0955: "Neutrinos"

[Wnderer]

@^
um what?

They found neutrinos traveling faster than light. How much faster? 1.0000248 times faster. That's about 4.6 miles per second. That's fast but it's not enough to make it to 3x10^8 meters/second or make to it 1 foot/nanosecond. It's just 299799893m/s compared to 299792458m/s. It's too big to be the ln(20.044). But 1.0000248 is equal to 666^(6/(6+6))^(6+6+6). I think it's funny.

Devil's Neutrinos.jpg (1.59 KiB) Viewed 3089 times

They went looking for the God Particle and found The Devil's Neutrinos. The Devil's Neutrinos coming soon on DVD and BluRay.

[gmalivuk]

Would you put a bet on them?

If I had any money, I would. CERN knows what they are doing, for the most part. It could be that this is a fluke error, or perhaps a misunderstanding of the results. But I doubt it.

You don't have to have money now you know. Loosing party just need to pony $$ in a few years, when it's resolved. $200?

Indeed. Just send me a promissory note, collegestudent22, saying you'll send me $200 in two years if this doesn't pan out, and I'll send you one promising to send you $200 in two years if it does. Surely you can manage to save up a bit less than 30 cents a day.

Deal?

[keithl]

There is a lot to go over in the Arxiv.org paper, but my preliminary take on it is that it is built on a rather large heap of assumptions about timing and distance. If the position of the Gran Sasso detector was incorrect by 20 meters, that would explain the discrepancy. If the length of the cables, and the prop delay of the digital circuitry, or metastability in the events triggering the counters was causing systematic errors, that could easily result in a 60 nsec error (look up "sparkle codes" in analog to digital converters)..

Further, they are differencing some rather wide probability distributions, using gaussian/normal assumptions. Stat 101 - the difference of two distributions is not the same as the distribution of individual differences.

But my biggest concern is that they are taking the time difference between two dissimilar events: the alleged neutrino production time at a target at CERN, and the detection time (in a very large detector) at Gran Sasso. All sorts of room for mis-measurement to occur.

If you want to establish time of flight, you should measure the start time with an identical neutrino detector much nearer CERN. An neutrino detector in a side tunnel of the Mont Blanc highway tunnel (?) might be appropriate. If it is off axis from the beam, the luminosity is reduced, making up for the inverse-square increase in neutrino flux. Then you can measure timing differences between the two detectors, for each event, and only count the events when there are pairs of detections. It is not as good as measuring the same neutrino twice (or using a really long distance, say to sn1987a ), but it eliminates a lot of assumptions.

Physicists tend to be wildly optimistic about their instruments, and make many untested assumptions about the gathered statistics. Having designed integrated circuits for Tektronix for two decades, their optimism is unfounded. Instruments do not measure, they produce signals in reaction to stimuli. If the engineers who designed the instruments did not test for those stimuli, or optimized for different stimuli or different output signals, users will get inaccurate results.

Many recent "startling discoveries" may be similar misinterpretations of instruments.

[iabervon]

So, this group is actually trying to measure something entirely different from the speed of neutrinos (flavor-changing behavior, in particular). But they have to demonstrate that the neutrinos they detect were the ones that were emitted, so that they can tell they're not just seeing the neutrinos of the wrong flavor that happen to be going through the area from a different source during the experiment. They want to check based on narrow time windows, so they're looking at when the neutrinos could arrive, based on the range of possible speeds given by other experiments. But they found that the neutrinos got there before they should have been able to (based on standard assumptions about neutrinos and relativity), which is a bit awkward for their otherwise robust demonstration that the neutrinos that turn up electron-flavored left CERN muon-flavored. So they're soliciting help in figuring out when they should expect to detect neutrinos, because nobody's worked that out exactly. They'd like nothing better than to be told "the surveying team screwed up the correction for the direction of gravity at your site when locating your detector relative to the GPS receiver", because then they get reasonable positive results out of their intended experiment. But if it seems like this is not the best possible experiment for determining the speed of neutrinos, it's because it wasn't supposed to be such an experiment.

[Wnderer]

I believe this statement in the paper is wrong.

The high-accuracy time-transfer GPS receiver allows to continuously monitor tiny
movements of the Earth’s crust, such as continental drift that shows up as a smooth variation of
less than 1 cm/year, and the detection of slightly larger effects due to earthquakes. The April
2009 earthquake in the region of LNGS, in particular, produced a sudden displacement of about 7
cm, as seen in Fig. 7. All mentioned effects are within the accuracy of the baseline determination.
Tidal effects are negligible as well.

That's not accuracy. That's precision. The ability to measure a quantity repeatably to very small amounts is precision. Accuracy refers to how closely a measured value comes to the true value. If the GPS receiver has some error in some constant than the distances would be off. The above statement gives no clue to how they verified the accuracy of the GPS. I'm not sure how dependent GPS is on our knowledge of the speed of light in the atmosphere or if they got some ratio-metric way of reducing that dependency. Maybe the world is 25ppm smaller than we think it is.

[gmalivuk]

Maybe the world is 25ppm smaller than we think it is.

This would be really surprising, honestly. Geodesy is a fairly mature science, and errors of 25ppm would be about an order of magnitude greater than the current accuracy of geoid models.

Yes, there's probably a systematic error in their measurements, but I really strongly doubt it's because we're wrong about how big Earth is.

[GuyOnTheInterweb]

My take on this:
1) Vacuum is not really vacuum, there is a bit of hydrogen in empty space, something like 1 atom per cubic meter.
2) We already know light slows down based on the index of refraction of the medium. That's how lenses work.
3) I'd interpret this to mean neutrinos really do travel at the maximum speed information can propagate.
4) We ought to start saying, "Faster than neutrinos!" instead of "Faster than light!"

1) I believe when light speed has been measured they have used "real" vacuum in the lab, not just bouncing about out in deep space. That vacuum tube would still include any virtual particles that in theory could interact with light. (I won't go into dark matter and dark energy here!)

Or could it actually be light slowed down by neutrinos? They do after all go through "everything" and are "everywhere" in space, so presumably also through those vacuum tubes. This should be easy enough to test, put some very long fibre optics in front of this neutrino beam they are making, and measure if the amount of neutrinos affect the light speed. (My bet is it does not).

The "extra dimensions" explanation (neutrinos did not go faster than light, but found some shortcut through those tiny extra dimensions of string theory etc) is interesting, but would require one of those to be a second time-dimension, and begs the question of why light (which until now has always taken the fastest path, even in QED) is *not* taking those shortcuts. Is it because the photons are massless?


Now my take on this whole story, after reading the rather impressive paper (http://static.arxiv.org/pdf/1109.4897.pdf) and watching the CERN presentation (http://cdsweb.cern.ch/record/1384486), is simply that something went wrong on calculating the distance, as it would need to be 20 meters further (out of about 731 km). This could either be something fundamentally wrong with GPS, or more likely, some systematic error in how they transposed the positions from the GPS marks over to the station. I think they did do this twice, even with accurate measurements of the geoid (which did vary throughout that tunnel!), but did not try any other control measurements. The British guy in the Q&A session suggests "Just drill a hole!" - this would certainly work.

I would also try to measure the positions of both sites using a non-GPS system, like Galileo or GLONASS, or try to not use any transformations to ETRF2000 geoids, just get the absolute positions relative to the satellites on both cases (ignoring the earth - as the neutrinos do). In theory you could download the raw GPS data on both sites at the same time (which should share view of many satellites) - they have already established a very high-accuracy time-link between the sites, and so you can then (online or offline) do the calculations of the exact distance between the two sites as if you were a massive GPS device.

[rhhardin]

Light plodding along, as attested by characteristic blurring
http://rhhardin.blogspot.com/2010/10/186000-miles-second.html

[CZeke]

Good points, weak strip. The premise is funny enough, but there's no punchline.

[K^2]

A tunneling particle can violate locality. Just take a plane wave solution under barrier, substitute in packet of your choice, and provided your choices don't cause too much packet distortion, you should see that time of transit can be less than for particle to propagate at speed of light.

If you can think of a better practical realization of this than neutrino passing through a mountain, let me know.

[ijuin]

The thing about tunneling is that it should be random for each particle, so therefore we should see a Gaussian distribution in the "excess speed" of the neutrinos--e.g. some would be 80 ns ahead and some would be 40 ns ahead. In other words, there should be some "blurring" of the arrival times. However, if the neutrinos were all arriving close to 60 ns early, with very little variation, then it would imply that a non-random effect is in play.

To toss in my own speculation, what are the relative altitudes of the neutrino source and detector? I am not speaking of altitude relative to the GPS satellites, but rather altitude relative to the geoid (the gravitational equi-potential surface http://en.wikipedia.org/wiki/Geoid ), and therefore their depth within Earth's gravity well. If, for example, the detector is at a lower altitude than the source, then general relativity dictates that time itself will pass slower for the detector than the source--and that therefore the detector will receive the neutrinos "early" compared to the transit time calculated from the source's frame of reference.

[K^2]

The thing about tunneling is that it should be random for each particle, so therefore we should see a Gaussian distribution in the "excess speed" of the neutrinos--e.g. some would be 80 ns ahead and some would be 40 ns ahead. In other words, there should be some "blurring" of the arrival times. However, if the neutrinos were all arriving close to 60 ns early, with very little variation, then it would imply that a non-random effect is in play.

But there is a distribution.

An early arrival time of CNGS muon neutrinos with respect to the one computed assuming the speed of light in vacuum of (60.7 ± 6.9 (stat.) ± 7.4 (sys.)) ns was measured.

So it's more like some arriving 70ns early, while others only 50ns early. The question is whether the packet distortion one expects from such a tunneling event is smaller than the distributions measured. You can't tell one way or another without running the numbers.

Part of the problem with actually running the numbers is that you'd have to make assumptions about neutrino mass and the effective interaction potential. That's two unknowns, with only one fitting parameter. Might not matter for packet distortion, though. With constant potential, the contributions from mass and interaction shouldn't really be any different, so all I should need is the propagator... I'll see if I can make the numbers work, but my brain is a little fuzzy right now.

[Wnderer]

Here is a coincidence. The distance was 730534.61 meters. The speed of the neutrinos was 1.0000248*(299792458)m/s so the time was 2.436740731ms. In that amount of time at the speed of light, the distance would be 730516.4932. The difference is 18.12 meters. According to this site.

http://public.web.cern.ch/public/en/Res ... GS-en.html

At the end of this decay tunnel, an 18 m thick block of graphite and metal absorbs the remaining protons, and pions and kaons that did not decay. The muons are stopped by the rock beyond. Only, the muon neutrinos remain to streak through the rock on their journey to Italy.

So there is an 18 meter block of granite and an 18 meter additional distance traveled.

[lgw]

Here is a coincidence. ...
So there is an 18 meter block of granite and an 18 meter additional distance traveled.

It would be particularly hilarious if this was all a measument error due to an intern getting confused about which side of the block of graphite he was recording the position of. Twice.

[K^2]

The granite block is located at the end of the decay tunnel. If it was on detector side, yeah, that would be something to look into. But it seems like they are measuring distance directly from decay detector (scintillator + photo multiplier?) to the neutrino detector. That would exclude possibility of making a mistake due to the granite block.

But I'm in theory. I'll try to track down some experimentalist in the dept to see what they think about it.

[philipquarles]

The Devil's Neutrinos

I liked their earlier albums better. You probably haven't heard those albums though, because they haven't come out yet.

It's clear that a lot of you (and almost certainly Mr. Munroe) have little to no experience with gambling. Allow a real degenerate to clue you in on a few things:
1. You cannot, cannot make bets with random strangers on the internet and expect them to pay up. Your expected value on a bet set up this way is less than 10% of what it would be if you used an escrow.
2. Most "normal" people really don't want to bet. This is especially true if you insist on posting the money (with a neutral party), agreeing to odds, and clearly establishing win conditions. People will happily make random posts with the word 'bet' in them and accept any money you want to send them as a result of those posts. Will they actually put up their own money to back up these posts? Not nearly as often.
3. It's much harder to set terms for bets than it would first seem. What if these results cannot be replicated by any other experiment, but the initial results also cannot be directly overturned (as with the Fleischmann–Pons experiment)? If you make a bet that requires the results to be clearly overturned for payment to be issued, how long will you wait? I would probably bet $200 that evidence from another experiment will be released that clearly calls these results into question within two years. But then you have to agree on an escrow that you can use for two years, and a judge to evaluate any new evidence. These things are hard to come by, and even harder to find for the minimal expenditure of time and money that would make a series of $200 bets feasible.

Believe me, if this technique worked, I would be doing it. I won about $350 betting that an airplane on mythbusters would be able to take off from a conveyer belt. Unfortunately, these opportunities are few are far between, and I doubt this neutrino data will present any profitable opportunities.

[holshy]

Strictly speaking, the CERN experiment does not by itself violate special relativity. Relativity permits particles that always travel faster than light. What is forbidden is for an object to be capable of going both below and above the speed of light (c). Particles that always move (relative to us) faster than light are called tachyons, and people have been looking for them, but most physicists were not really expecting to find one. If neutrinos turn out to be tachyons, it will be one of the great "right-under-our-noses" events in particle physics history.

There are problems with tachyons. They allow you to communicate, in principle at least, instantaneously with anyone else moving below c. This fact combined with special relativity can cause major problems with causality. You may determine that some kind of event always "causes" a different event some distance away. For it to really be the cause of the other event, you want it to be earlier than the other in all reference frames. When tachyons are let into the picture, you can no longer do that. The result must precede the cause in some frames if tachyons are the intermediaries over that distance in that time interval. So strictly speaking, relativity and tachyons together violate causality, and we don't really want that to happen.

Major props for being the first person to bring this up.

The upshot is that you can pick 2 of the following to be true, but not all 3

• Relativity (as Einstein imagined it)
• Causality (as the Greeks imagined it)
• FTL travel (as... Star Trek... imagined it )

A lot of the tubes seem really focused on "this would disprove relativity". They forget that it could, alternately, disprove the common convention of causality. However, I'm still betting on an instrumentation error and FTL being the odd man out.

[tijis421]

http://jtotheizzoe.tumblr.com/post/11322648997/gravity-may-have-thrown-off-faster-than-light

Yeah, so... Someone owes Randall money. lol

[nothinglikeit22]

all this talk of gps's... geocaching anyone?

[BJReplay]

http://arxiv.org/PS_cache/arxiv/pdf/1110/1110.2685v3.pdf

Lorentz contraction and detector movement.

Start spending those bets you collected, you won't need to pay out.

[thegretstar]

Psh. Continental drift? For such a smart guy, Randall really messed his one up!!
Thank you, i'm so glad i wasn't the only one who noticed this.

As a geologist, the tooltip for this comic (and the figure caption in the paper) really irked me.

GPS stations are used all the time in geology/geodesy. They have been used to measure gross tectonic plate motion and local strain accumulation along plate margins and major faults, among many other things. However, they do not measure "continental drift"! Continental drift was a theory of Earth evolution in which the continents "plowed" through the oceanic plates. This theory was postulated by Alfred Wegener in the 1910s. During the 1960's, the theory of Plate Tectonics was formed and it continues to be the accepted theory for the crustal evolution of Earth. Plates don't drift, they move, subduct (dive under each other), and collide as a collective whole.

[steve waterman]

I did never get why everyone thinks that relativity HAS to be right anyway. After all, it makes two assumptions that are AFAIK impossible to prove: That the speed of light in vacuum is constant and that the laws of physics work the same in every frame of reference.
For the first:

c=\sqrt{{1}\over{\epsilon_0\cdot\mu_0}}

but who can be sure that the fundamental constants of nature are really constant and that \epsilon_0 and \mu_0 have always had the same values?
And for the second: Isn't that even a contradiction to the multiverse interpretation of quantum physics, where the same thing can have different outcomes?

DISCLAIMER: I'm not a physicist.
I'm not even a scientist.

I decided to take down my thoughts regarding fundamental constants here...and only have them at my site.

[steve waterman]

Randall...I would like to support you in your cause here at xkcd personally.

That is, I would like to take up your bet, however 200 dollars is too steep
for me at this time and i would offer a wager of 117 dollars.

"pointless arguments about Galileo"...
how apt..."noting too, math has a point and physics is pointless, is how I envision it."

Prior to just now seeing this particular comic, and only being a member since my 64th
birthday on this November 14th, when you wrote 0955 "map projections "
and it included my waterman map, I mentioned
that i had others avenues of interest, like Physics.
Randall...is this 117 Canadian dollars challenge acceptable to you ?

Will I be allowed to present my math logic here ?
Or we could exchange, outside this forum, bu another means.
I would do it differently, than appears on my site, btw...

Noting, It can be seen on my site..in a mathematical form...from here
http://watermanpolyhedron.com/june2011VERSION.HTML

it is short enough to read through, and has some pictures, and a first read through
might take all of 5 minutes. Hopefully, you will able to walk yourself through it,
with a second read...'okay, lets take it line by line and see if I can follow each line and agree
or disagree with that singular line."

[scarletmanuka]

... assorted physics-like gibberish...

Mr Waterman, I believe you should stick to cartography.

[Eebster the Great]

What if the permittivity of a vacuum was the isoperimetric quotient of a sphere, 1/(36 pi)...and not an approximation through measurement ?
Then the speed of light would, by using your equation, since the permeability of a vacuum is 4 pi...turns out to be exactly 3, and not their agreed-to-by-committee-decree of an "exact value" of 2.99792458 !

wat

The speed of light was not "decided" by a committee, it was fixed. The units are defined based on the speed of light, not vice-versa. And the value is 299 792 458 m/s, not 2.997 924 58 in any useful units.

So I stopped reading after this.

[Starlight Sonata]

I hadn't thought about this comic in a while...so imagine my surprise when I opened up my December's copy of Physics Today and saw, right there on pg. 8 under an article on superluminal neutrinos...this comic.

Randall Munroe, I wish I was you.

[Eebster the Great]

I hadn't thought about this comic in a while...so imagine my surprise when I opened up my December's copy of Physics Today and saw, right there on pg. 8 under an article on superluminal neutrinos...this comic.

Randall Munroe, I wish I was you.

Is it any surprise the author of an article in Physics Today reads xkcd?

[repeatingdecimal]

What do a neutrino and I have in common?

Spoiler:

We're both constantly penetrating your mom.

[mdziel]

So, Randall, tell us how much you made this time

http://news.sciencemag.org/scienceinsider/2012/02/breaking-news-error-undoes-faster.html

[steve waterman]

i felt a connection to the "neutrinos" comic and the Galilean transformations

Original poems were inspired by two of the xkcd comics...
"chess photo"
"a bunch of rocks"
http://www.watermanpolyhedron.com/galarticles.html

the corresponding math logic and related diagrams can be seen on this page
http://www.watermanpolyhedron.com/june2011VERSION.HTML
[ give it a minute to load up please ]

Another poem, more related to time and light, was inspired by Randall's
"purity" (from same page as the above two poems )

In essence, the Galilean transformation ignores that coordinates remain FOREVER fixed wrt to their own system.
This certainly is not obvious, when looking at their simple equation x' = x- vt.

Imagine two coincident Cartesian systems, having point P at (2,0,0) in A and point Q at ( 2,0,0) in B....
and we relocate B so that it is now not coincident with A...
if we transform point P in A to system B...is that new point called point P in B
...or called point Q in B ( as the Galilean would suggest ) ?

Anyone who knows how a Cartesian system works....got an opinion/response to the simple direct question above ?

[deocar]

just saw #955 for the first time.
sure, they now claim there's a perfectly normal explanation for the "false" ftl travel, but i think the $200 has been lost. how else can you explain this (on the same day as i first saw the ftl story)?
http://www.bbc.co.uk/news/world-europe-15032614
to make it more clear how this is simply the backside effect of those neutrinos, i've made a handy map:
http://maps.google.com/maps/ms?msid=211962374319432562114.0004ada3b4b9b5d7a8053&msa=0&ll=47.901614,2.285156&spn=17.75026,40.78125

[mfb]

For others:
BBC-Link
Google maps link

While the map is funny, the events are actually not in one line, not even approximately - consider the 3rd dimension .

[deocar]

"consider the 3d dimension"!?

surely not. to any reasonable approximation, the 3d dimension is inconsequential. See: http://xkcd.com/848/ i presume the effect would only be magnified when, from the point of view of the neutrinos, the energies involved in the rest of the system are so small. should be obvious - after all, the man did burn. qed and whatnot.

[J Thomas]

For others:
BBC-Link
Google maps link

While the map is funny, the events are actually not in one line, not even approximately - consider the 3rd dimension .

Perhaps it would be interesting to see whether other anomalous events have come up elsewhere on that great circle.

[gmalivuk]

And remember that it only counts if deocar happened to read about it the same day as the ftl "news".

[mfb]

"consider the 3d dimension"!?

surely not. to any reasonable approximation, the 3d dimension is inconsequential. See: http://xkcd.com/848/ i presume the effect would only be magnified when, from the point of view of the neutrinos, the energies involved in the rest of the system are so small. should be obvious - after all, the man did burn. qed and whatnot.

If you draw a line from the CNGS detector to CERN and beyond, it would go over ireland at a height of ~100km. That is a back-of-the-envelope calculation, but I think the value is between 50 and 200km.

[J Thomas]

"consider the 3d dimension"!?

surely not. to any reasonable approximation, the 3d dimension is inconsequential. See: http://xkcd.com/848/ i presume the effect would only be magnified when, from the point of view of the neutrinos, the energies involved in the rest of the system are so small. should be obvious - after all, the man did burn. qed and whatnot.

If you draw a line from the CNGS detector to CERN and beyond, it would go over ireland at a height of ~100km. That is a back-of-the-envelope calculation, but I think the value is between 50 and 200km.

Why would you expect strange effects on the opposite side of the detector? Is this one of those quantum mechanics arguments that says if you detect neutrinos in one direction, then something-or-other must have gotten shot off in the opposite direction?

I can imagine an argument like that being plausible, but I can also imagine arguments that make so little sense that I'd just as soon assume that strange things happen on the plane formed by CERN and the detector and the center of the earth. I can't make a good argument for it, but I can imagine arguments for the backward straight line that wouldn't be any better than a vague handwaving argument for that plane.

[mfb]

No, you don't expect anything there. It was just a joke, and even that works in two dimensions only.

[The Moomin]

No, you don't expect anything there. It was just a joke, and even that works in two dimensions only.

The three dimensional argument has fallen flat?

[jpers36]

No, you don't expect anything there. It was just a joke, and even that works in two dimensions only.

The three dimensional argument has fallen flat?

There was no depth to it.