Trajectory in Dipole Parallel-T
Observation
In the comparison of the created maps form Thick Tracker (which I am working on) and the maps of Cosy Infinity , I have noticed a difference between the dipole maps.
The COSY 10.0 input: CompBend.fox cosy.fox
In the closer look: the Thick Tracker algorithm uses the parameter ElementLength, which is the value L of the Input.in- file.
class ElementBase
...
/// Get design length.
// Return the design length defined by the geometry.
// This may be the arc length or the straight length. <-----
inline
double ElementBase::getElementLength() const
{ return getGeometry().getElementLength(); }
The ElementLength in a SectorBend is the straight Trajectory in between the entrance and exit point of the ref particle, as documented.
BUT if I now compare the THICK TRACKER with PARALLEL-T, like in the issue #215 (closed), it won't match this smoothly.
The two map generated Codes differ in the use of either the straight or the curved trajectory for the creation of the dipole map. (In the curved implementation I just changed the map, but not its position to visualize the effect of this change.)
Suggested Solution
After that comparison, I think that there might be a bug in the SectorBend, which probably is just the use of the straight instead of the curved trajectory.