CSBEND question
Moderators: cyao, michael_borland
CSBEND question
Can you please explain the sign convention for E1 and E2 for a CSBEND?
-
michael_borland
- Posts: 1933
- Joined: 19 May 2008, 09:33
- Location: Argonne National Laboratory
- Contact:
Re: CSBEND question
Michele,
The convention is the same as MAD. If E1=E2=0, then it is a sector magnet. If E1=E2=ANGLE/2, then it is a rectangular magnet; the sign of E1 and E2 is the same as the sign of ANGLE in this case.
Having E1 or E2 with the opposite sign from ANGLE would be a very unusual case.
--Michael
The convention is the same as MAD. If E1=E2=0, then it is a sector magnet. If E1=E2=ANGLE/2, then it is a rectangular magnet; the sign of E1 and E2 is the same as the sign of ANGLE in this case.
Having E1 or E2 with the opposite sign from ANGLE would be a very unusual case.
--Michael
Re: CSBEND question
Mike,
The question of sign convention on E1 and E2 has been causing us some grief in getting our decks cleaned up. The problem we are trying to solve is arriving at a consistent convention for E1 and E2, not just in terms of edge focusing but also geometry. Elegant only cares about edge focussing (that is all it is using E1 and E2 for correct?). The entrance and exit angles are also important in calculating the pathlength through the dipoles. From a geometrical sense E1+Angle = E2. The MAD definition (page 40 in the CERN document) is that E1 and E2 are positive for transverse focusing which is causes grief from the geometrical sense.
So according to MAD, E1 and E2 are essentially always positive, and for a "standard" positive bending dipole E1=E2=angle/2. For a negative bend, the traverse effects are the same so I would have thought that E1 and E2 are still positive??? The edge effects go as tan(E1)/R, so if E1 changes sign focussing changes to defocussing.
Unfortunately CEBAF has some special cases where the dipoles are infact tilted and the incoming and outgoing angles are not equal. T
The question of sign convention on E1 and E2 has been causing us some grief in getting our decks cleaned up. The problem we are trying to solve is arriving at a consistent convention for E1 and E2, not just in terms of edge focusing but also geometry. Elegant only cares about edge focussing (that is all it is using E1 and E2 for correct?). The entrance and exit angles are also important in calculating the pathlength through the dipoles. From a geometrical sense E1+Angle = E2. The MAD definition (page 40 in the CERN document) is that E1 and E2 are positive for transverse focusing which is causes grief from the geometrical sense.
So according to MAD, E1 and E2 are essentially always positive, and for a "standard" positive bending dipole E1=E2=angle/2. For a negative bend, the traverse effects are the same so I would have thought that E1 and E2 are still positive??? The edge effects go as tan(E1)/R, so if E1 changes sign focussing changes to defocussing.
Unfortunately CEBAF has some special cases where the dipoles are infact tilted and the incoming and outgoing angles are not equal. T
-
michael_borland
- Posts: 1933
- Joined: 19 May 2008, 09:33
- Location: Argonne National Laboratory
- Contact:
Re: CSBEND question
Arne,
At least in MAD 8.1 (the last version I used), the dipole edge angles are done the same as elegant. In particular, a rectangular dipole always has E1=E2=ANGLE/2, even when ANGLE is negative. The reason is that the edge focusing focal length is 1/f = -tan(e)/rho, where rho is a signed quantity. I.e., if ANGLE<0 then rho<0.
I checked this just now with the PAR lattice, which has ANGLE<0 and E1=E2<0. It's unstable if I put the wrong sign for E1 and E2. If I have negative E1 and E2, both elegant and MAD agree on the beta functions.
The path-length for an on-axis, on-energy particle won't depend on the edge angles. This is true by construction, since the orientation of the end faces is defined by pivoting about the trajectory for an on-axis, on-energy particle.
--Michael
At least in MAD 8.1 (the last version I used), the dipole edge angles are done the same as elegant. In particular, a rectangular dipole always has E1=E2=ANGLE/2, even when ANGLE is negative. The reason is that the edge focusing focal length is 1/f = -tan(e)/rho, where rho is a signed quantity. I.e., if ANGLE<0 then rho<0.
I checked this just now with the PAR lattice, which has ANGLE<0 and E1=E2<0. It's unstable if I put the wrong sign for E1 and E2. If I have negative E1 and E2, both elegant and MAD agree on the beta functions.
The path-length for an on-axis, on-energy particle won't depend on the edge angles. This is true by construction, since the orientation of the end faces is defined by pivoting about the trajectory for an on-axis, on-energy particle.
--Michael