[Halld-cal] NIM article for FCAL beam test

[Kei Moriya]

Hi Elton,

Thanks for the detailed comments. Before I can revise
a new draft, I had some questions regarding your comments.
I hope to have a revised draft soon after I have all of the
necessary information at hand.

> It is worth perhaps referencing a review article on calorimetry for a
> context of lead glass detectors also.

Would you know a good review article on calorimetry using lead glass?

 > p. 1. sec 2.1 first paragraph
> p. 1 right column middle, needs rework
> magnetic shield, and acrylic light guides have been installed to encase ?
> optical guides? Perhaps simply state that a small air gap was replaced
> with an acrylic cylinder to improve light collection and allow for
> better magnetic shielding. [Should perhaps mention the expected maximum
> field at the pmts, was it 150 G?]

I think the wording was confusing so will fix this. I believe
the magnetic field strength is an estimate, do you know if there is
any detailed value I can use?

Beyond the wording, is there something you do not agree with?

 > p. 1. right column
> further down
> may wish to "beam bunch" instead of bucket
> photon rate: 10^8/s is the photon rate in the coherent peak. [The photon
> rate is orders of magnitude higher]
> The use of the timing to identify the beam bunch is desirable, but is
> not a requirement. In fact, it is difficult to argue that the bunches
> can be identified cleanly at low energies.

I fixed the text to say coherent peak rate is 10^8. For the latter part,
I'm not what your comment here is trying to get to. Do you want me to
fix something?

> p 2 Figure 3
>
> The blue-gray area of the trigger counter scintillators does not appear
> to be an array of counters. However, it does nicely show the outline of
> the prototype. I suggest trying to make both visible.

The two arrays overlap almost completely, so it is difficult
to separate them. Do you have any ideas of how this could be done?

> I would emphasize early on that the test was performed in parasitic
> mode, so we did not have control over beam parameters, such as beam
> energy or intensity. Early on, you should specify the running conditions
> or periods (e.g. beam energy and beam current). Also important is the
> radiator thickness (50 microns?)

I've asked Andy Sandorfi about this, hopefully there will be a reply soon.

> p 4
> Figure 6, questions
> What happened to the fADC top left (replaced by remote paddle)
> Why are the baselines so high on fADCs 1,4 and 2,5?
> Why do most of the channels show zero (should they not have a pedestal
> value?)

The fADC at the top was replaced by mistake, this is corrected now.
The pedestals in each channel were always somewhere between 100-200 ADC
counts. Most channels show 0 because this run was taken in raw threshold
mode, where only channels that had ADC heights above a certain
pre-determined value were read out. The top two rows did not have this
set correctly, and some channels are read out regardless of signal.

> p 4 left, second paragraph
>
> ...8 samples around p?  the pedestal?

Yes, p is defined as the pedestal a few paragraphs above.
The paragraph starts "The cut-off applied...".

> p 8 Figure 12.
> Why are there such large offsets between blocks, up to about 5 ns? Since
> they are looking at the same shower they should be at the same time. I
> guess there could be different delays in the pmt transit time.

The timing offsets are rather large, and as you point out up to 5 ns.
The values for specific combinations of blocks do not depend on Sp,
and are constant. This may be due to PMT transit time, and also
may be due to different cable lengths or delays (we did not do a check
of the timing of signals when building the setup).

> p 9 right, below eq 2
> "Figure 14 shows AN EXAMPLE OF the resolution of a single module." (If I
> understand what you mean)

No, the resolution for a single module can be determined up to 4 times,
using different modules it is clustered with. The results we show are
then the weighted average of those measurements.

> p. 9 right discussion on units. I think you should make the conversion
> to energy, and quote a in units of ns/GeV. However, it is also nice to
> have the number in terms of ns/mVm, which gives the size of the signal.

For the timing resolution, we can't convert to energy since each block
gets illuminated by different energies for each event. The run used
for this analysis had ~1275 MeV electrons coming in, but for each
event we do not know how much energy was deposited in each block.

> p. 9 References.
> Earlier I listed the E852 and Radphi papers.
> You should also reference the JLab fADC. I don't have a handy reference,
> but make the citation and we can find a good one.

Yes, if you find a good reference let me know.
I know there is this conference proceedings:
http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=4436457&contentType=Conference+Publications&refinements%3D4294967269%2C4291944822%26matchBoolean%3Dtrue%26searchField%3DSearch_All%26queryText%3D%28p_Title%3AFlash+ADCs%29
H. Dong et al., "Integrated tests of a high speed VXS switch card and 
250 MSPS flash ADCs" (2007)
I've asked Fernando if there is a newer reference.

Thanks again for the comments.
Cheers,
	Kei