Email from Mike:

From jmichaelfinn@cox.net Thu Sep 23 11:32:16 2004
Date: Mon, 20 Sep 2004 02:23:25 -0400
From: John M. Finn 
Reply-To: finn@physics.wm.edu
To: carlini@jlab.org, 'Jim Birchall' 
Cc: 'Juliette Mammei' , 'Klaus Grimm' ,
     'Mark Pitt' , 'Neven Simicevic' ,
     'Greg Smith' , 'Allena Opper' ,
     'Tony Forest' , 'Norman Morgan' ,
     'Mike Finn' , 'David Armstrong' ,
     'Yongguang Liang' ,
     'Shelley Page' , 'Dave Mack' ,
     'Richard Jones' 
Subject: Some comments on the optics

    Hi,
    I have spent most of the weekend writing an event generator for Klaus to
put into his GEANT 4 simulation. But I also have spent some time thinking
about the spectrometer optics, waving my arms in the air, and doing a good
deal of funny looking right hand rules calculations to the amusement of my
daughter. I have finally caught up with my email and have looked at your
generated results with interest. Thanks to all for your hard work.
    Let me summarize what I believe lies at the root of our problem: The
collimator issue was a distraction, in retrospect, but not a fundamental
problem, and one that is well on its way to being resolved. However, we took
an unexpected 25% hit in rate when we had to reduce the phi acceptance due
to transverse defocusing effects. And, maybe not everyone was fully aware of
how quickly radiative losses can add up. Not having radiative losses in the
calculation earlier gave somewhat of a false picture of our situation.
Attempts to recover from these rate losses by opening up the theta
acceptance were not entirely sufficient to compensate. The recent studies of
the inelastic backgrounds indicate, possibly, that there is less there than
we might have hoped. Also, the beam envelope is not quite fitting within the
useful region of the spectrometer as much as one would like. This may just
be reality, and, if so, we should take our lumps and move on. 
    Part of the problem appears to lie in the strong positive curvature of
the entry face, which limits the useful theta acceptance of the device. A
flatter cut at an appropriate angle could have introduced more transverse
focusing at the entry face to compensate for the transverse defocusing of
the exit face. The later was unavoidable to get a short focus with the
limited field strength of a normal conducing magnet. 
    Since it is too late to change the design of the magnet, the question
arises as to whether we can hope to improve the situation by considering
some small adjustments to the optics. The kind of changes I have in mind
include one or more of the following studies. Feel free to add your own
suggestions.
    1) The entry face problem might be alleviated by moving the coils
slightly further away from the beam line, lowering the entry point on the
magnet. A small shift of about 5 cm outwards, with the target at its current
location, should completely eliminate (or at least greatly reduce) any
transverse defocusing effects on the entry face at the higher scattering
angles of ~11-12 deg. However, the major problem is on the exit face and
this may only help to partially alleviate a bad situation. It would also
require a complete retune to realize any benefit. I will drop this
suggestion immediately if the magnet builders object. 
    2) Somewhat the same effect might be achieved by moving the target
downstream, towards the dump, on the order of, say, 25 cm. This, of course,
changes the focus at the rear, and we may not have enough current in reserve
to compensate.
    3) If we need more field strength to improve the overall focus, we might
accomplish this by optimizing the experiment to run at smaller scattering
angles. The smaller angular spread would require less effort to bring into
convergence. This would involve moving the target upstream, which would move
the focal plane in closer to the spectrometer. This works by push more of
the transverse envelope back into our field of view, by not allowing the
divergence to fully develop. At first glance, by pulling the target back, we
should lose acceptance, unless there are significant trade offs to be
realized in the transverse dimension. This might be more readily acceptable
if we come to the conclusion that we really can't use the higher scattering
angle stuff anyway, as the simulations seem to naively suggest.
     I don't think that any of above suggestions will necessary result in a
dramatic effect. We seem to have reached a rough balance in the available
trade offs. Jim and Roger have already explored a good deal of the phase
space. Without a good magnet model to work with, I really can't promise any
great hope. I make the above suggestions in the spirit of all the good work
that has already been done, and in the interest of knowing that all
possibilities have been put on the table. I regret not asking for a more
complete optical analysis earlier, when it could have been done to better
effect and be better received.
    One final point, if, when all other avenues have been exhausted, what
would be so wrong about either requesting more beam time, or settling for a
slightly larger error bar of, say, 4.5%. If these are the cards that Mother
Nature has dealt us, what other choices would we have? We would still have
the best low energy test of the Standard Model realizable in the near
future. 
    Regards,
    Mike Finn