Page 45 of 76

Re: Several biggest errors of particle physicists.

Posted: Sat Apr 03, 2010 7:36 am
by Xymox
Na... Everything was fine, just really really curvy lines..

Its unexplained.. We are not privy to the science data...

I just figured Ivan might enjoy knowing about these events...

Re: Several biggest errors of particle physicists.

Posted: Sat Apr 03, 2010 7:55 am
by Mailo
Actually, the explanation is very simple. No particle detector actually measures tracks of the particles. They all consist of layers which have a certain probability to detect the passage of a particle through the layer (called a "hit").
In LHCb, there are no layers inside the magnet (the white block slightly left of the middle of the image), only before and after. The detected passages of particles are marked by purple and green x.
After an event is measured, a software routine then runs over these "hits" and tries to assign them to particles passing through the detector. In this case, that software routine obviously failed and needs to be improved.
Granted, this assignment is rather difficult if many particles were present. Not all of them leave hits in all layers, so your software needs to take into account these "missing" hits, the angle the particles change direction in the magnet depends on the yet unknown momentum, etc.
If you look at the first image you showed, you can see the software failing on the lowest purple "x" on the right side of the magnet. A straight line would have fit just fine through these points, yet the software insists on putting these wiggly lines through them (and remember, there is no information at all about the particle position between these purple x, so it is all fantasy).

Stephen, you need to develop a thicker skin concerning this. There is simply no way a magnetic field as strong as necessary for this could suddenly appear out of nowhere. On the other hand, there ARE people out there who will stop at nothing to convince others to believe their BS.

Is it a bad omen that my first post on this forum is in this part of it? :mrgreen:

Re: Several biggest errors of particle physicists.

Posted: Sat Apr 03, 2010 9:06 am
by Kasuha
I guess it's rather some problem in the interpolation software. Most of these screwed lines are interpolating no more than three dots on almost straight track. The dots are what the machine actually detected.

Re: Several biggest errors of particle physicists.

Posted: Sat Apr 03, 2010 9:42 am
by Stephen
Good, so it seems like these images are nothing more than errors in the software.

Re: Several biggest errors of particle physicists.

Posted: Sat Apr 03, 2010 10:14 am
by Anitusar
Yes, its the Software. Remember, all lines in this pictures a reconstructed predictions by the Software. A lot of parameters enter the reconstruction process and sometimes the results go rogue. What actually happened are the measurements presented as x and bars in the calos. From these you try to determen, what happens and this gets more complicated the more energy/measurements are involved.

Re: Several biggest errors of particle physicists.

Posted: Sat Apr 03, 2010 11:46 am
by spencer
i think the software may be running as planned.

This looks likes a high order regression line drawn through a few data points, with a missing data point near the front end of the detector.

i would first look for a hardware failure/error.

Re: Several biggest errors of particle physicists.

Posted: Sat Apr 03, 2010 12:06 pm
by Anitusar
spencer wrote:i think the software may be running as planned.

This looks likes a high order regression line drawn through a few data points, with a missing data point near the front end of the detector.

i would first look for a hardware failure/error.
No, you would not see Hardware failures in these pictures (only complete detectors missing). These weird lines probably are just very bad matched tracklets. Multiple turns in a track would mean very strong multiple scattering, which can not be reconstructed by the software (it would increase the combinatorics to the point where no reconstruction is possible).

The software furthermore does only fit parabulas, as higher order fits would not make sense.

Re: Several biggest errors of particle physicists.

Posted: Sat Apr 03, 2010 12:38 pm
by spencer
Anitusar,

Thank you for the additional info and keeping this discussion in line.

Much appreciated.

Re: Several biggest errors of particle physicists.

Posted: Sat Apr 03, 2010 1:45 pm
by Allan
Hi. Has anyone else noticed that RICH1 and RICH2 are not reporting any hits as they were earlier?

I wonder if the lack of information these detectors could be partial to blame for strange results.

Allan

Re: Several biggest errors of particle physicists.

Posted: Sat Apr 03, 2010 2:24 pm
by Anitusar
Allan wrote:Hi. Has anyone else noticed that RICH1 and RICH2 are not reporting any hits as they were earlier?

I wonder if the lack of information these detectors could be partial to blame for strange results.

Allan
The Rich1/2 do not play part in the track finding algorithmus, so they can not be the source for any weird tracks. The Rich detectors are important for particle identification, which takes place after the tracks are reconstructed (You need a track pointing to the Rich information before you can merge these two measurements).

But indeed, there are no Rich measurements present, hmmm ?

Re: Several biggest errors of particle physicists.

Posted: Sat Apr 03, 2010 2:25 pm
by Allan
Hi Here is quotation from the documentation on the RICH detectors;

"The two RICH detectors are responisible for identifying a range of different particles that result from the decay of B mesons, including pions, kaons and protons."

Without have this basic information about a collision, the software is probably guessing about tracks rather then actually calculating them.

Allan

Re: Several biggest errors of particle physicists.

Posted: Sat Apr 03, 2010 2:30 pm
by Anitusar
The flow is

measurements -> track seeds -> tracks (this is what you see in the online displays) -> particles (add particle info) -> decay chains (combine particles)

Re: Several biggest errors of particle physicists.

Posted: Sat Apr 03, 2010 2:56 pm
by CharmQuark
It's nice to have you around again Anitusar :D this makes a very happy CharmQuark :dance:

Re: Several biggest errors of particle physicists.

Posted: Sat Apr 03, 2010 3:21 pm
by Allan
Hi

A quotation from LHCb log; "HCAL is back with new timing recipe."

Anyone have any ideas what this means?

Are they about to make brownies or something with a new recipe? :rolleyes:

Allan

Re: Several biggest errors of particle physicists.

Posted: Sat Apr 03, 2010 3:29 pm
by Anitusar
:icon-idea: A recipe is a certain set of settings for the frontend electronic. A timing recipe is therefore the relative timing adjustments between the electronics within a subdetector and the global detector. :icon-idea: