Update 3 corrects a small error in the side view.

The AutoCAD Mech Desktop 2004 file is magnet_tube_k.dwg
e166_positron_spectrometer_top.jpg = top view
e166_positron_spectrometer_side.jpg = side view
e166_positron_spectrometer_end.jpg = end view along beam

These files are also stored on my web page:
http://puhep1.princeton.edu/~mcdonald/e166/

The .dwg file now has layers.  The ones relevant to the things I added are
center lines
coils
dimensions
pole tips
solenoid 1
solenoid 2
targets
vac chamber
yoke

The color code is
black = SS
red = iron
blue = aluminum
green = copper

Note that the iron yoke is now simply a pair of rectangular plates, 1" thick.  They are held apart by rectangular Al spacers at the ends, completely outside the bend magnets.

The magnetic forces are only a few hundred pounds....

I show only a few dimensions, but these are the critical ones.  They are set by Alexander's design of the magnetic field.

Alexander:  We need you to confirm that these dimensions are "good enough".  Note that I give the dimensions in inches, while your drawing was in mm.

Particularly important is the location of the intermediate focus, where we will put the sliding Al jaws to range out the off-momentum positrons.  I show this as 8.87" = 225 mm.  However, I have the impression that the intermediate focus is actually closer than this to the first bend magnet.  

Please advise.

Roman:  Note how I indicate that the upstream half of the vacuum chamber is surrounded by 1" thick Al to range out the lost electrons and positrons.  We hope that you can do a GEANT study as to how effective this is.  Would it be better to use stainless steel, or even lead?

How far down the straight-ahead vacuum pipe do we need to surround this in Aluminum (not yet shown on my sketch.

We can, of course, add even more material, as suggested by the simulation...

--Kirk