<html>
<head>
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">Matt, I have nothing against referring
to Radphi in this article! If we are going back and rebuilding
something, why not do it better? I just want to get the story
right. Maybe the misunderstanding is all mine, that will be up to
you to decide.<br>
<br>
Thank you for reminding me of this technical note 850. In
reference #7 that is cited in the draft paper, which is reference
#1 in doc 850, it states (Optical Coupling Study section):<br>
<blockquote>Two calibrations were <br>
performed, one with an air gap of order 1 mm between <br>
the lead glass and photomultipliers, and one with optical <br>
grease. For each condition the electron energy peak was <br>
calibrated to peak at 5 GeV. and a Gaussian fit done to <br>
give an estimate of the total energy resolution. The <br>
quantity sigma/E was found to be 0.05 with and without <br>
optical grease. A significant increase in the number of<br>
photoelectrons collected would have decreased the stati- <br>
stical term in the shower energy resolution. narrowing <br>
the ratio sigma/E. No such narrowing was observed. <br>
</blockquote>
Is it your view that the stated conclusion is not warranted by the
observations? A 5 GeV shower might not be statistics dominated.
If not, no observable narrowing might occur as a result of an
increase of photostatistics by a factor of ~3.<br>
<br>
I have no problem that a favorable comparison is made with Radphi,
provided that the factors contributing to the improvement are
properly identified. What I see claimed in this paper is that the
statistical term improvement from 7.3%/sqrt(E) to 5.x%/sqrt(E) is
entirely explained by the insertion of the optical coupling
between the block and the phototube. I am wondering outloud
whether is it really plausible that the difference comes entirely
from photostatistics. Keep in mind that the E852 calorimeter in
test beam -- that had the same glass-air-glass coupling scheme as
Radphi did -- reported 6.0%/sqrt(E) for the statistical term, (see
footnote 19 in the Crittenden et.al. article and the associated
reference). If you took that at face value, wouldn't you conclude
that about half of the improvement comes from the light coupling
scheme and half from other sources?<br>
<br>
-Richard Jones<br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
On 4/9/2013 11:45 AM, Matthew Shepherd wrote:<br>
</div>
<blockquote cite="mid:22DB8969-3426-41AC-A9D7-4B04D220EF7A@indiana.edu" type="cite">
<pre wrap="">
Richard,
We attempted to study this in detail several years ago as we were designing the light guides.
You may examine:
GlueX-doc 850
Table 1 is the table of interest. There we indicate that a cylindrical light guide and cookie provides a factor of three improvement in photon collection efficiency over the air gap used in E852 and RadPhi. When we wrote this, we consulted you on what the contribution of photo-statistics was to 7.3% RadPhi statistical term and then tried to assume that we just improve that component of uncertainty.
In my opinion we could remove all direct comparison with RadPhi from the NIM article and simply state that we are attempting to demonstrate that we can achieve the GlueX design resolution (which is really based on experience from E852 and RadPhi, but we don't need to make the connection so explicit in the paper).
Matt
</pre>
</blockquote>
<br>
</body>
</html>