Hello,
On Mon, 19 Jun 2000, Flemming Videbaek wrote:
>With real data gets lots of warnings about dx == 0. Have to deal with this
>needs comments/questions
>why does this arise?
>the Warning method is in this case from TObject and do in contrast to the
>one from BrModule NOT have a limit count and thus continuies ad infinitum.
>This is no good either? though very relevant. The solution is to find why
>this appears.
I think at least part of the problem has been that BrTPCClusterFinder
so far has not discarded one-pad clusters arising from remains of cluster
deconvolution. These will have zero width and position uncertainty (dx).
The problem has been observed from time to time in simulated data as well.
I have added a (temporary) fix in BrTPCClusterFinder::DeconvoluteClusters()
so that a subcluster is only added to the list if it has status different
from BrTPCCluster::kNoiseHit. This seems to eliminate all the error
messages at least for a few sample "real data" MTPC1 events that used to
give lots of warnings.
Best regards,
Trine
This archive was generated by hypermail 2b29 : Thu Jun 22 2000 - 08:36:40 EDT