Pages

Thursday, February 23, 2012

It was a loose cable...

So the rumor is that the whole faster than light neutrino result was actually due to a loose cable. While there may have been other issues with the result, the article that I linked to above claims that it was due to "a bad connection between a fiber optic cable that connects to the GPS receiver used to correct the timing of the neutrinos' flight and an electronic card in a computer." Apparently there was a delay in the equipment that was unaccounted for that equaled roughly 60 ns.

Having dealt with high precision time measurements before I know how difficult this can be. All electronic equipment has a delay and if you don't calculate it right your experiment will be off. It could also be that they had the wrong length of fiber (hey it happens!), but to get a delay that long they would have to have about 18m of extra fiber (again, it could happen). So I am banking on it being an electronics problem that caused the delay. So if this proves to be the case .... wow. Cautionary tale.


18 comments:

  1. Yeah, I was pretty sure there was a mistake though I was secretly hoping there wasn't. Having a faster then light particle definitely would make life more interesting. :)  

    ReplyDelete
  2. I, for one, am a fan of better publicizing scientific results, but this may have been one case where it might have been better not to call the press conference.  This just makes CERN/OPERA look silly.

    ReplyDelete
  3. I actually just saw the article and had the same thought.  It's amazing that since last Fall, we've seen several theory papers explaining why this should be expected, and several more experimental papers explaining some error or another that should have been accounted for, but in the end, it was just a loose cable.  Of course, as the article said, "New data,
    however, will be needed to confirm this hypothesis."  All in all, very crazy.  

    ReplyDelete
  4. "It's amazing that since last Fall, we've seen several theory papers explaining why this should be expected"

    OH MY GOSH BILL!!! You have no idea how bad it is.  Most particle theorists I know (and I am not trying to slam them as I have tried to work with them and have a particle paper myself) spend a significant amount of time trying to explain any anomaly as if "we should expect this anomaly because..." no matter how screwy it is.

    Why, because if it turns out to be true everyone is going to cite them and say they are genius even though, like a paid attorney, their genius amounts to crafting any argument that makes the case convincingly whether reality agrees or not.

    One of my good particle friends graduated with his most highly cited paper being one of these "we should expect this anomaly is true" papers and guess what... as soon as he this anomaly became the next example of why half of all 3-sigma detections are false.

    Anyways, I am sure this behavior happens outside of particle theory, but at least one this is certain: it's alive and well in particle theory!

    ReplyDelete
  5. Interesting. Darn loose cables! I suppose this is a good reason for checking and rechecking equipment before opening your mouth.

    ReplyDelete
  6. Kind of sounds like the economists that "predicted" the housing crisis. Given a large enough number of people saying anything and everything, one of them is bound to be right. Now if they can just replicate their results...

    ReplyDelete
  7.  This often happens in astrophysics as well.  Any sufficiently odd observation will have half a dozen explanations from theorists within a couple months.  From the theorists' point of view, getting one wrong is usually forgotten but getting one right can get you a tenure-track position.  Of course if any one person does this too much it hurts their credibility, but the rewards encourage it to an extent.

    ReplyDelete
  8.  To be fair, I'm sure they did recheck everything they could think of and then checked it all again before they made the announcement.  60 nanoseconds is a ridiculously small margin and there are literally thousands - if not millions - of things that could have caused it.

    ReplyDelete
  9. JS,

    How long before the announcement that trillion data points were analyzed by a software that had a design error, so no Higgs at this time, but may be after 100 trillion measurements...

    ReplyDelete
  10. They are struggling to keep the funding.

    ReplyDelete
  11. Reminds me of Alice in Wonderland.

    ReplyDelete
  12. This is an insult to economists!  They readily admit that economics is a dismal science and they sometimes practice voodoo!

    ReplyDelete
  13. NN,

    I recall there were statements that they checked everything many a times including the instruments, and I hope the cable was part of the instrumentss.

    ReplyDelete
  14. When I was in experimental research, my rule was that all results must be reproducible after dismentling the set up and reset-up by another team member(s).  Internal double check.  If the second team could not reproduce, the first team was allowed to troubleshoot where the second team may have gone wrong. 

    ReplyDelete
  15. Final comment: loose cables wiggle, and, could easily add or substract many nanoseconds, and fairly randomly.  

    ReplyDelete
  16. Haha!  I love your summary of the situation.  My comment was, admittedly, a gross understatement.  I guess I was mostly referring to the few papers on this subject I've gone through in any detail.  Anyways, it'll be interesting to see where everything lies once the dust has cleared. 

    ReplyDelete
  17. That would be a very hard rule of thumb to follow with something like OPERA or CERN.  For a table-top experiment, that might be a good idea, but for anything of extreme cost or complexity, that doesn't seem realistic.  (I can just see the budget proposal.  "You've spent $10 billion and 20 years and finally found the Higgs boson.  Now, before you even tell anyone, you want to spend another $10 billion and 20 years taking the whole thing apart just to re-build it again to make sure you were right?!?!?") 

    ReplyDelete
  18. In science, all results must be reproducible independently.  We had tabletop fusion that was not or might have been duplicated depending on whose paper you may read.  Even for CERN, independently reproducible rule applies, however difficult it may be.  It is about science, not belief.

    ReplyDelete

To add a link to text:
<a href="URL">Text</a>