Re: Understanding HDGeant Hits (fwd)

[Elton Smith <elton@jlab.org>]

---------- Forwarded message ----------
Date: Mon, 29 Oct 2001 15:34:43 -0500
From: Richard Jones <Richard.T.Jones@uconn.edu>
Reply-To: jonesrt@uconnvm.uconn.edu
To: Edward J. Brash <brash@uregina.ca>, hdgeant@zeus.phys.uconn.edu
Subject: Re: Understanding HDGeant Hits

Hi Ed and all,

Since these questions are bound to come up again, I have posted it
to the hdgeant list.

> I am trying to understand the output from the HDGeant simulation.  Most of it
> makes sense, but with regards to the BC, I am a little confused.  There are hit
> entries for individual phi modules, with energies and times for both the
> upstream and downstream ends of the module, which makes sense.  These are
> labelled with a <shower> tag in xml.  Then, after this information is presented,
> there are some entries like:
>
>         <barrelShower E="0.0143335" t="2.787" z="25.3923" phi="0.442111" />
>         <barrelShower E="0.172093" t="28.2692" z="121.109" phi="0.917504" />
>         <barrelShower E="0.375096" t="2.97893" z="118.624" phi="-1.13804" />
>
> Can you please explain the rationale behind this information?  Is it some
> compilation of the previous information?

There are two kinds of tags defined for each subsystem: detector hits (what you
would see in real events after calibration has been applied) and Monte Carlo
truth tags like <barrelShower...> and <fdcPoint...> which say where the track
actually went through (3D point) what the energy was before attenuation, etc.
The latter tags would not be present in real events, and eventually have to be
ignored by analysis code.  They may prove to be useful in developing analysis
code and measuring its performance.  That is why I put them in.

> Also, how are the times defined?  That is, when is t=0?  I had thought initially
> it was at the event generation time, but this cannot be right, since many of the
> shower entries have negative times.  Unless this is a resolution effect?

You are right that t=0 is the instant of the photoreaction in the target.
So then, how do you get negative values of t in the BCal?  The answer is in
the definition of how the time delay for propagation inside the fibres gets
included in the final t for a BCal hit.  My decision was to define the time
offset in the BCal so that [t(upstream)+t(downstream)]/2 is the ToF for the
particle, independent of where it hits the counter.  If the particle hits
right in the middle of the counter then 0 propagation delay gets added to
each end.  If it hits at theta=70deg then the upstream time gets a negative
-delta_t and the downstream gets a +delta_t for propagation.

Although it looks surprising at first, it is the simplest convention for
2-ended ToF counter alignment.  Because of arbitrary delays in readout channels,
the only way to get meaningful time scales is to decide on some convention and
then use calibration to remove the spurious offsets.  My suggestion is to use
a known ToF (say for a photon shower) to fix [tup + tdown]/2 and a known effective
speed of light in the fiber to fix [tup - tdown]/2.

Finally, this question is up to your group and I.U. group working on the ToF.
If you want me to add L/2c (or any other constant) to all of the hit times then
I will do that.


Richard Jones