[BTeam] Lambertson multipoles

[Jay Benesch]

In late 2014 I modeled the Lambertson at Arne's request.
https://jlabdoc.jlab.org/docushare/dsweb/Get/Document-10105/15-005.pdf

There is a large skew quad and a significant normal sextupole resulting 
from the Lambertson in all beam paths, but Hall B suffers most from skew 
quad as HPS demonstrated.  Mike Tiefenback recently asked me to design a 
Panofsky quad with could be mounted to null the Lambertson skew quad. 
See results in
https://jlabdoc.jlab.org/docushare/dsweb/Get/Document-117253/16-007.pdf

BPMs were installed between the vertical MBDs over a decade ago at the 
suggestion of Johan Bengttson.  I must have been asleep during the 
discussion because the beams are not supposed to be at zero vertically 
in the A and C lines until C01 girder, after the MBDs take out the 
residual vertical angle from the transport recombiner.  Steering to zero 
at C00 screws up steering downstream.

In response to Mike's query I decided to spend ~700 CPU hours 
calculating the fields in the Lambertson for all combinations of A and C 
passes, 25 in all.  The paths used to calculate multipoles were started 
at A (0,2.2,-150) ; B (0,0,-150) ; C (0,-2.2,-150) even though the A and 
C beam pipes in the Lambertson are horizontally offset from zero by by 
about +- 0.35 cm.  Ops doesn't offset the A and C beams from the center 
of Lambertson viewer - it stacks them in a vertical line.  I attempted 
neither to get the entrance heights nor the angles for each pass right 
at z=-150; I used one orbit parallel to the Z axis starting at the 
indicated locations in calculating multipoles.  Since I have the full 
field models I could get fancier if someone can convince me the effort 
would be worthwhile.

Since the A and C channels are offset from zero and the orbits I chose 
are not, I calculated all 25 combinations rather than just the 15 
on/above the diagonal.  Decide for yourself from the attached whether 
the offsets are significant.

I suggest the the B team consider installing girders consisting of 
relocated BPMs, Panofsky skew quads, and sextupoles relocated from C08 
girders about midway between the Lambertson and C01 girders.  BPMs would 
be zeroed via the Panofsky quads and the design vertical offset at that 
location then put in GOF so Ops could steer to zero on C00 and C01 
girders using the MBDxC00V correctors.  Skew quad and normal sextupole 
would be set according to my Lambertson model results.  For conventional 
solenoids, dipoles and the combined function PEPPo spectrometer my 
models agree with measurement to better than 0.25%; for the multipoles 
I'd estimate 5%.  Nulling skew quadrupole and normal sextupole within 5% 
will improve matching and beam quality.

If you deem the effort superfluous, it was only about a day of my time 
and 100 kWh wasted.

Jay


-------------- next part --------------
A non-text attachment was scrubbed...
Name: Lamb_25sims_multipoles.ods
Type: application/vnd.oasis.opendocument.spreadsheet
Size: 16398863 bytes
Desc: not available
URL: <https://mailman.jlab.org/pipermail/bteam/attachments/20160424/e534d284/attachment-0001.ods>