Gravitational waves are very well described in the book by S. Weinberg   http://www.amazon.com/Gravitation-Cosmology-Principles-Applications-Relativity/dp/8126517557/ref=sr_1_1?ie=UTF8&qid=1455411165&sr=8-1&keywords=gravitation+and+cosmology.  I strongly advise everybody  to read Chapter10 of this book prior to rushing with whatever comments. 

Please answer the question, do not advertise your own work when it's not an answer

The output from the photodetectors combines the two arm optical signals so what happens thereafter is probably immune to the effects in principle.

They also quote having ~100k other sensor signals for checking and the wave signal would appear in various forms in some of those while noise and false detection sources might appear in others. Validating all of those is part of the work that has taken so many months since the event.

Thank you George

My (very poor) understanding is that the passage of the GW alters space-time: the impact on the arms of the interferometer changes their length, which creates interferences that can be measured.

But when the GW reach the measurement devices (whatever they are : photodetectors, phototransmitters, optoelectronic and electronic components and boards), they also locally alter space-time. There may be an impact on these devices (length change of the tracks inside an integrated circuit or an optical transducer, or something else) which might change the measurement result's accuracy.

Is this correct? Is this too small to be accounted for?

All the other sensors are equally affected because they would suffer from the same space-time alteration. So even so many sensors does not mean the result is accurate.

FA: Is this correct? Is this too small to be accounted for?

It is not correct in principle because the solid material of which devices are made is held together by EM forces so completely unaffected. The wave only affects the mirrors because they are suspended in a way that isolates them from any mechanical determination of the positions at the frequencies of interest.

It is also far too small to matter. Remember the signal is a proportional length change orders of magnitude less than the size of a proton over four km so how much change would that be in the few mm inside a component or cm of a sensor? Another five or six orders of magnitude down. The effects of local temperature variations causing expansion and contraction would be vastly greater, if they were significant, none of the electronics would work at all.

Fabrice,

Yes. Not only do gravitational waves distort space, they also distort time (or said more accurately, the rate-of-flow of time). GWs cause ripples in space and ripples in time. If a GW with a frequency of 100Hz passes through a detector, the rate-of-flow of time will fluctuate at 100Hz. Of course, you would not be able to measure this locally with only one detector; you would need at least two detectors separated by a large distance. Linking the two detectors and comparing their outputs will show a phase difference in time fluctuations.

The gravity waves are fluctuations in the stress energy of space time that occur in actions which are not balanced by reactions except by propagation of the fluctuations to far distances.

The gravity waves do affect their measuring devices, and one branch more than the other, which leads to a detectable signal.

Compare the detectors to radio amplifiers. A fairly ordinary amplifier can demodulate a signal energy that is one part out of 40 decibels of noise energy. That is one part of 10 to the 40th power. The key is that the amplifier is constructed to recognize a pattern either frequency or amplitude, and reject the remainder. The detectors are just an extreme version of amplifiers looking for predictable patterns of amplitude and frequency.

Another example is a mechanical oscillator containing a mass supported by springs and driven at a specific frequency. The gravity wave affects the mass more than it does the springs causing a momentary change of amplitude and frequency in the oscillator. The LIGO event occurred quickly which requires the mass and springs to be microscopic, about the size of photons in the detectors. Then the oscillators are like zero point assemblies of vacuum potentials, or virtual states if you prefer, altering the beams in the detectors and the branched tubes that contain them.

"A fairly ordinary amplifier can demodulate a signal energy that is one part out of 40 decibels of noise energy. That is one part of 10 to the 40th power."

Jerry: 40 dB correspond to a factor 100 in signal amplitude.

Your reference factor of 100 applies to current or voltage in an amplifier, but not both at the same time. Power gain of 40 decibels is a 10,000 factor in amplifier modern design.

I referred to sensitivity ratio of signal acceptance to noise rejection ratio in a demodulator like a part of a tuning circuit, not the current, voltage or power gain of an amplifier.

It's been nearly 50 years since I studied electrical engineering in college. As I recall tuning circuits and demodulators were in one course, and amplifier design and gain were in a different course, both of which were core requirements for one of my degree programs. That was a long time ago, so I defer the question about 10 to 40th power noise rejection to readers with recent and greater expertise.

Maybe some reader who is more involved in LIGO could give a clarification of what sensitivity signal to noise the detectors have and what amplifier gain is involved after demodulation.

Thanks for your comments which are helpful in the clarification. I suspect the noise rejection is very large at LIGO, comparing two detectors in different places, intensified by use of templates for pattern recognition predicted from years of intensive calculations.

In the previous comment I should have said the tuning circuits and demodulators associated with a fairly ordinary amplifier can select a signal out of noise ...

Thierry, I would advise you to read  carefully book by Cornelius Lanczos

https://books.google.com/books/about/The_Variational_Principles_of_Mechanics.html?id=ZWoYYr8wk2IC  about proofs related to time. Also, keep in mind that it would be impossible to build any accelerator, to fly to any planet, etc.should people would have REAL problems with measuring time. Speaking of which, how astronomers determine the distance to galaxies, etc? How, after all, you determine the distance to the Moon, or to the satellite which is on the orbit...For the person who is in General Science Journal, these are important topics , especially, if one is thinking about histories of this or that discovery connected with time, e.g. how for the first time the size of earth was estimated.Etc.

Peter Hahn wrote: If a GW with a frequency of 100Hz passes through a detector, the rate-of-flow of time will fluctuate at 100Hz.

Peter, I thought plane waves (which are really all we will see because the range is always far larger than the detector size) had only transverse spatial components, no time component. Do you have a reference for your information please?

JD: Maybe some reader who is more involved in LIGO could give a clarification of what sensitivity signal to noise the detectors have and what amplifier gain is involved after demodulation.

Jerry, the actual systems at LIGO are incredibly complex and I'm not involved but they do publish the overall noise graph in the linked page, see figure 3.

https://losc.ligo.org/events/GW150914/

GD Thanks for reply. I believe the figure 3 you referenced is the output signal to noise of the GW event for each of the two detector locations evaluated individually. It can be copied and displayed on a larger format to make reading easier. In the output channel the signal should be much greater then the noise as shown in the figure. The input signal to noise must be enormously more chaotic and not decipherable except by comparing the two locations and looking for predictable patterns. For input of raw data the noise is usually much greater than the signal. It is the job of the circuit designers to reject nearly all of the noise while conserving almost all of the signal.

 In some of the LIGO technical documents going back 5 years or more the tuning and demodulation sections were based on greater noise reduction than common practice by using templates of predicted wave frequencies and amplitudes changing in a pattern and time frame. It is similar to but enormously more complex than communication electronics.

GD about the gravity waves at the detectors, I believe the shape should be very nearly planes of equal gravity potential because of great distance, waves modulated in both frequency and amplitude, longitudinally not transverse, but with a time component. This is not proven but is open for suggestions. It follows fairly conventional science. Changes to a gravity field are usually explained as longitudinal waves, compared to transverse polarization waves of AC electromagnetic fields.

Hi Jerry,

JD: I believe the figure 3 you referenced is the output signal to noise of the GW event for each of the two detector locations evaluated individually.

No, they are purely the noise output spectra. You can easily see the spikes due to local mains hum and harmonics. It would be interesting to overlay the GW150914 signal onto that chart, it would be a line rising out of the noise then dropping back. You can see the signal in Fig. 1, the frequency is the vertical scale and the colour indicates the energy which is related to the strain.

JD: It is the job of the circuit designers to reject nearly all of the noise while conserving almost all of the signal.

If you have a single frequency or narrowband signal, that can be done by filtering but the source frequency is a chirp here, rapidly rising. A chirp can be searched for using their range of templates but in the case of GW150914, the signal was 24 times greater than the noise. In fact they were lucky in a way because the signal was so large, it was almost out of the range for which templates had been produced.

JD: I believe the shape should be very nearly planes of equal gravity potential because of great distance ...

They are certainly indistinguishable from plane waves as far as we are concerned.

JD: ... waves modulated in both frequency and amplitude, longitudinally not transverse, but with a time component. This is not proven but is open for suggestions. It follows fairly conventional science.

No, all the descriptions I have seen in textbooks and on the web say that GR predicts that the amplitude is the peak strain and that the waves far from the source are purely transverse spatial with no time component.

JD: Changes to a gravity field are usually explained as longitudinal waves, compared to transverse polarization waves of AC electromagnetic fields.

GR predicts they are transverse like EM but the format is quadrupole whereas EM waves are dipole. If you have some alternative pet theory, I couldn't comment.

GD ...  I find 250,000 references to support your transvers gravity wave, and another 250,000 references to support my longitudinal gravity waves. So I guess the question is not resolved, except that the powerful gamma ray burst arrived 0.4 seconds after the large gravity wave departed, maybe suggesting that longitudinal waves travel a bit faster than transverse waves as is observed in other types of vibrations.

Jerry, there are a lot of cranks out there who have their own aether-based theories and spend a lot of time promoting them, the number of hits on Google is not an indicator of accuracy. If you check a text book, you'll find there is no controversy at all, gravitational waves are transverse, quadrupole. I've attached a random link I found with a quick look but you can probably find many more like it.

http://www.tapir.caltech.edu/~teviet/Waves/gwave.html

GD Thanks for the link.

We should make a distinction between two topics that seem to be getting mixed together. One topic is how a graviton vibrates, which is the subject of most references, but not the topic I am discussing. I don't argue vibration of gravitons. Another more interesting topic and my intent in this discussion is what happens to a sparse three dimensional field of small mass particles as the gravity wave passes through them. They accelerate and decelerate slightly in rapid sequence, but in which directions first and second?

Quadrupole form of a graviton is not in dispute by those who support quantum gravity. I do support quantum gravity considering all the known causes of gravitational curvature have quantum character. Quantum gravity in its present form is out of favor with many scientists for other reasons.

Graviton quadrupole gives the basis of experiments on field effect and gravity induction using interference patterns of radio microwaves and polarized lasers. Since photons are dipole and gravitons are quadrupole in theory, two photons should interact out of phase to make a graviton. The photons have provable transverse polarization, but precisely in the components that are canceled out in the interference pattern. The publications of curl free vector potential lead to conclusions, but are not the main theme of this discussion.

For this discussion there is a historical sticky bead argument for a transverse gravity tide attributed to Hermann Bondi but originated in 1957 by Richard Feynman who was not willing to put his name on it.  Reading the fine details that example gives a sliding bead on a stick perpendicular to propagation of a gravity wave. The character of the wave is such that it adds a different potential to the middle of the stick than it adds to the ends, causing the bead to slide back and forth on the stick as the wave arrives and departs.

The example fails when the stick is curved into a circle with plane perpendicular to the propagation, but with 20 beads equally spaced.  There are no centers or ends of the stick to receive greater or lesser potential from the wave. We are left with several choices to make, depending on the character of the wave.

Peter Bergmann published two chapters on  this topic in 1968. If the stick circle is close to the black holes and near to  the plane of their orbits , he claimed the wave should be polarized causing a stick ring to become elliptic shaped, first on one axis then on another. This is the transverse point of view based on theory.

The longitudinal component is based on celestial mechanics which I studied under Dr Olesiak (RIP) from Warsaw University. Center of gravity is based on [ mr ] while acceleration of gravity is based on [ m/r2 ]. First an example is given for the proof using a three body problem in which the stick ring will eventually become the third body. When two large bodies are orbiting with a third small body in the same plane the three bodies approximately orbit their mutual center of gravity as long as they are far apart. The difference between approximation and reality occurs when the small body has a close encounter with one of the other bodies. Its orbit is changed in a predictable and verifiable way, not a theory. The reality is that the gravity potential at a distance in the plane is not exactly the same when the two large bodies are equidistant as it is when one body is farther away than the other. The acceleration of gravity changes at any distant place as the gravity waves sweep past it. The ring stick is accelerated more strongly one moment and less strongly the next in the direction opposite of the wave propagation.

A gravity wave is just a momentary passing change in the local strength and direction of the gravity field. Polarization is OK to argue as a theory. Vibrating gravitons are OK too, even in 8 different modes that are offered to me, as long as it is theory or provable. I borrow an argument from Feynman, science should say what physical and measurable thing a gravity wave does as it passes by.

Please, read this paper http://arxiv.org/abs/gr-qc/0501041

Hi Jerry,

JD: We should make a distinction between two topics that seem to be getting mixed together. One topic is how a graviton vibrates, which is the subject of most references, but not the topic I am discussing. I don't argue vibration of gravitons. Another more interesting topic and my intent in this discussion is what happens to a sparse three dimensional field of small mass particles as the gravity wave passes through them. They accelerate and decelerate slightly in rapid sequence, but in which directions first and second?

Good point, I've not seen any detailed work on gravitons, I am only talking about gravitational waves. All the on-line stuff I've seen is talking about those and the descriptions I've seen of gravitons say they would be a massless spin-2 boson. A massless particle travels at the speed of light so couldn't have any oscillatory longitudinal motion so I don't know what you've seen on those.

JD: For this discussion there is a historical sticky bead argument for a transverse gravity tide attributed to Hermann Bondi but originated in 1957 by Richard Feynman who was not willing to put his name on it. Reading the fine details that example gives a sliding bead on a stick perpendicular to propagation of a gravity wave.

Right, that's the correct conventional view.

JD: The example fails when the stick is curved into a circle with plane perpendicular to the propagation, but with 20 beads equally spaced. There are no centers or ends of the stick to receive greater or lesser potential from the wave. We are left with several choices to make, depending on the character of the wave.

True but you can remove the stick and just consider free-floating particles. The addition of the stick was only to demonstrate that the wave must carry energy. When you consider free test particles, you get the transverse motion patterns shown in the Wikipedia article, these accord with all the text books I've seen too. See the link and read in particular the second paragraph which makes clear the direction of propagation.

JD: The longitudinal component is based on celestial mechanics which I studied under Dr Olesiak (RIP) from Warsaw University. Center of gravity is based on [ mr ] while acceleration of gravity is based on [ m/r2 ]

That sounds as though he was trying to use f = Mm/r2 which is the Newtonian formula. That had instantaneous effect so did not allow propagating waves.

The variation due to the inverse square law gives you the tidal component which drops off as r-3 whereas the strain amplitude of gravitational waves falls off as r-1. That is why any longitudinal effects become negligible at more than a few times the orbital radius of the binary system.

JD: I borrow an argument from Feynman, science should say what physical and measurable thing a gravity wave does as it passes by.

I totally agree with that, the Wikipedia page specifically shows the effect.

P.s. I've just started looking at the paper Arkady posted and it looks excellent, I would recommend reading it too. Page 7 seems to be particularly relevant to this discussion.

https://en.wikipedia.org/wiki/Gravitational_wave#Effects_of_passing