Re: Hall D / Tagger magnet orientation (fwd)

[Elton Smith <elton@jlab.org>]

---------- Forwarded message ----------
Date: Thu, 29 Jun 2000 11:07:54 -0400
From: Jay Benesch <benesch@jlab.org>
To: Elton Smith <elton@jlab.org>, jonesrt@jlab.org
Cc: douglas@jlab.org, harwood@jlab.org, lebedev@jlab.org, andrew@jlab.org
Subject: Re: Hall D / Tagger magnet orientation

Elton and Richard, 

	I just spoke with Dave Douglas about the problem because I hadn't come
up with a satisfactory concept after talking to Elliott. I understood
from him that the problem is tagger resolution - having the beam at a 20
degree angle increases the effective horizontal width.  
	Bottom line from my conversation with Dave is that the best choice is
vertical tagging since the beam is so much smaller vertically than
horizontally.  Next best choice is to spend considerable time (weeks),
beginning with degradation of the vertical beam properties and
reclaiming the ten meters I removed from the line, so the required
focusing to get a round spot at the collimator is closer to equal for
horizontal and vertical and both can be kept smaller at the radiator. 
The third choice is to throw out the design entirely and start over,
including a horizontal bend somewhere upstream or downstream of the ramp
so one can use a five quad rotator to exchange horizontal and vertical
properties.  This will couple vertical to horizontal dispersion, which
is why one needs the horizontal bend.  Dave says one could also use a
helical ramp and do everything in it, but then the hall would have to be
at an angle to the linac and the hall likely won't fit on the site. 
There isn't room to do the rotation after the ramp.  (Hmm. 
superconducting quads??)

Constraints and comments

1.  incoming horizontal emittance is five times vertical, so horizontal
size is (root five) larger than vertical

2.  beam is translated 50 cm down from linac height in the spreader and
left there, with vertical dispersion, past the point of tangency.  The
vertical dispersion isn't nulled until the top of the ramp.  Therefore
anything one does to exchange X and Y before the top will couple
vertical to horizontal dispersion and require a horizontal bend and at
least couple of quad cells up top to compensate - and likely the room to
do that doesn't exist.  

3.  Space exists to kick the beam horizontally about 50 cm after the
spreader, putting it behind the magnet stands, but I was told not to do
that.  I investigated because my thought when I got started on this a
year or more ago was to bend horizontally, vertically, and horizontally
again, so I could cancel both vertical and horizontal dispersion just
after entering the North Stub.  IF Claus is willing to relax this
constraint, one could bend horizontally in the extraction region to pick
up horizontal dispersion and then completely redo the optics to rotate
the beam and cancel both dispersions by the top of the ramp.  This would
take me a month or more with help from Dave or Valeri.  

4.  Vertical emittance is degraded only 15% in the present design.  I
can increase this to ~60%, so the X:Y ratio is only 3:1, by removing two
of the bend dipoles.  I can also increase the vertical beta at the
dipoles to reduce the ratio more.  Removing the dipoles precludes
operation above 12 GeV.  

Bottom line (again) The only simple and elegant solution is vertical
tagging.  This improves resolution by more than a factor of two and gets
the electron beam pointed where it belongs,  down.  

						Jay

Elton Smith wrote:
> 
> Hi Jay,
> 
> As you probably know, we had tentatively decided to orient the tagger such
> that it would continue to bend electrons in the horizontal plane, but also
> partially downwards so that the electron beam dump would be below ground
> level. The angle of the bend plane (relative to the horizontal) is
> approximately 20 degress. We realized that this is large enough that the
> phase space of the beam will clearly need some quads to compensate. Do you
> have any quads in the design for this purpose and/or could you add them to
> the bend string? Could you give a rought cost estimate for this?
> 
> Thanks, Elton.
> 
> Elton Smith
> Jefferson Lab
> elton@jlab.org
> (757) 269-7625