The notion of time is not even properly defined. What clock must be tested, the coordinate time clock or the proper time clock? How is the information transported? How is our living space deformed. At long distances the Hubble effect is taking its role. What is the reason and what does it to clocks?

What are photons? How long takes the passage/emission/absorption of a photon. How can its frequency/energy be measured?

In short, Yes.

The idea of confirming some thing that is just a theory is just to check its accuracy over time and distance.  The theory is just a better approximation of gravity and therefore all you would be doing is saying it is closer or further away from what you thought its limits were.

However theories do not imply that you understand the force of gravity.  It only implies you have a better approximation of the force over time and distance.  Gravity is still a force that has not been explained.

George Van Hoesen

Physics is mainly a descriptive science. It hardly explains anything. It can definitely not explain the dynamic coherence of the objects that appear in universe.

I would like to involve some people who agrees in order to:

a) make a joint project of the experimental setup for the tests

b) make a joint requst to fund at lest the first experiment which is easier to perform.

Hans,

I did not talk about time, but atomic clock retardation. Just wanted to be very low profile: read gauges of one device and compare it with another one numerically. Check some results from some specific device which Physics assumes to be suitable to measure time.

Dear Alex,

thanks for your link which reports:

"In a bizarre experiment using the most accurate atomic clocks ever invented, researchers showed that clocks run faster if they are raised by just 12 inches."

it is the gravitational time dilation which has been experimentally verified several times. Your link testifies that there are very accurate atomic clocks used to perform clock retardation experiments.

The first experiment I propose is Kinematic not gravitational. It regards for example a very high speed automobile which accelerates on Death valley from 0 to 500 mph and has one clock in the trunk and on in the hood.

Regarding the second experiment the tidal effects are negligible so won't be revealed by present instrumentation. My question is to check if the clock retardation is only due to tidal effects or it is the same as the gravitational time dilation as in the link you provided.

Very nice question.

Thanks Otto,

the equally accelerated clocks is actually something crucial:

a) would discriminate on the accuracy of the equivalence principle in the form initially proposed by Einstein (1907 and 1911) the link  gravitational field / accelerated frames . It will be tested if there is any match between gravitational time retardation and whatever results from the equally accelerated frames.

b) would set a definitive word on the relativity of simultaneity.

Alex,

A significative value can be obtained in a setup with  100 meter distance between the clocks and acceleration of 2 m/s for each second for 180 seconds:

 the mag lev train if you check

 https://www.youtube.com/watch?v=x-NXyPPaf3Q  

goes about like that reaching 430 km/h in 3mins =180s

Delta T/T = gH/c2  ; Delta T = T *gH/c2

the term gH/c2 should be 200/10^16 =2*10^-13  

being T=180s

so that    Delta T = 180s  *2*10^-13 = 360*10^-13 = 3,6*10^-11s

which is a detectable in the gauges of the clocks if you put in sync two oscillators having a precision higher than 10^-14 which is something quite affordable nowadays.

If a value results lower than 10^-12s  from the difference between the gauges of the clocks, then the prediction of SR on accelerated delayed clocks does not hold.

i`ve found out that, if mass, space and time must be adjusted by using 1-v²/c², then v² itself has to be adjusted and re-adjusted and re-re-adjusted... the Problem is not that the Terms (series) become negligible, but they are divergent! This is my own idea and it it brand-new!! 

Alex,

the prediction of SR is that in accelerated frames clocks manifest retardation proportional to the acceleration itself and their reciprocal distance along the direction of motion. SR is based on the Poincare' group which supports accelerations. Lorentz group does not.

Is it okay if I predict that Einstein will be exactly confirmed?

The third implication he foresaw, gravitation light slowdown, will be harder to confirm, but it will be the most interesting one to check. Because one may then be able to confirm also the new prediction of of a vertical (longitudinal in the experiment) size increase.

Dear Otto,

I did not propose this to confirm and disconfirm Einstein. I proposed this to clarify on a subject which deserves a test, it is very important but was never tested, so welcome to everybody.

Alex,

it depends also on the distances between the clocks H, not connected by light beams at all. They both have to communicate with a central unit in order to be set in sync(optic fiber). They travel like passengers, they are put in sync initially where they are located, left in the same position in the train since the beginning of the experiment.

I am interested in reading the final clock retardation, the sum of several femtoseconds which accumulate and that is for sure possible to detect.

For example if the trip lasts 180 seconds and each second in the tail clock provides one femtosecond delay, at the end I will have at least 180 femto which is possible to detect.

Dear Prof. Behnam.

thanks a lot. Yes your link regards a further and more accurate confirmation of results which had already a confirmation. My question is specifically about two experiments on atomic clocks which I don't think have been performed and I would like your opinion on its feasibility.

To your first experiment: is there any consistent model which predicts anything else? The reason to make an experiment is mainly to decide between alternative possibilities. From SRT follows what I have said. Galilean physics, of course, leads to a quite different prediction. But Galilean physics has been refuted by countless experiments. Ether theories have been parametrized, and several experiments set strict limits on the possibility of deviations from SRT.

@ Ales: ``classical SR is not valid for accelerated frames.'' That is correct. But there are no problems with accelerated motion. You calculate with proper times, as shown initially by Einstein and later by Minkowski. The world's highest accelerations, in particle accelerators, are well described by SRT.

What indeed does not work well in SRT, is to use an accelerated reference frame. But reality can always be described in any reference frame and does not change. So you can describe anything you like in an  appropriate inertial reference frame. Describing accelerated motion, which is what Stefano wants to do, is no problem whatever.

Francois,

I agree with you that in order to be 120% ok for presenting an experiment an alternative model should be proposed which could justify the alternative results expected.

The aim is not to confirm a new theory which does not exist but to see if the present holds or not in a crucial test. Should be more than enough to test the present model and see to what extent it holds.

It will be the exact  same philosophy adopted on the experiment reported in the thread by Benham, with the difference that this should be the first to test such feature in accelerated motion.

Why identical objects should show a different behavior only for being placed at rest at a certain distance (different numbers on their gauges) and having had  the same history of motion (after being set in Sync)?

Einstein on the base of the Doppler Shift present in accelerated motion affirmed that the two clocks should present an "actual delay". The perceived frequency shift in accelerated frames  is quite an experimental evidence, is it due to the frequency shift of the oscillators (clock retardation) or the frequency shift of the radiation? It is not the same thing. This experiment will discriminate also on the actual nature of the perceived frequency shift of radiation exchanged between bodies  with same proper acceleration.

For experiment 1) the two clocks must --- by translation invariance --- behave identically, when viewed from the unaccelerated frame. However, as they gain speed their common perception of simultaneity will change.  After the acceleration phase this will amount to a time shift of dt = vx/c2, regardless of acceleration history (as long as translation invariance in the stationary frame is maintained).

Compared with the (around) 500 THz frequency of lasers in the visible range, i.e to = (1/5) 10-14  s, we must have vx = 180 m2/s for dt = to.

A possible experiment would be:

1. Rent a train car, 20 m long(?),  with a (modest) locomotive, and two sufficiently stable lasers in the visible range.

2. Set up the lasers in the back and front of the car, and mirrors producing a stable interference pattern between the two in the middle of the car. As long as the car is at rest this interference pattern should remain stable if the lasers are stable enough, but when the car is set in motion the interference fringes should start to move, due to the changed notion of simultaneity --- and corresponding detuning --- in the instantaneous rest frame of the car. With a full period each time the velocity is increased by the modest amount of 9 m/s.

Alas, this is probably a close replica of the 2010 experiment by Chou et. al. already mentioned. 

Note added: There is of course a question of the mechanical stability of the train car: Can the distance between the two lasers be kept constant to sub-micron accuracy?

“…Both theories of relativity are profound concepts…”

- both theories are “profound concepts” first of all since in the theories it is postulated that there is no absolute Matters spacetime and so interactions between material objects are determined essentially by “spacetime geometry”, when this geometry, which, besides, is pseudoEuclidian/pseudoRiemannian and has imaginary (spatial or temporal) dimensions, is transformed by material objects and reference frames [so implicitly – by some “observers”].

All these SR/GR postulates seem as at least questionable, for example seems nobody till now observed imaginary space or time, and have no any experimental confirmations.

Thus indeed direct testing of the theories is experimental observation of the fact that Matter’s spacetime is absolute [5]4D Euclidian manifold where every material object moves with concrete absolute spatial speed. And such experiment is possible  now – see https://www.researchgate.net/publication/259463954_Measurement_of_the_absolute_speed_is_possible

Cheers

Article Measurement of the absolute speed is possible?

Test 1 is probably impractical as stated, the time dilation would be very small but the resulting velocity difference would create a large Doppler shift if measured from an inertial platform. Measured on the accelerating platform, the result is just a repeat of Pound-Rebka via the Equivalence Principle so I doubt you'd get anyone interested. The EP itself is best tested in other ways, there's a lecture linked that lists the history of this, see slide 10 for a list. I don't think MiniSTEP has been funded so you would need to raise a proposal that was cheaper and could provide better accuracy than that.

Test 2 is easily checked using existing communications satellites. In fact it is a known problem, satellite to ground comms is easy but satellite to satellite is harder because they don't stay in sync for exactly this reason. In other words, it is a test that has already been performed and GR gives an exact prediction.

http://www.tat.physik.uni-tuebingen.de/~kokkotas/Teaching/Experimental_Gravity_files/EP_06_05.pdf

Aleš Kralj: One should be clear: classical SR is not valid for accelerated frames. So, no need to invalidate it. :-)

F. Leyvraz: @ Ales: ``classical SR is not valid for accelerated frames.'' That is correct.

Those statements are in fact not correct. SR does deal with accelerated frames (with a bit more effort) and Rindler Coordinates (see the link) are specifically defined in such a frame. That is the key point about the Equivalence Principle, it turned uniform gravity into an SR problem in an accelerated frame.

• "The Rindler coordinate system or frame describes a uniformly accelerating frame of reference in Minkowski space. In special relativity, a uniformly accelerating particle undergoes hyperbolic motion. For each such particle, a Rindler frame can be chosen in which it is at rest."

Aleš Kralj: This is why SR it is called "special".

It is called that because it copes with accleration which is only equivalent to uniform gravity. It doesn't cope with tidal effects which are not the same everywhere in space.

https://en.wikipedia.org/wiki/Rindler_coordinates

George,

"Test 1 is probably impractical as stated, the time dilation would be very small but the resulting velocity difference would create a large Doppler shift if measured from an inertial platform"

which Doppler shift are you talking about? What I propose is just reading gauges of atomic clocks at two instants.

1) the only signals exchanged will be

     a)  SYNC signal when the system is at rest on the surface of the earth, it will be given by some device RECORDER set at the same distance from both clocks, in the middle of the train.

      b) when the acceleration is over there will be a read gauges  by the RECORDER,  the train is an IRF at that point, so no signal will be frequency shifted at this point. The recorder is not on the ground but moving in the IRF as well.

George,

"Test 2 is easily checked using existing communications satellites. In fact it is a known problem, satellite to ground comms is easy but satellite to satellite is harder because they don't stay in sync for exactly this reason."

According to GR two free falling clocks following the same trajectory should stay in sync because they belong to the same IRF . Orbiting clocks on different orbits cannot stay in Sync because they belong to different IRFs.

George,

"Those statements are in fact not correct. SR does deal with accelerated frames (with a bit more effort) and Rindler Coordinates (see the link) are specifically defined in such a frame. That is the key point about the Equivalence Principle, it turned uniform gravity into an SR problem in an accelerated frame."

yes infact Rindler and Eric Lord before him (as pointed out by Kare Olaussen) dealt with curvilinear coordinates suitable for accelerated frames. The result in accordance with SR is that the leading clock of an accelerated configuration should go faster, a sort of a confirmation of the setup of the Gedanken experiment proposed by Einstein in his paper of 1911 when he illustrated the application of the EP.

Sorry Stefano, for test 1 I was thinking you wanted to do the measurement from a stationary reference.

You could use the Japanese "Bullet Train". The maximum speed is quoted as 320km/h and carriage length is 25m so if sync'd when stationary, clocks at each end of the carriage would be 3.2*10-13s or 320fs out of sync at full speed (it's 1fs per km/h, a very expensive speedometer).

As Kare said, the amount is vx/c2 so doesn't accumulate and is independent of the acceleration history. It will revert to zero when the train stops. Essentially you would be testing the relativity of simultaneity directly.

You would need to suspend the clocks and fix their separation mechanically, a long rigid pole for example, to prevent stress in the carriage structure "stretching" the metal.

If the satellites don't do what you want, I don't understand what you are proposing for the second test. Clocks in free fall if aligned vertically and separated by distance H will show differences due to gravity and their varying speed. Separated horizontally, the tidal effect will cause them to move together. What is your suggestion exactly?

George,

Sorry Stefano, for test 1 I was thinking you wanted to do the measurement from a stationary reference.

Isn't a moving train at constant speed a stationary RF?

You could use the Japanese "Bullet Train". The maximum speed is quoted as 320km/h and carriage length is 25m so if sync'd when stationary, clocks at each end of the carriage would be 3.2*10-13s or 320fs out of sync at full speed (it's 1fs per km/h, a very expensive speedometer).

better Maglev 600 km/h and set at 100 meters why limiting to one carriage. One could lay fiber optics from the head to the tail. (I do not need a speedometer...)

As Kare said, the amount is vx/c2 so doesn't accumulate and is independent of the acceleration history. It will revert to zero when the train stops. Essentially you would be testing the relativity of simultaneity directly.

yes but I don't want the train to stop (return to its initial status) and read the results, I want to read the gauges when the train is running at full speed but not accelerating.

You would need to suspend the clocks and fix their separation mechanically, a long rigid pole for example, to prevent stress in the carriage structure "stretching" the metal.

The acceleration is not so high it is 1m/s per second (3km/h per second) I don't think there is a problem for the stress of the clocks.

Clocks in free fall if aligned vertically and separated by distance H will show differences due to gravity and their varying speed.

Are you sure, difference calculated according to which (math) relation?

Dear Stefano Quattrini,

1.The time dependence on the velocity of the objects is main result of the Special and General Relativity and it agrees well with the observations.

2.The motions of the cosmic objects  are noninertial. The Special Relativity deals with the inertial motion. General Relativity replaced the rigorous description of the noninertial motion by the introduction of the Space curvature to take account the hypothetic gravity field (with several matching parameters).   In the most simple case, when almost all parameters are equal  to zero,   the Newton law  of the Universal Gravity was obtained. For the more complicated cases of stars and Galaxies, the obvious results were not obtained.     

3.Both the Kepler law and the equations of the General Relativity can’t explain the observed rotation curves of stars. Instead of that, the presence of hypothetic dark matter and dark energy were postulated.

           Best wishes,   Tamara Kudykina.

Dear Tamara,

yes infact there is little to test about speed but something to know about its derivative which is detectable by accelerometers.

The Special Relativity deals also with non inertial motion. Lorentz Transformations deals with IRF , the group is included in SR   with the Minkowsky spacetime which implements the Poincare' group dealing also with accelerated motion. The LHC accelerator is based on the predictions of SR which are confirmed daily.

SQ: Isn't a moving train at constant speed a stationary RF?

Yes, but I thought you wanted to do the measurement from the station platform, not on the train. No matter, we've cleared that up.

SQ: better Maglev 600 km/h

True though the fastest commercial system seems to be the Shanghai Transrapid at 430km/h, you want to minimise the cost I assume.

SQ: and set at 100 meters why limiting to one carriage. One could lay fiber optics from the head to the tail.

If you can eliminate variations due to thermal variations in the fibres etc., you might get away with that but for sub-pico second measurements, that's not going to be easy. I was assuming you would use a direct laser beam (e.g. laser diode) to convey the signal, you probably need some sort of interferometer setup to make the measurement.

SQ: yes but I don't want the train to stop (return to its initial status) and read the results, I want to read the gauges when the train is running at full speed but not accelerating.

OK, you used the word "accumulate" earlier and I worried you might think the difference grew with the duration of the experiment.

GD: You would need to suspend the clocks and fix their separation mechanically, a long rigid pole for example, to prevent stress in the carriage structure "stretching" the metal.

SQ: The acceleration is not so high it is 1m/s per second (3km/h per second) I don't think there is a problem for the stress of the clocks.

The clocks would be more affected by vibration, Maglev would be a good idea from that point of view, but I was talking about the distance between the clocks varying due to drag on the carriage which would depend on speed, not acceleration. If you can use fibres, that would get round that problem (but introduce others).

GD: Clocks in free fall if aligned vertically and separated by distance H will show differences due to gravity and their varying speed.

SQ: Are you sure, difference according to which (math) relation?

a(r) = GM/r2 versus a(r+H) = GM/(r+H)2 hence their speeds will differ. The Doppler difference will be much larger than time dilation.

George,

GD: Clocks in free fall if aligned vertically and separated by distance H will show differences due to gravity and their varying speed.

SQ: Are you sure, difference according to which (math) relation?

a(r) = GM/r2 versus a(r+H) = GM/(r+H)2 hence their speeds will differ. The Doppler difference will be much larger than time dilation.

Yes you are referring to tidal effects though.

It seems that the redshift predicted between them should be virtually 0.

Dear Arno,

the Sagnac correction is used in that case. Some argue that it is out of the SR.

GD: The Doppler difference will be much larger than time dilation.

SQ: Yes you are referring to tidal effects though.

Yes.

SQ: It seems that the redshift predicted between them should be virtually 0.

Try calculating the two, my statement is just an estimate. If you release the two from reat at the same time, how does each effect then grow as they fall?

AG: Yes. Let a satellite fly without so-called srt-correctIon and see what happens then.

There's a story that some non-scientists were sceptical (though that may be apocryphal) but for whatever reason, the first test flights of GPS satellites had a switch to enable the correction. When they were launched it was off and the clocks ticked faster than the ground clocks by the predicted amount. Then they switched on the correction and they then matched the ground clocks.

SQ: the Sagnac correction is used in that case. Some argue that it is out of the SR.

GR gives a single correction figure of 38us/day. Some people make a comparison with a hypothetical clock hovering at the same altitude but not orbiting. Such a clock would tick faster by 45us/day due to gravitational time dilation at that altitude, and the GPS clock would move past it at a speed that caused slowing by 7us/day as a purely SR effect.

The Sagnac Effect is valid for signals sent round the planet between satellites where it causes difficulties with synchronisation but it doesn't really make much sense applying it to a single satellite.

Dear George,

The clocks would be more affected by vibration, Maglev would be a good idea from that point of view, but I was talking about the distance between the clocks varying due to drag on the carriage which would depend on speed, not acceleration. If you can use fibres, that would get round that problem (but introduce others).

good point.

Dear Kåre ,

thanks a lot for partecipating.

"the two clocks must --- by translation invariance --- behave identically, when viewed from the unaccelerated frame. However, as they gain speed their common perception of simultaneity will change.  After the acceleration phase this will amount to a time shift of dt = vx/c2,"

I think it is better to perform the test all inside the train not in reference to the observer at rest, as already mentioned with George.

I'm interested in the objective values registered by the two clocks, reporting them in the same middle point in the IRF of the train when it proceeds at constant speed.

Correct me if I'm wrong but the clock retardation of the tail clock should be proportional to the proper time of the head Thead

Delta(Ttail)=Thead(1+gH/c2)  (1)

I do not have to consider  vx/c2  since my measurements are done in the train

and (1) is involved in  the equivalence with a gravitational configuration with gravitational potential of gH.

Stefano> I do not have to consider vx/c2.

But that is what you have! You don't want the "1+" on the right hand side of eq. (1), Thead  g = v (the final velocity), and my x is your H. As I said, the whole effect is just a net rotation of the line of simultaneity, adiabatically slow for realistic accelerations and train sizes. Although you may interpret it as a (time dependent) gravitational redshift. That is just two ways to understand the same physics. (According to Feynman, we need to come up with a few more ways if it is a really profound phenomenon.)

My calculations, and suggested experiment, do refer to the situation on the train, taking for granted that the notion of simultaneity is that of the local restframe on the train. I kind of doubt that anyone will bother to do it. Perhaps if the effect could be used to come up with a patent for an ingenious speedometer, or similar device of practical use :-D.

A similar setup testing much of the same physics (with the addition of a little topological twist) is a rotating device, where the acceleration can be maintained for much longer time.  Such setups are already been used in practical devices.

And yes, they are based on the Sagnac effect. There may be some who argue that this is out of SR. I disagree. There may be some who argue that the earth is flat also.

Previously we said:

GD: You could use the Japanese "Bullet Train". The maximum speed is quoted as 320km/h and carriage length is 25m so if sync'd when stationary, clocks at each end of the carriage would be 3.2*10-13s or 320fs out of sync at full speed (it's 1fs per km/h, a very expensive speedometer). As Kare said, the amount is vx/c2 so doesn't accumulate and is independent of the acceleration history. It will revert to zero when the train stops. Essentially you would be testing the relativity of simultaneity directly.

SQ: yes but I don't want the train to stop (return to its initial status) and read the results, I want to read the gauges when the train is running at full speed but not accelerating. I'm interested in the objective values registered by the two clocks, reporting them in the same middle point in the IRF of the train when it proceeds at constant speed.

Now you say:

SQ: Correct me if I'm wrong but the clock retardation of the tail clock should be proportional to the proper time of the head Thead

SQ: Delta(Ttail)=Thead(1+gH/c2)  (1)

When the train is "proceeding at constant speed", the acceleration g is zero ;-)

SQ: I do not have to consider  vx/c2  since my measurements are done in the train.

Done on the train as you describe, if the clocks are synchronised when at rest, the time difference (measured in the centre or at either end, it doesn't matter where) when travelling at speed v with a separation of x will be:

Thead - Ttail = vx/c2

which is 1.72ps for clocks separated by 100m travelling at 430km/h. That's about 0.5mm at the speed of light meaning you need to maintain a length stability of about 0.3mm in fibre. That seems quite reasonable.

George,

When the train is "proceeding at constant speed", the acceleration g is zero ;-)

yes but in the gauges of the clock the delay (different progressive number count), if any, should remain present if exists (getting accumulated during the acceleration phase provided that  clocks did have different paces during the acceleration phase). If  acceleration ceases the clocks will again indisputably posses the same rate since they belong to the same IRF so in that condition, but they should be out of sync.

The crucial point is to avoid to stop the counting   when the train is again at rest, then for sure I would not be able to find differences.

a) they can be stopped by an optic singal through the fiber and then both clocks brought together for the comparison.

b) could be compared directly reading the values from the mid point of the train.

I think a) is better.

SQ: If acceleration ceases the clocks will again indisputably posses the same rate since they belong to the same IRF so in that condition, but they should be out of sync.

Yes, that's what Kåre and I are saying, the accumulated difference, the amount they are out of sync, is given by vx/c2.

SQ: b) could be compared directly reading the values from the mid point of the train.

Right, each clock would just produce a constant frequency signal, usually 10MHz but a higher frequency is better (probably essential). What you do is measure the phase difference when the train is at rest and again when it is moving and you average each measurement over a long observation time to reduce the Gaussian noise. That is what the fibres from the clocks carry to a common point where the phase difference is measured. If for example you could get a direct feed from the front end of the clock at 9.19263177GHz, that's a period of ~1.0878*10-10s or 108.78ps, so at 430km/h measuring a time difference of 1.72ps corresponds to a phase shift of about 5.69 degrees relative to the value at rest.

Stefano, I just wondered if you had realised the connection between the measurement we are discussing and gravitational time dilation. The measurement is best done at constant speed so that you can integrate to eliminate noise. However you could do it at a range of speeds, say every 10km/h. Plot a graph of phase shift φ versus speed v and you'll get a straight line.

If you then consider what you would get for constant acceleration dv/dt=a, the rate of change phase difference would be dφ/dt, but rate of change of phase is angular frequency, so dφ/dt represents an observed frequency difference between the clocks even though, at each measurement, the frequencies of the clocks are identical.

Put in a few constants (separation x and signal frequency f) and you therefore have a direct measure of the frequency difference due to acceleration in an SR scenario which will be the same as the gravitational frequency shift. That would be a direct demonstration of the Equivalence Principle :-)

“…direct measure of the frequency difference due to acceleration in an SR scenario which will be the same as the gravitational frequency shift. That would be a direct demonstration of the Equivalence Principle …”

- besides at least two points: (i) – with a great probability the real gravitational frequency shift [and “gravitational time dilation”] is two times lesser then the GR claims, and (ii) – the “gravitational frequency shift”/ dilation exist till there is a gravitation, if the gravitation is turned out, the shift/”dilation” disappear,

when after an acceleration the shift/”dilation”

[of course not “time dilation”, but a changing of  internal processes’ rates in concrete accelerated material object. That is true for “gravitational time dilation” also – there is no any “spacetime curvature” and corresponding “time dilation”, again only internal processes in concrete object are slowed down by the gravity]

- remains be constant.  Moreover, the processes’ tick rate depends on – what absolute speed the object has after an acceleration, so if the “gravitational time dilation” is always negative, an “absolute” deceleration results in a “time acceleration” …

Cheers

SS: ... the real gravitational frequency shift [and “gravitational time dilation”] is two times lesser then the GR claims ...

The GR formula has been confirmed many times. The linked image shows the amount versus altitude and indicates the approximate ranges for "Low Earth", GPS and geosynchronous orbits. The amount of the net effect has been confirmed in all of those. If your alternative ideas produce the same "net orbital time gain" curve by some other compensating factor then all is well, if not it is wrong.

At the end of the day, observation determines veracity, and GR passes that test.

I just gave a new prediction made by Einstein's early gravitation theory ( https://www.researchgate.net/post/CHAOS_THEORY_SAVES_PLANET ):

Does anyone see an opportunity for a direct experimental check?

Otto, you need to edit your post to remove the brackets and colon or it breaks the link.

OER: Does anyone see an opportunity for a direct experimental check?

You should ask that in your own thread, this one is about Stefano's proposals.

I appreciate your help, George.

GD: “…the GR formula has been confirmed many times…"Low Earth", GPS and geosynchronous orbits…”

- again in the experiments quoted, as that was in the Pound-Rebka-Snider experiments as well, a sum of two effects is measured – a [very probable] change of photons’ frequency at motion between points with different gravitational potentials and difference of internal processes’ rate in material objects that are in these points because of the gravity impact. The sense of the  “Burj Khalifa experiment"  [SS posts above] just is in that in this experiment only one effect, i.e. the gravity impact on electronic clocks

 [the gravity impact on pendulum clocks will be opposite – the upper clock ticks slower then the clock on first floor, instead of the GR’s “gravitational time acceleration” here is a “gravitational time dilation”]

is measured; and further, by using  the results of the known experiments, that allows to answer also on the question –  change photons their energy in the gravity field or not.

And answers on these two fundamental questions cost no more then $1000 000, thousand times cheaper then, say,  the LIGO experiments, which by no means can “confirm the GR” – the gravity force is the 4-th fundamental Nature force that seems as rather similar to the EM force; and so it seems quite natural that at some changing of gravitational charges [g-masses] configurations some gravitational waves can be created. To estimate any difference between the GR and this “a la EM” effect is necessary to know – what initial masses values and initial masses configuration were involved in the process, what is practically impossible…

Cheers

Dear all,

Something interesting emerged on the thread dedicated to the paper of Eric Lord about the SR theory.

I modified my questiona above and I would like to see what is your opinion about the following.

Postulate of SR: the speed of ligth is constant in vacuo equal c in all inertial reference frames.

This has been experimentally verified though for the two way speed of light. Some may raise doubts about the opportunity to accept that the one way speed of light is implied necessarily by the two way measurement but this is something which I am not interested in because there is something which I think is more important to reason aobut.

There is not an explicit derivation in accelerated frames or there has been no experiment regarding what occurs for NON INERTIAL REFERENCE FRAMES, in accelerated motion, it seems that light has always been considered to follow the same rule irrespective of the RF. 

This is a degree of arbitrariness which has to be experimentally tested and it is strictly connected with the relativity of the simultaneity.

SQ: There is not an explicit derivation in accelerated frames or there has been no experiment regarding what occurs for NON INERTIAL REFERENCE FRAMES, in accelerated motion, it seems that light has always been considered to follow the same rule irrespective of the RF.

That is not the case. A trivial example is the speed of light between two stars in the Andromeda galaxy in the non-inertial frame co-rotating with the Earth. Andromeda is just under 3 million light years from us so in that frame moves in a circle of that radius every 24 hours. That's a speed of ~20 million light years per 24 hours. Linear acceleration produces similar effects but it is most obviously seen in rotation.

George,

I will be more accurate: same linear accelerated motion in absence of gravitation.

Rotational acceleration would present a paradox connected with clock retardation which is not acceptable, unless you admit that speed of light is not constant. The rotational configuration proposed in the paper below has nothing to do with Sagnac since there is not any exchange of radiation between the clocks.

Article Time dilation in accelerated frames and gravitational potentials

About the feasibility of experiment 1):

As calculated by George, one must be able to measure a time difference dt of order a few picoseconds or smaller out of the total duration t of the experiment. With 1 g acceleration t would need to be about 15 s, probably more for a whole train. This means a relative accuracy

   dt/t of order 10-14. 

This seems modest compared accuracies of optical clocks. The 2010 Science article by Chou et. al. mention a frequency accuracy df/f of order 10-17, but they are doing long duration measurements in the frequency domain. I don't think any time difference dt can be extracted from their method.

I mis-remembered their experiment when I said that my proposal was similar to theirs, mistaking it with a gravitational redshift experiment by Muller, Peters and Chu (Nature 463, 926, 2010) using interference of matter waves. They achieve an accuracy of  7 x 10-9 (i.e., not quite good enough), and their method anyway does not look applicable for this case.

The method I proposed apparently belongs to a technique called optical heterodyning, https://en.wikipedia.org/wiki/Optical_heterodyne_detection. From some brief readings it seems that its greatest obstacle is to achieve a sufficiently long time coherence of the lasers (inversely proportional to the laser linewidth). I.e., one would need two lasers with sub-Hz linewidths.

Hence it seems that an atomic clock, like the https://en.wikipedia.org/wiki/Deep_Space_Atomic_Clock, is a better option. It is said to achieve an accuracy of 1 ns in 10 days, which translates to a relative accuracy of 10-15, i.e sufficient. But it is unclear from the wikipedia article if it can resolve picosecond time differences directly. If it can, it could be possible to convince the Jet Propulsion Laboratory to run an experiment. It should be considered sufficiently different from the existing GPS observations to warrant a try, and it might be used for educational and public outreach purposes.

Who wouldn't like to be informed that the clocks in the front and end of your Maglev train have gone out of sync, because of the relativity of simultaneity?

Stefano> Postulate of SR: the speed of ligth is constant in vacuo equal c in all inertial reference frames.

I have been teaching that the speed of light c is an invariant for many years, but that is obviously not quite correct, since speed is a dimensionful quantity. I think the correct statement (which one may want to promote to a postulate) is:

It is possible to choose time and space coordinates locally such that the speed of light is the same in all directions at all points.

It is well known, to those who know it well, that there can be topological obstructions to the global construction of such coordinates.

SQ: ... in accelerated motion, it seems that light has always been considered to follow the same rule irrespective of the RF.

GD: That is not the case.

SQ: I will be more accurate: same linear accelerated motion in absence of gravitation.

Rotational acceleration would present a paradox connected with clock retardation which is not acceptable, unless you admit that speed of light is not constant.

I guess I should have been clearer too, what I meant was that it was not true that the speed of light is assumed to be the same in accelerated frames, it always varies with the details depending on the form of acceleration. For linear acceleration, typically there is a term of the form ax/c involved where a is the proper acceleration of the observer (frame) and x is the distance from the origin to the pulse of light under consideration measured in the direction of the acceleration vector.

Incidentally, the best confirmation of Lorentz Symmetry I've seen is the linked article.

http://www.readcube.com/articles/10.1038/ncomms9174

 Dear George,

"For linear acceleration, typically there is a term of the form ax/c involved where a is the proper acceleration of the observer (frame) and x is the distance from the origin to the pulse of light under consideration measured in the direction of the acceleration vector."

is it the Poincarè group you are referring to? 

The postulate at the base of the Lorentz Symmetry is the same as the Lorentz Transformations which rely on inertial reference frames, how can you extend the concept to non-inertiality?

Sorry don't understand.

Dear Kare,

I was replying to you but I lost everything. I will reply you tomorrow. Thanks a lot for your support meanwhile.

SQ: how can you extend the concept to non-inertiality?

The LTs tell you how coordinates change under a "Lorentz boost", a simple change of speed of the coordinate system. You can use that to calculate the effect of a change of speed by increment dv and apply it repeatedly every time increment dt. The limit as dt->0 is the way the coordinates change as seen from a frame which is uniformly accelerating at dv/dt. This approach is simply calculus applied to the LTs.

For another approach, look at Rindler Coordinates (see the link and its references).

https://en.wikipedia.org/wiki/Rindler_coordinates

The postulat of Coonstance of the speed of Light : "

the speed of light is a constant of value c in vacuum in ALL INERTIAL REFERERENCE FRAMES veryfied in the mesasurement of the 2 ways speed of light in vacuo

Would it be advisable to ask for additional experiments to confirm some predicitons of Special and General relativity? - ResearchGate. Available from: https://www.researchgate.net/post/Would_it_be_advisable_to_ask_for_additional_experiments_to_confirm_some_predicitons_of_Special_and_General_relativity [accessed Aug 29, 2016]."    

is conformed for any small regions enclosing the given 4-dimensional point in a curved space-time. See, any textbook on General Relativity.

@Valentin,

I agree with you that it would be better to check with some clocks on the ground along the path but this would overcomplicate the experiment, risking to make it unaffordable.

Dear Alex:

You are aware that this accepted constancy of c in the vacuum is not a global but only a local constancy?

Valentin, your suggestion results in comparing a moving clock to stationary clocks which tests speed-dependent time dilation. Stefano's proposal was to test gravitational time dilation or time dilation due to acceleration of two clocks moving at a fixed separation (in the clock's frame). Your method would be testing a different consequence of the theory.

VD: Well, I didn't mean to check the times "on the fly" (i.e. to compare the moving clocks to the stationary clocks while the train is moving). What I meant is: only capture the indications of the moving clocks, by using the stationary detectors,

I don't follow that, do you mean you want to stop the train for each capture, then drive it farther along the line before stopping to capture the next reading? If not, the train must be moving as it passes each capture point which makes it difficult to define the exact location of the train clock (to a fraction of a mm) at the moment of the capture.

stationary clocks do not exist.

Assume you want to transmit state information about the current tick of your clock, to a device receiving only a few "ticks" as the clock fly by, with a tick rate of1012-1015 ticks per second, as required for Stefanos experiment. How would one do it, and what would the required data rate be?

Transmission of single ticks would not provide information about the global count. But maybe one could do with a single byte: This would provide time information within a given 256 tick epoch (the epoch might be inferred/guessed by other means). This leads to a data rate of 1 Tbyte/sec --- 1 Pbyte/sec. In comparison, the running LHC experiments is said to store about 1 Gbyte/sec, about the limit of what their systems can handle (this is information from a few years ago).

Hence, I am skeptical that any direct comparison of two absolute times will be feasible with existing technology, to the required time resolution. Interference experiments makes it possible to monitor drifts in relative time to sufficiently high resolution, but requires long time coherence of the clocks.

relativity theory has already introduced adjusting functions like: square root of 1-v²/c² (let`s name it Sr(v,c)) where c is variable too. an absolute velocity can`t be postulated (relative to what?), and on the other hand speed must be adjusted by multiplying with Sr(v,c). thus v=(Sr(v,c))²*s/t and acceleration=(Sr(v,c))³*s/t². ergo: 1. acceleration should not be absolute and 2. the adjusting term includes v². there is v once again to be adjusted by operating with Sr, where v²appears anew, again and again... iteration! furthermore, when matter is in motion, mass, as well as time and space,velocity and acceleration must also be adjusted by operating with (iterated) adjusted velocity. the problem is not that the Sr terms are negligible, but the Sr(Sr(Sr(...))) series are divergent! acceleration (as speed variation in time) is not absolute, speed (as space variation in time) is relative. acceleration and gravitation are not equivalent.

space-time is not homogeneous especially in time. the system (i.e. universe) is not homogeneous due to singularities, bangs & crunches. operating in discontinuities is avoided by using test functions and set functions. physical entities shall be described by using functionals of complex function argument (i.e. the distributions theory).

see my work: unification of relativity and quantum theories

1.I thought it was one of Poincare`s ideas, used by Einstein: see. Lorentz transformations.

2.in my opinion time isn`t linear.

With the theory of relativity a new miracle came into the world: at very high speeds systems are ageing slower; even though c in the formula e=mc² needn’t be seen as velocity, because c² > c(axiom violation!!) and c² should be measured in km² per square seconds[1] (i.e. two-dimensional time(!); while every function should depend on time- as time is motion whereat it obviously depends on two variables: acceleration and velocity (in other words: variation of space and variation of variation of space (in time- referring to itself)).

[1] if planck-time (10^-43) hypothetical is the smallest possible object and time is defined to be one-dimensional, then 1/s² (10^-86) surely should be less than 10^-43! (except that: Planck-etalon is considered unit, but in this case the probability theory must be revised.)

Valentin,

“…in general everything which has inertia "moves" (in reference to something else), so in that sense the clocks cannot be deemed "stationary" in the sense of absolute rest…”

- that isn’t so. Every material object moves with the speed of light in the absolute Matter’s [5]4D Euclidian spacetime simultaneously by two ways: (i)  - along 5-th, i.e. “true time”  dimension and (ii) - in different directions in the 4D Euclidian sub-spacetime. Thus there can be at least 3 “absolute rests” – no [5]4D motion – the case that doesn’t exist in Matter, since everything material moves; the rest relating to/in the 3D space and the rest relating to/in the 1D “coordinate time” [further “time”, "temporal", etc.] dimension [4-th coordinate/axis]

There are well known particles that are at rest relating to the coordinate time dimension – that are photons, which move in the absolute 3D space only with [absolute] speed of light. The “having rest mass” T- particles are created at impacts that have non-zero temporal component and so always move along the temporal axis; this motion is the increasing of the coordinate time interval that accompanies changes of such particles/bodies…internal states. Thus the T –particles can be at absolute 3D spatial rest and in this case they move in time with maximal speed, i.e. with the speed of light.

Clocks are some objects that show how their internal state changes; as usual they are made from T-particles and so at 3D spatial rest tick with maximal rate. If a clock moves in 3D space also with a speed V, its speed along the temporal axis decreases, and is equal, as that Pythagoras established to c(1-V2/c2)1/2; thus clocks thick slower (unstable particles live longer, etc).

Cheers

@ Valentin: if you do what you say, that is, measure the moving clocks from the tracks, SRT predicts you will obtain that they all show the same time. If, on the other hand, you synchronisethe two moving clocks on the train, then SRT predicts you will find them desynchronised.

VD: No, as I said, the detectors will be many and placed along the tracks. Each detector will capture the information from the clock which passes by that particular detector.

OK, that makes sense (a bit) but previously you said:

VD: Well, I didn't mean to check the times "on the fly" (i.e. to compare the moving clocks to the stationary clocks while the train is moving).

That seems to be exactly what you propose. The hard part is sampling the train clock at a known point when it is moving.

Anyway, you then go on:

VD: The detector records the data and the time of capture. After the experiment ends all the data from all detectors will be analyzed for comparisons.

What data do you mean? The only information that can be captured from the clock is the time.

To be honest, you seem to be introducing some serious experimental difficulties and at the end you would not be measuring what Stefano is trying to test. His method of two clocks being continuously compared on the train is far simpler and gives a direct, if sampled, confirmation of the effect of acceleration.

VD: @ George Dishman, Please read again, I said twice: data is to be processed "after the experiment ends"

Yes, and I asked "What data do you mean? The only information that can be captured from the clock is the time."

For some suggestions for the on-train version, see the end of this post.

VD: First, I'll quadruple T because we can use a= 0.5 m/s^2 so the total time quadruples.

The time for the train to reach full speed would be increased so you would have four times as long to make the measurement, but the quantity Stefano wants to measure is ax/c2 so it would be reduced by a factor of four. Increasing the separation of the clocks is a better approach but you

VD: Then I'll double that total time again because the train needs also to decelerate, which can be controlled to happen at the same a= 0.5 m/s^2.

It could, but during that time a = -0.5m/s2 so the time difference is reversed. It's a good idea to measure it in both directions though.

VD: Then I'll triple the final speed if I use a technology as in HyperLoop - the vacuumed tube, etc..

Is that a real system currently commercially available? The figure of 430km/h is for the Shanghai Transrapid system. I think the HyperLoop is only at the stage of staring to build test tracks.

VD: So we can get a final Delta T of ~100*10^-11s which means ~1 ns.

No, you just get one quarter of Stefano's number using the Shanghai Transrapid at quarter of the acceleration. However, the proposal is still workable with Stefano's numbers.

In reality, you would keep the cost down by getting permission to put the equipment on the train but letting it run to the normal timetable. In addition to the clocks, you would carry an accelerometer or inertial guidance system to integrate the speed. You probably also need a gyroscope as most of the track curves, especially at the western end.

http://www.transrapid.de/en/archiv/4/shanghai_strecke.gif

Note: 2 m/s2 for 180 seconds means 360 m/s final velocity.

We didn't learn any physics in primary school where I grew up, but it was common knowledge among 10-year olds that sound moves 1 km in about 3 seconds.

I don't recommend approaching NASA or the Jet Propulsion Laboratory with proposals involving 100 meter long trains being accelerated to supersonic speeds along straight line tracks.

@George,

"The LTs tell you how coordinates change under a "Lorentz boost", a simple change of speed of the coordinate system. You can use that to calculate the effect of a change of speed by increment dv and apply it repeatedly every time increment dt. The limit as dt->0 is the way the coordinates change as seen from a frame which is uniformly accelerating at dv/dt. This approach is simply calculus applied to the LTs."

Sorry but it is this which is under scrutiny. A simple change of speed is the acceleration, is the energy change of a body referred to a fixed IRF... you regard as banal, I would not, since we are talking about prinicples here...

Dear all,

so it seems that the situation of a Maglev train which goes from 0 to 430 km/h (as the Shanghai transrapid as George suggests) accelerating at about 1 m/s per sec (or 3.6 km/h per second) can be a reasonable choice.

a) the experiment would last at least around 2 minutes

b) the acceleration would not be too high to produce significant stress as Kåre pointed out.

c) the costs would not be limited as George pointed out.

d) the interest in the experiment would be also pedagogical as you suggest and could be useful for the train company to host it as Kåre suggested.

I would add one more pedagogical feature then. The twin clock retardation or different ageing has been officially verified only in the HK experiment. By keeping a additional  atomic clock at the departure station it would be possible to observer the clock retardation of the clocks on board of the train after every return trip.

The twin effect should be used as a test for a necessary condition, assessing the accuracy of the clocks on board of the train. Infact by monitoring the speed of the train with a GPS it would be possible to estimate the total retardation in a round trip of the clocks on board against the clock at the departure station and verify if the clocks on board follow the predicted retardation which can be found in many text books of relativity.

According to the calculations the accuracy should be more than

vx/c2 = 430*100/10^16=4,3*10^4/10^16= 4,3*10^-12 s  which is a relative estimated retardation of 4,3 picoseconds in a total time  of about 120 seconds

Since with a Rubidium clock it is possible to reach the femto seconds 10^-15 it should not be a big deal to reach a good experimental setup.

KO: 2 m/s2 for 180 seconds means 360 m/s final velocity.

There's a slip up somewhere, the speed advertised is 430 km/h which I make 120m/s. I think I multiplied by 3.6 instead of dividing earlier in the thread. Thanks for spotting the error Kåre.

GD: "This approach is simply calculus applied to the LTs."

SQ: Sorry but it is this which is under scrutiny.

Stefano, I was specifically replying to your earlier question:

SQ: The postulate at the base of the Lorentz Symmetry is the same as the Lorentz Transformations which rely on inertial reference frames, how can you extend the concept to non-inertiality? Sorry don't understand.

The way you extend it is simply by calculus but that was separate from the discussion of how you might test it.

SQ: Since with a Rubidium clock it is possible to reach the femto seconds 10^-15 it should not be a big deal to reach a good experimental setup.

Be careful, while the stability might be that good, measuring a phase shift of that amount won't be as easy. However, I do think with a bit of careful design, your experiment could be done. It is only indicative of course, it won't reach the accuracy of other tests of Lorentz Invariance but that was never the aim.

Yes, for instance, verifies the repulsive gravitation. ,

Stefano> 4,3 picoseconds

Let us get the numbers right: v=430 km/h and x=100 m means that

dt = vx/c2 = (430/3.6)*100/(3*108)2 = 0.133 picoseconds

So, for an experiment of duration t=100 s, one finds dt/t = 10-15, which is about the accuracy of the Deep Space Atomic Clock. My continuing worry is the 0.1 ps time resolution, which I feel pretty sure the DSAC does not have.

I don't recommend approaching NASA or the Jet Propulsion Laboratory with Valentins 3 arguments either.

Valentin> Use the same settings as in the P-R experiment, but place the devices at the respective ends of the train and then set the train in accelerated motion.

Something like that could perhaps be done, even within a single train car. One problem would be to keep the acceleration sufficiently constant for sufficiently long time. Another is to be able to perform the experiment to sufficient accuracy within the available acceleration time. Finally it would not quite correspond to the test Stefano wants to perform.

From a scientific point of view a direct measurement in the time domain could be more interesting, because it is a different approach.

VD: Reason 2.) Let's think what v means? It doesn't mean much for the physics of the accelerated clocks, unless you consider an absolute v [which me and Sergey would welcome]. Under SRT concepts, that v is not something gained by both clocks in an absolute sense, it's just something relative only to the ground (See also Reason 1.).

That's a good point, we should have been clearer about what it means. When we set up the two clocks and each gives a signal at some usefully high frequency, there will be a random but constant phase shift between them. Suppose we start the train moving at a known but low speed, v1, just to put tension into the couplings to get the distance between the clocks to the right length then we measure that phase just calibrate the 'zero' point. Then we accelerate up to full running speed v2. The time difference will be dt = (v2-v1)x/c2, the "v" we have been using is the change in velocity between the two phase readings.

VD: Finally, here's a crazy idea: wouldn't it be better to perform a horizontal Pound-Rebka experiment? (In case none was already performed...) Use the same settings as in the P-R experiment, but place the devices at the respective ends of the train and then set the train in accelerated motion. How about that?

That's what the rest of us have been talking about throughout. The train allows a larger distance between the clocks as the train is 153m long while Pound and Rebka's tower was only 22m high. You can try to do it with uniform acceleration or you can plot the phase versus speed and take the derivatives to get the same result without having to control the train (that keeps the cost down).

George,

"That's what the rest of us have been talking about throughout. The train allows a larger distance between the clocks as the train is 153m long while Pound and Rebka's tower was only 22m high. You can try to do it with uniform acceleration or you can plot the phase versus speed and take the derivatives to get the same result without having to control the train (that keeps the cost down)."

it is more obvious in that case that you would have first order Doppler effect between the head and the tail predicted by the theory. In that case you would not test the oscillators as atomic clocks  but as photon emitters and you will see what happens to radiation between equally accelerating devices. I don't know if this expeirment has been performed either, but it is not significative as the one with atomic clocks.

The Pound and Rebka experiment with photons for gravitation is not exactly the same as the one shown in the paper of 2010 testing dilation of atomic clocks at different heights.