Gravitation constant G can vary 0.1% in 6 years

[John Cuthber]

John..., why not address the science in the OP?

Because I couldn't find any.

 

 

So I addressed the title.

"Gravitation constant G can vary 0.1% in 6 years"

And, as far as I can tell that's impossible because, for example, pendulum clocks rely on g being constant (locally) which implies that G is constant.

The clocks work, so G must be substantially constant (or, at least, it must not vary by as much as 0.1% over 6 years).

It's not that clocks are the only things that measure G but they are among the best known.

Gravimeters of various types do it and (at that level of precision) even a spring balance would do the job.

just about every electronic balance would stop working properly- and people would notice because they run calibration checks on them.

 

Incidentally, someone has been measuring g in the same place for years.have a look at fig 34 here

https://jila.colorado.edu/sites/default/files/assets/files/publications/precision%20measurement%20of%20gravitational%20quantities.faller.pdf

The direct measurement doesn't show a 6 year periodicity

Edited August 23, 2015 by John Cuthber

[Strange]

Because I couldn't find any.

 

Really?

 

Well, here is a link to the paper itself, if that is more useful:

 

 

About a dozen measurements of Newton's gravitational constant, G, since 1962 have yielded values that differ by far more than their reported random plus systematic errors. We find that these values for G are oscillatory in nature, with a period of P = 5.899 +/- 0.062 yr, an amplitude of (1.619 +/- 0.103) x 10^{-14} m^3 kg^{-1} s^{-2}, and mean-value crossings in 1994 and 1997. However, we do not suggest that G is actually varying by this much, this quickly, but instead that something in the measurement process varies. Of other recently reported results, to the best of our knowledge, the only measurement with the same period and phase is the Length of Day (LOD - defined as a frequency measurement such that a positive increase in LOD values means slower Earth rotation rates and therefore longer days). The aforementioned period is also about half of a solar activity cycle, but the correlation is far less convincing. The 5.9 year periodic signal in LOD has previously been interpreted as due to fluid core motions and inner-core coupling. We report the G/LOD correlation, whose statistical significance is 0.99764 assuming no difference in phase, without claiming to have any satisfactory explanation for it. Least unlikely, perhaps, are currents in the Earth's fluid core that change both its moment of inertia (affecting LOD) and the circumstances in which the Earth-based experiments measure G. In this case, there might be correlations with terrestrial magnetic field measurements.

http://arxiv.org/abs/1504.06604

 

 

 

And, as far as I can tell that's impossible because, for example, pendulum clocks rely on g being constant (locally) which implies that G is constant.

 

I don't think anyone is suggesting that G is changing.

 

 

 

Incidentally, someone has been measuring g in the same place for years.have a look at fig 34 here

 

That article says that the situation with regard to measurements of G is "not good" and mentions about half a dozen experiments to measure it. He also notes that the measured values differ by a few parts in 10⁴ (much more than the errors in measurement). This seems entirely consistent with the paper referenced by the OP.

Edited August 23, 2015 by Strange

[Sensei]

If either g or G varied then the clocks wouldn't work and people would notice.

 

g is varying even from place to place on Earth.

 

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

 

g vary as a function of latitude.

That's it, it depends on distance to Earth's center.

But if it does so with Earth, it must also does with Sun, and Moon, and Milky Way center.

Once we're the closest to the Sun at noon, and the furthest from the Sun at midnight.

 

ps. I would start analyze from checking in what exactly days of year were G calculated in the past (function of distance to Milky Way center)..

 

ps2. Variation of g is used during exploration of resources https://en.wikipedia.org/wiki/Gravity_gradiometry

Edited August 24, 2015 by Sensei

[John Cuthber]

 

 

I don't think anyone is suggesting that G is changing.

 

Gravitation constant G can vary 0.1% in 6 years

It seems like the OP does.

 

g is varying even from place to place on Earth.

 

 

 

I know, but it is the variation in time that is under discussion here. I already explained that.

 

The spatial variation of g is big, but, once you build an observatory and bolt a clock to the wall, the temporal variation in g is small.

If the pendulum clock keeps track with the rotation of the earth (as measured by the apparent motion of the stars) then the clock must be keeping good time.

If either g or G varied then the clocks wouldn't work and people would notice.

[Strange]

Gravitation constant G can vary 0.1% in 6 years

 

That is just sloppy journalism (as noted by swansont in post #3).

[John Cuthber]

Is acsinuk a sloppy journalist?

I'm getting confused here.

 

Anyway, as I pointed out, if you look at the local value of g it doesn't vary much- but it is vastly easier than G to measure accurately.

So, it looks like there must be another explanation.

I have a relatively simple one based on years of looking at "error budgets" and such like.

Let's have a quick look at the data.

 

the data in the graph have error bars based on a calculated uncertainty.

If those were all correct then all the error bars (strictly 95% of the error bars) would overlap some value that was actually correct.

 

They don't.

So there's either some lack of constancy of G, or the error bars are not correct.

 

Imagine that you replotted the data but made the error bars 5 times bigger.

Then they would (nearly) all overlap and you could claim that the middle of the overlap was a consensus value.

 

That would only happen if there were (ironically) errors in the error bars.

but the error bars are the result of calculating the effects of things they know about.

So, if there is some mystery 6 year oscillation- then it's something they don't know about but.

if there's something they don't know about, then the error bars are wrong- specifically they are too short- and nobody knows by how much.

so you can't use those error bars to justify the sine wave through the data.

To cut a long story short, if the data is so bad that there might be a 6 year oscillation then the data is too bad to demonstrate that oscillation.

 

My simple guess is this- they modelled the errors as normally distributed but they are in fact a fat tailed distribution (and believe me there are lots of those in science)

https://en.wikipedia.org/wiki/Fat-tailed_distribution

 

Any takers?

 

 

 

[Strange]

Is acsinuk a sloppy journalist?

I'm getting confused here.

 

I don't know why you are confused, he simply quoted the headline (I think that these are often not written by the author of the article, so there is extra scope for poor wrtiting).

 

 

Imagine that you replotted the data but made the error bars 5 times bigger.

 

Do you have any reason to do that (other than covering up the pattern in the data)?

[John Cuthber]

No, I don't, but it would be interesting to look at how inaccurate estimates of uncertainty typically are.

But your question misses the point; do they have any reason for that sine wave?

 

Even a factor of 3 would make the data look more like a constant than a sine wave, and it's not as if the sine wave goes through all the data's error bars anyway.

 

Is some mysterious 6 year cycle more likely than the estimates of uncertainty being optimistic? Nobody knows, but people often miss out error sources and thus underestimate the overall error

As I said, even if the 6 year effect is real, the people reporting it didn't allow for it (because they didn't know about it) so their assessment of error is wrong.

 

Even if they are right, they are wrong.

Edited August 24, 2015 by John Cuthber

[swansont]

But the point of the Shortt clock was that you didn't keep tweaking it because you didn't need to.

These were not quite that hands-off.

 

"Many Shortt clocks have shown a progressive slowing down, as if the pendulums were slowly lengthening or as if the bobs were settling. On one of the clocks at Greenwich an inver bob was substituted in 1929."

 

The Precise Measurement of Time. Alfred L. Loomis.

Monthly Notices of the Royal Astronomical Society. 91 (1931): 569

 

He also measured how as you evacuated the chamber of a Shortt clock the period changed, because the pendulum drags air along with it. As the pressure goes down (or goes back up), the effective location of the mass of the pendulum changes. So a leak is a problem, and leaks happened.

 

He's the one that measured the effects of the moon on pendulum clocks by comparing them with crystal oscillators (that's the next paper in the journal). They had to account for earth tides and the effect of the mantle compressing and relaxing from sea tides, since they were close to the ocean. The effect was several parts in 10^8. So presumably if there was geologic settling or upheaval of the same order of magnitude as the tidal motion, you would also see this, and we know this effect is happening — it's one of the factors affecting the earth rotation rate. Currently of order 1 cm/year https://en.wikipedia.org/wiki/Post-glacial_rebound

 

Speaking of which, if there was a change in G, the rotation rate of the earth would change. If G got smaller the earth would tend to expand, and the centrifugal bulge would be larger, so we would slow down. Any amount that was common to the slowing of the clocks would be something we wouldn't measure with astronomical measurements.

 

I don't know why you are confused, he simply quoted the headline (I think that these are often not written by the author of the article, so there is extra scope for poor wrtiting).

 

It doesn't look like he did. The headline talks about the measurement, not the constant itself. They also don't say it's 0.1% — the actual fluctuation is less than half of that.

[Strange]

It doesn't look like he did. The headline talks about the measurement, not the constant itself. They also don't say it's 0.1% — the actual fluctuation is less than half of that.

 

You are right. I don't know why I thought that - apologies all round!

[John Cuthber]

They were not perfect, but they were good enough to show that the Earth's rotation wasn't constant.

They were certainly good enough to show up an effect at 0.1% in six years (and I'm happy to round 0.06% or so to 0.1% for the sake of convenience, though I'm slightly embarrassed to have only just checked it)

A good clock would keep time to 1 second in a year and would run for a few years before it failed- for whatever reason.

So, unless they were really unlucky and picked a local extremum of that cyclic variation they would expect to see a change of something like a third or a sixth of the amplitude.

Even if that that amplitude is just 1% of 1% ( 1 in 10^4) in six years than it's about 1 in 10^5 for a typical year.

There are 31 million seconds in a year, so the error should be something like 5 minutes.

A clock that keeps time to a second a year should be able to see an error of 0.001% in six years with a factor of about 300 in hand.

 

Of course, if some mechanism keeps the pendulum clock and the earth's rotation locked, then no method will detect a difference between them.

 

I guess it's unlikely that anyone seriously times the earth with a pendulum clock any more (perhaps slightly sad but...)

But they do time it with some very good clocks (take a bow Swansont et al) and we know how much it varies. And the variability is about what the very good pendulum clocks said it was so unless the variability of the earth's rotation has suddenly changed we know that the pendulum clocks must have been pretty near right.

If G was varying then g would also have changed and that would have messed up the pendulum clocks. It didn't.

 

I just think my explanation - they are not as good at estimating errors as they think they are- is at least as credible as the cyclical variation (in something unspecified) they have suggested- not least because on one hand, that variation would have been spotted by other means and on the other hand, other data show that people are bad at estimating errors.

 

(I used to be part of the team run a QA scheme and believe me they think they are a lot better than they actually are)

[acsinuk]

Your quote "Of course, if some mechanism keeps the pendulum clock and the earth's rotation locked, then no method will detect a difference between them."

Yes, there is a common magnetic locking mechanism and it is a variation in this that sets relativistic time or time dilation. Looking from the magnetic centre of the galaxy the sun has a relative velocity that sets the solar time reference frame. BUT we are rotating our sun so from the galactic centre point of view our relative time is changing a little as we rotate the sun which changes our relative velocity

Now gravity in the MKS system is effected if time changes thus the variation recorded are correct. We need several labs to make make regular recordings so that we can verify the exact variation accurately.

[John Cuthber]

What I meant was that if you have two methods for measuring time, and they are both altered in the same way then you will never notice the alteration.

so if you were using orbital periods and pendulum clocks to keep time and G changed you might not notice because it would affect both "clocks" in the same way.

But a quartz clock or one based on the radioactive decay of uranium or something would make it clear that something had changed.

[Phi for All]

Your quote "Of course, if some mechanism keeps the pendulum clock and the earth's rotation locked, then no method will detect a difference between them."

Yes, there is a common magnetic locking mechanism and it is a variation in this that sets relativistic time or time dilation. Looking from the magnetic centre of the galaxy the sun has a relative velocity that sets the solar time reference frame. BUT we are rotating our sun so from the galactic centre point of view our relative time is changing a little as we rotate the sun which changes our relative velocity

Now gravity in the MKS system is effected if time changes thus the variation recorded are correct. We need several labs to make make regular recordings so that we can verify the exact variation accurately.

 

!

Moderator Note

NO Speculating in the mainstream sections, please!

 

You started out asking about legitimate science. If you want to go out onto the ice, please start a new thread in Speculations.

 

There's no need to respond to this in this thread, but you can report it if you object.

[acsinuk]

Sorry for that. What we need is more accurate regular measurements of G by several labs to get to the truth.

[John Cuthber]

How much would G need to vary before we noticed the change in the orbit of the moon?

Certainly 0.1% would mess up the tide tables.

How about how big a change in G would it take before our TV satellite dishes were all pointing in the wrong direction?

[imatfaal]

In a two body gravitational system the seperation varies inversely with G (note1). We can (and do) currently measure the earth moon distance to an accuracy of 3x10^-2 metres over a distance of 3.85X10^8metres; 3cm in 385,000 km via the apollo 11 lunar ranging equipment. We would immediately notice that level of change

 

BTW this level of accuracy far exceeds the level of accuracy we have of G anyway which is only accurate to a few thousandths of a percent of its value.

 

1. I think it also changes the eccentricity - which might with a bigger shift make more of a difference

[swansont]

In a two body gravitational system the seperation varies inversely with G (note1). We can (and do) currently measure the earth moon distance to an accuracy of 3x10^-2 metres over a distance of 3.85X10^8metres; 3cm in 385,000 km via the apollo 11 lunar ranging equipment. We would immediately notice that level of change

 

 

Not immediately. The precision of the ranging experiment took several years' worth of measurements to tease out.

[imatfaal]

 

Not immediately. The precision of the ranging experiment took several years' worth of measurements to tease out.

 

Ah - I did read that the amount of reflected light is perilously small but I hadn't understood it took that long/that many mesaurements to get the precise reading

[swansont]

Ah - I did read that the amount of reflected light is perilously small but I hadn't understood it took that long/that many mesaurements to get the precise reading

I went to a colloquium on this a while back. The error in any one measurement is several cm, as I recall.

[John Cuthber]

One of you says "The error in any one measurement is several cm", and the other "an accuracy of 3x10^-2 metres"

There's not much in it, and the nearest Km would still easily be close enough to spot a change of 0.1%.

[acsinuk]

In a 2 body gravitational system the G varys with increase or decrease in separation of the 2 bodies. So as our sun rotates the black hole hub at the centre of our galaxy then we should expect its G and our G to vary in sympathy with our stars changes in distance from the galactic hub.

Further, as our planets rotation the sun, it takes us 150 million km closer or further from the galactic hub every year but this would be a tiny amount compared with the suns variations from the galactic hub.

[Strange]

In a 2 body gravitational system the G varys with increase or decrease in separation of the 2 bodies.

 

No it doesn't. It is a constant.

 

"Constant: staying the same : not changing"

http://www.merriam-webster.com/dictionary/constant