You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
hpIntersect: build the CentralHelix plane window from the true helix axis #237
hpIntersect narrows the search with a window around where Traj::linearize meets the plane. For LoopHelix, linearize is the helix axis, so every plane crossing lies in that window (within R·tanθ along the axis). For CentralHelix, linearize returns the tangent, so tmid ± dt does not bound the crossings.
After #235, the axis-ray crossing is taken in either direction and inrange_ is reported against the caller's range, but the window itself is unchanged. In oksuzian's random comparison on #235 (~19,887 low-momentum CentralHelix plane queries), 1,930 still return nothing while stepping the whole query finds the crossing (2,426 on pre-#235main). LoopHelix planes do not show this.
Mu2e CentralHelix fits today are high-momentum cosmics (bend radii of metres), so most take the no-loop branch; #233 found no Mu2e disagreement in 134,639 queries. This is still a KinKal correctness hole for low-p CentralHelix / large tilt.
Suggested fix
Build the CentralHelix window from the true helix axis (center(t) along bnom, as LoopHelix::axis does), then keep the same tmid ± dt / query intersection logic. Optionally share that axis helper with LoopHelix.
Activity