Science a GoGo's Home Page Forums General Discussion Physics Forum Lets try science discussion

I saw this image today and it got my attention from the article http://phys.org/news/2014-06-method-gravitational-constant.html on the calculation on the universal constant G over a 32 year history



Those are actual error bars on those results that the scientists have calculated on their experiments. Now G is very hard to isolate so either a lot of different scientists were being overly optimistic in there error analysis or there exists a problem.

I don't know the answer but it does bear some thinking about and some investigation and reading.

it doesn't look like the universal gravitational constant
is cooperating very much unless its trying to say something
other than what the experimenters wanted it to say.

it almost looks like its more like a universal variable constant.

if there is such a thing yet!

That is indeed one of the number of possible ways of explaining it. I gave you greater than expected experimental error in the post above , so we therefore have 2 reasons so far.

See if you can actually think of any other explainations?

The first step in evaluating a problem is to list all the possibilities.

there just might be hope for you yet , orac.

you are right there is no way to find a exact solution
without first listing all possibilities.

you can find a close approximation which should say to you
that something was not accounted for.

as far as finding a exact universal constant of gravitation goes all we can do is use a pre-defined constant that is a close approximation of G that is the result of experiments and data gathered by the instruments that we construct.

and because there is no such thing as a true variable constant that has a +/- error range built into it , the number that we use to represent G will always be a variable to a degree even though we call it a constant.

however in the future as we break away from the chains that garbage science has on us and we begin to explore the cosmos
a precise number for G will be demanded.

I didn't start out to figure out why the results vary so much, I started out trying to figure out just how much the variation is. Checking out the chart the lowest end of the lowest error bar is at just about 6.67009. The highest end of the highest error bar is just about 6.67508. The average would be 6.67285. The range on either side would be very close to .00250. That gives us a maximum variation of +/- .037%. Keep in mind that this is the total variation possible from all the measurements included in the chart. While we would really like to get better than that it really isn't too bad. Particularly when we are trying to measure a very small signal in a very large noise. At least all the measurements are in the same order of magnitude. That suggests that we are closing in on the correct value.

The fact that the latest measurement,using an entirely different technique, is lower than all the others raises some questions. But it is hard to evaluate where the differences come from without being able to completely understand the technique. Possibly the technique doesn't work as well as they think. Also possibly their technique is just better than the other techniques and they are getting a more accurate number. As they refine their technique we may get better numbers.

Now that one group has come up with a completely different technique it may inspire others to do the same. Then when we get several different techniques working we will be able to come up with a more accurate number for G.

Bill Gill

We can't just explain it away like your first statement Bill those results have error margins on them. Do a refresher on what the bars mean http://en.wikipedia.org/wiki/Error_bar and the result is 1.5 standard deviations from the published and accepted value.

Lets assume nothing is wrong which is the most likely case that still means there are systematic errors that numerous groups have overlooked or under-rated in their experiments. There is a problem and you can't pretend there isn't because those results are well outside each others range. If the Higgs was that far outside it's range would you allow the claim of discovery to stand and it is is far trickier measurement.

We are looking at vastly different things but lets look at the Higgs result in something sort of similar display ... image courtesy of jester



I haven't got a clue on why the results disagree I am pulling each experiment paper and looking at the systemic error estimates to see if I can spot whats going wrong as I suspect are a group of people after the article which was the point in publishing the result.

I also note we are late to this party it was actually picked up last year (http://www.scientificamerican.com/articl...ate-slide-show/). Note the take home message which I totally agree with

I wasn't trying to reach any particular conclusion in my statement. There is no possibility that I can figure out what is going on. What I was doing was trying to get a handle on the magnitude of the problem. Obviously the reported values of G are kind of all over the place, although they do at least cluster in the same ball park. Even the values measured by the more traditional techniques have a lot of variation. The fact that the new value is lower than any of the others does require some explanation, but at the same time there is the one that is higher than the others. Based on the scatter in the measurements it is obvious that measuring the value of G is not a simple project. There are a lot of variables that have to be taken into account, that is what I was referring to as noise. Since the measurement is influenced by the mass of every item in the area, and by any electromagnetic fields in the area it is really hard to make sure that none of them are influencing the result. The problem is the extremely small value of G. With a value that small almost anything that takes place around it will cause errors.

Bill Gill

Do you think the Higgs which barely interacts with matter at all and you are trying to glean from femtobarns of extreme high energy collision data is easy?

I and obviously at least one or two others have been surprised at the inability to nail the value with more accuracy when compared to the Higgs it would rate as trivial. The issue is lack of dollars, time and effort probably because it was not really causing any problems and offered no real hope of any new physics.

The reason for writing the articles they have is to mount a case for funding to research the problem and I suspect they will get it because the results are starting to get problematic.

What is really interesting is when you consider the above against the nice work by Julija Bagdonaite et al which shows the electron to proton mass ratio has not changed in 7 billion years.