bug in MuonSpoiler magnetic field

[Jochem Snuverink]

As reported by Belgin. Slicing a 5m MuonSpoiler into 5 1m pieces gives different tracking results.

[Jochem Snuverink]

There were 2 issues in the develop version:

1) field attached to wrong pointer, so that magnetic field was zero. this was a recent change

2) field calculation was wrong:

G4double BFactor=itsBField/LocalR.mag(); Bfield[0]= LocalR.y()BFactor; Bfield[1]= -LocalR.x()BFactor;

idea is right, however LocalR.mag() is the 3d radius instead of the 2d radius. z-component is not relevant and will give arbitrary results.

it has been changed to:

G4double LocalX = LocalR.x(); G4double LocalY = LocalR.y(); G4double LocalRadius = std::sqrt(LocalXLocalX + LocalYLocalY);

Bfield[0]= LocalY/LocalRadius * itsBField; Bfield[1]= -LocalX/LocalRadius * itsBField;

tracking for the single and sliced spoiler is now consistent.

[Jochem Snuverink]

see comment 1.