Floor coordinate

[Hossein_Ghasem]

Hi Michael,

I had a question about floor coordinate calculation with elegant. We would need to have 3 horizontal steerers and 3 vertical chicane magnets in one of the storage ring straight section (in K18 beam line). The steerers and vertical chicanes have been defined in the lattice file as given below;

k18_steer_1 : csbend, l = 0.15, angle = 0.002, k1 = 0.0, e1 = 0.001, e2 = 0.001, N_KICKS=40, SYNCH_RAD=0
k18_steer_2 : csbend, l = 0.20, angle =-0.004, k1 = 0.0, e1 =-0.002, e2 =-0.002, N_KICKS=40, SYNCH_RAD=0
k18_steer_3 : csbend, l = 0.15, angle = 0.002, k1 = 0.0, e1 = 0.001, e2 = 0.001, N_KICKS=40, SYNCH_RAD=0

k18_chic_1 : csbend, l = 0.15, angle = 0.0005, k1 = 0.0, e1 = 0.00025, e2 = 0.00025, TILT = "pi 2 /", N_KICKS=40, SYNCH_RAD=0
k18_chic_2 : csbend, l = 0.15, angle =-0.0010, k1 = 0.0, e1 =-0.00050, e2 =-0.00050, TILT = "pi 2 /", N_KICKS=40, SYNCH_RAD=0
k18_chic_3 : csbend, l = 0.15, angle = 0.0005, k1 = 0.0, e1 = 0.00025, e2 = 0.00025, TILT = "pi 2 /", N_KICKS=40, SYNCH_RAD=0

The chicanes includes of 90 degree tilt to rotate the dipole along longitudinal axis and properly simulate vertical chicane. When I calculate floor coordinate for straight of K18 (ID_K18 beam line) I got below plots which look reasonable. Layout of chicanes and steerers are schematically shown as well. The bottom plots show the horizontal and vertical bumps. Both horizontal and vertical coordinates are nearly zero after the third steerer and the third chicane.

Once I calculate floor coordinate for whole ring (ringRF beamline), I don’t get reasonable results in the vertical plane (Please see below plots) and the 3D plot shows that the orbit is off by almost 200 um at the end of ring.

I expected to only see a horizontal line with a bump at K18 in the top right plot and a closed loop in the bottom plot. It looks that there is additional kick or bend even by each ring dipoles after the vertical bump cancellation which is not sensible. In other words, the ring horizontal dipoles after K18 vertical bump gives a bending or kick in the vertical coordinate which is odd. This is shown in below plot.

Just for your info, I used below command to calculate floor coordinate;

&floor_coordinates
filename = %s.flo
X0 = 0.0;
Z0 = 0.0;
theta0 = 0.0;
include_vertices = 0;
vertices_only = 1;
magnet_centers = 0;
store_vertices = 0;
&end

I would appreciate it if you let me have your valuable comments or suggestions. Did I miss some thing in the calculations?
Many thanks!
Regards
Hossein

[michael_borland]

Hossein,

The problem is that when you interleave horizontal and vertical bending, you don't necessarily get back to the same coordinate system in the end. The reason is that each bend's TILT parameter is defined in the coordinate frame defined by all the upstream bends.

There are two ways to fix this: (1) you can rotate some of the interior bends to null out the three angle coordinates (theta, psi, phi) at the end (2) you can order the bends in such a way that bends in the horizontal (vertical) direction only occur when the beam is traveling in the original horizontal (vertical) plane.

I think the second approach is preferred because it doesn't depend on the angles. An example of the second approach is to split the second vertical bend, then optimize the drift distances. Please see the attached files.

--Michael