Nevertheless, we continue to gather data, and we notice that the round-trip time varies with time. There is a diurnal dependence, a monthly dependence and an annual one, among others. Theorists develop an elaborate theoretical model of the experiment with many adjustable mathematical parameters, including one for the speed of light. When we have many data points spread over a period of many years, we try to fit the model to the data, adjusting the model parameters as necessary to match the data with the predictions of the model.

If the time-varying components of the data are not important, the fitting of the model to the data would not work, of course. One reason might be too much noise. However, if the quality of the data is good, then we may have an interesting or "tight" range of values for each parameter including the speed of light.

Actually, every experiment with important time varying data is modeled that way. Other examples are the Shapiro round-trip radar time delay experiment and binary pulsars. Many other ongoing experiments depend on the speed of light, so I feel fairly confident that we do have a good grip on its one-way value even though we have not yet measured it directly.

For details see the article I had linked to earlier and have relinked here for convenience:
relativity.livingreviews.experiments

Loren Booda Mar15-04 06:27 PM

Martin,

Such clocks' asynchronicity may realize the GR effect of the expanding universe on nearly local time, or even the curvilinear movement of the observer himself relative to each clock.

Martin Miller Mar16-04 08:01 AM

The SR Question of the Century
 
'outandbeyond2004' wrote:
"Many other ongoing experiments depend on the speed of light,
so I feel fairly confident that we do have a good grip on its
one-way value even though we have not yet measured it directly."

If we correctly word your above, then we have the following:
"Many other ongoing round-trip experiments depend on the
round-trip speed of light, so I feel fairly confident that
we do have a good grip on its round-trip value."

We certainly have a firm grip on light's round-trip speed - it is
invariant and isotropic; however, as I will prove below, the round-
trip speed has nothing to do with either light's one-way, two-clock
speed or with clock synchronization.

The conceptual keys behind the following experiment are two, namely,
(i) a ban on Einstein's rigged clocks (which are _forced_ by a
baseless definition from man to obtain one-way 'invariance'), and
(ii) the use of two frames sharing one light source (something
which is never done in relativity texts).

The physical facts upon which the experiment is based are also two,
viz., (i) the simple fact that light's motion through space is not
infinitely rapid, and (ii) the corollary that light must always
take a finite amount of time T to move between any two points.

The Apparatus (2 tables, 4 clocks, 1 light bulb):
[diagram given below]

>>> There are two narrow, adjacent tables, each 30 meters long and 1
meter high (per a ruler at rest wrt each table). Table F is fixed
relative to the lab, whereas Table R can roll relative to the lab.
>>> There is an unstarted clock [C] each end of each table.
>>> There is a light source [S] (e.g., a light bulb, represented by the
asterisk) aligned with the left end of Table F at table-top height.

At the start, Table R's left end is to the left of Table F's left
end, as shown:

...[C]=======Table R=======[C]
........[S]
........[C]=======Table F=======[C]

In the 2nd scene, the source emits a light ray to the right just
as the left clocks meet in passing and start on zero:

........[C]=======Table R=======[C]
........[S]~~>ray
........[C]=======Table F=======[C]

In the 3rd scene, the light ray reaches Table F's right-hand clock:

.............[C]=======Table R=======[C]
........[S]~~~~~~~~~~~~~~~~~~>ray
........[C]======Table F======[C]

In the final scene, the light ray reaches Table R's right-hand clock:

.................[C]=======Table R=======[C]
........[S]~~~~~~~~~~~~~~~~~~~~~~~~>ray
........[C]======Table F======[C]

Since not even a light ray can be in two places at once, and
since (as was mentioned above) not even a light ray can move
infinitely rapid), the ray must take a finite amount of time
T to travel between the two right-hand clocks.

Here is what the experiment tells us about light's one-way
speed relative to each of the two frames:

***************************
Light's one-way speed wrt Table F
= 30m/100ns = c

Light's one-way speed wrt Table R
= 30m/(100+T)ns =/= c
***************************

In other words, experiment shows that light's one-way speed
varies with frame velocity.

So how did Einstein obtain his one-way invariance? He did so
by incorrectly placing the _same_ time x/c on both right-hand
clocks when they were 'hit' by the light ray.

Why did Einstein force clocks to obtain one-way invariance?
He did this because he improperly extrapolated invariance
from the round-trip case.

Ironically, exactly none of the above has anything at all to
do with the real problem which is How to correctly measure
time? (And this problem was also ignored by outandbeyond's
talk of "how useful time-varying data can be.")

In other words, the real problem for today's space-time
theorist is How can clocks be correctly synchronized?

Here is a hint: Without using more than one clock, one
must be able to guarantee that the clock-starting entities
have truly equal speeds relative to the clocks.

Let's quit talking about "how useful time-varying data can be,"
"how great the GPS system is," "how wonderful it is that we
have atomic clocks," "how the MMx took away the aether," "how
Einstein made time relative," "how wonderful special relativity
is," "how the MMx proved round-trip invariance," etc., etc.,
etc., and let's concentrate on the real issue of absolute clock
synchronization because as of now, we cannot even correctly
measure the speed of a bug relative to a log.

russ_watters Mar16-04 09:29 AM

Wow, Martin - what you suggest is a test of relativity that ignores the effects of relativity. You completely miss the point of SR by not applying it to your apparatus: the moving table experiences a tme dilation and as a result, it still takes 100ns (according to the moving clocks) and still measures C to be C.

This is why I asked you before how big your apparatus had to be and how accurate the clocks had to be to test this. Add also how fast the apparatus has to move and how you synchronize clocks in different frames of reference. Try doing the calculation: I think you will find that for an apparatus like the one you suggest, time dilation (or your "T") is less than the precision of the experiment and thus outside its ablity to detect. But with a big enough (and fast enough) apparatus, you will detect a time dilation.

And if you plan on arguing against time dilation, please remember that time dilation is no longer just an untested prediction of SR, but is in fact, experimental data.

outandbeyond2004 Mar16-04 01:20 PM

Martin, I don't suppose you ever got caught speeding. You'd be put in an asylum for the insane if you tried to convince the patrolman with statements like this:

\\
as of now, we cannot even correctly
measure the speed of a bug relative to a log. //

Good driving, Martin!

Kidding aside, it wasn't Einstein's original idea that the speed of light is constant to every inertial observer. Rather it is a theoretical consequence of Maxwell's equations for electromagnetism. These equations in turn are based on the results of countless experiments that were done long before Einstein's birth. Einstein did have the original thought that the laws of mechanics should be recast to be based on the same speed of light constant as the laws of electromagnetism are.

Martin, your problem may be that you are basing your theory on your everyday experience, instead of experiments done by others. It can be difficult to appreciate how every working GPS device is confirming the truth of GR and SR. It is indeed hard for people to accept that the speed of light is absolutely the same to every inertial observer. So very, very bizarre.

Martin Miller Mar17-04 02:07 PM

The SR Question of the Century
 
Note to russ_watters:
Time dilation cancels out in the experiment because the
two tables speeds through space are equal.

outandbeyond2004 noted:
"Kidding aside, it wasn't Einstein's original idea that the
speed of light is constant to every inertial observer. Rather
it is a theoretical consequence of Maxwell's equations for
electromagnetism."

You are mixing up light's propagational motion through space
with Einstein's definition of clock synchronization. Maxwell's
equations say nothing at all about light's one-way, two-clock
speed, and Einstein's definition says nothing at all about
Maxwell's equations. (If you believe otherwise, then tell us
how Maxwell synchronized clocks.)

Since you have claimed that light's one-way, two-clock speed
is the same for all inertial observers, you have the burden
to prove this, so let's see if you can.

You have also claimed the "truth" of SR, so let's see if you
can prove that its clocks are correctly synchronized.

And I could not help but notice that you had no counter to my
very simple and very direct experimental proof that Einstein's
clocks are not correctly synchronized, and that light's one-way
speed is not invariant.

I have presented a simple and direct experiment, and you have
(to put it bluntly, so pardon me!) presented absolutely nothing.
To repeat, the burden of proof is on the claimer. (This means
that I did not even need my experiment!)

Let's please stop runnning around in circles, and let's see
some 'meat on the table' for Einstein's case! Where's the
beef??


All times are GMT -5. The time now is 09:27 AM.

Powered by vBulletin Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
© 2014 Physics Forums