Hey Sho, Maurik,

I wrote this code so I'll look into it. But I'm really busy now so I might only have time in the next days or so. Hope there is a way around it for now (skip event?).

/Pelle


On May 13, 2014, at 1:14 PM, Sho Uemura wrote:

HPS2014OfflineTruthRecon.lcsim is not appropriate for data produced with 
HPS2014ReadoutNoPileup.lcsim; I've made a new steering file that removes 
all the SVT time cuts, and this is what should be used.

This resolves the problem you were seeing with low track efficiency, but 
the "t" crash is still there, so recon crashes on the third event. Ball's 
in Pelle's court.

On Tue, 13 May 2014, Hansson Adrian, Per Ola wrote:

I think this code can behave badly in special cases. I think it's used to calculate, generally, an intercept point in 3D space between a helix and an arbitrary plane. I had issues when the track momentum was too low which I presume meant that the track actually never intercepted the plane anywhere. My memory fails to tell me if this is the error I got. I think I fixed that problem in the MS code by checking analytically if the helix would intercept the plane and only if that was true use this (iterative) method to find the true intercept point. This might be an edge case of some kind in that process but I'm a little surprised that no one have seen it before.

Can you send more detailed info on the sample, particles, etc.

Thanks,

Pelle


On May 13, 2014, at 12:41 PM, Sho Uemura wrote:

Can look. Can you make available the LCIO file output by the readout
simulation?

On Tue, 13 May 2014, Chad Civello wrote:

Hello All,

I am wondering if anyone is familiar with a reconstruction error that I am
running into. When attempting to run reconstruction with the
HPS2014OfflineTruthRecon.lcsim steering file, the process quits with an
error caused by finding a negative value of the variable "t". Upon further
investigation, it seems that a negative "t" will cause an imaginary path
length of the track, calculated in WTrack.java. The physics behind
calculating t is beyond me, involving combinations of a momentum vector, a
magnetic field unit vector, and a unit vector of a plane. I am able to step
through 590 events before this error occurs, so it seems to be something
with this individual event that is stopping reconstruction from continuing.

The file being run through reconstruction is created from running an event
file in SLIC with the HPS-Proposal2014-v5-6pt6 detector, and then passing
this file through readout simulation with the HPS2014ReadoutNoPileup.lcsim.
The output error is listed below.

As a test on the steering file and drivers, I ran a smaller input file
through the same reconstruction without an error. However, out of 165
triggered events, I am only seeing 14 tracks. Is there a reason that so few
tracks would be found? Thank you for the help.


Regards,

Chad Civello
University of New Hampshire

Event 585
Event 586
Event 587
Event 588
Event 589
Event 590

getPathLengthToPlaneApprox ERROR t is negative (-0.09475479432494227)!
p 0.30696444517303073 rho -0.0014649536585467351 a -4.4968868700000003E-4
A -3.2568853514089034E-4 B 0.8943517247214046 C -686.7142264081292
track params: 0.2785713187164182, -0.12804764158077803,
-0.01513249730159967, 0.30696444517303073, 16.49545255281126,
35.88633038244529, 0.656055763097405,
xp [      702.45,      71.390,     -27.331]
eta [     0.99953,    0.030515,      0.0000]
h [      0.0000,      0.0000,     -1.0000]

########################################################################
Use REPLY-ALL to reply to list

To unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1

########################################################################
Use REPLY-ALL to reply to list

To unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1

########################################################################
Use REPLY-ALL to reply to list

To unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1

<HPS2014OfflineNoPileupRecon.lcsim>

########################################################################
Use REPLY-ALL to reply to list

To unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1