Size the degrader mother volume to its contents

[oksuzian]

Sizes the pion degrader's mother volume to its contents.

The bug

The mother's half-width was computed as 2*filter_hl + frame_hl + 1 while its
children were placed from a different expression, so the mother overran its
contents downstream — by 1.9 mm in nominal geometries and 9.65 mm with
geom_run1_b_v40.txt, whose filter is 8.75 mm half-length rather than 1.00 mm.

On v40 that overrun produced two real Geant4 overlaps:

Overlap is detected for volume Degrader:0 (G4Box) with protonabs3:0 (G4SubtractionSolid)
  overlap at local point (219.934,-404.385,-2025.8) by 3.7 mm  (max of 73 cases)
Overlap is detected for volume degraderSupportPlate:0 (G4Box) with Degrader:0 (G4Box)
  overlap at local point (287.173,-70.8224,22.55) by 2.3 mm  (max of 175 cases)

Both reproduce exactly from the geometry constants. The support-plate case:
the old v40 mother's downstream face sits at 4258.85 + 24.85 = 4283.70, and
the plate's front face at 4235 + 245 - 186 - 6.25 - 0.1 - 6.25 = 4281.40 —
a 2.30 mm overlap, matching what Geant4 reports.

Nominal was never affected: its mother face at 4252.70 is already 8.7 mm clear
of a plate face at 4261.40.

The fix

Size the mother as filter_hl + frame_hl + 0.1 and place the filter at its
upstream edge with the frame directly downstream. With surface checking
enabled (g4.doSurfaceCheck = true), v40 goes from 2 overlaps to 0.

Effect on nominal running

The degrader is built in nominal geometries — degrader_v02.txt sets
degrader.build = true, and "off by default" means rotation = 120.0, swung
out of the beam rather than absent. So this is not dead code and the nominal
gdml dump does change. It changes in a bounded way:

No material moves. The Degrader mother is filled with DSVacuum and sits
inside DS2Vacuum, which is also DSVacuum — resizing it displaces nothing.
The filter, frame, rod and counterweight all keep their exact absolute z:

volume	absolute z, nominal	absolute z, v40
filter	4236.00 → 4236.00	4243.75 → 4243.75
frame	4243.35 → 4243.35	4258.85 → 4258.85
rod	4243.35 → 4243.35	4258.85 → 4258.85
counterweight	4243.35 → 4243.35	4258.85 → 4258.85

The normalized gdml diff against geom_common.txt and geom_run1_a.txt is 39
lines each and identical between them. 34 of those lines are file reordering
(the filter is now emitted before the frame; degraderFrame's solid is
byte-identical). The remaining changes are the degraderOutline box dimension,
the degraderMother placement, and three compensating child offsets that exist
precisely so the absolute positions above stay fixed. x and y are untouched
throughout, and no volume outside the degrader assembly appears in either diff.

Step boundaries do shift even where material does not, so re-simulations will
not be bit-reproducible against earlier output.

One deliberate deviation from the Run1B branch

The branch also rewrote location2InMu2e to reuse the mother-centre
expression, which dragged degraderCounterweight upstream by 0.90 mm in
nominal and 8.65 mm on v40 — leaving it non-coplanar with the rod it
counterbalances. That reads as incidental rather than intended, so this PR
keeps the original dgr_z0 + 2.0*filterDims.at(2) + frameDims.at(2). Neither
overlap involves the counterweight, and the 0-overlap result was re-measured
after the revert to confirm the fix still holds. @sdifalco / Run1B authors:
please confirm this was not deliberate.

Minor, noted not fixed

The filter's upstream face is exactly flush with the mother wall — all 0.2 mm
of slack sits downstream. Legal and overlap-free, but coincident surfaces are
generally discouraged, and on v40 that filter is in the beam. Happy to bias the
filter 0.1 mm downstream if reviewers prefer.

[FNALbuild]

Hi @oksuzian,
You have proposed changes to files in these packages:

• Mu2eG4

which require these tests: build.

@Mu2e/write, @Mu2e/fnalbuild-users have access to CI actions on main.

⌛ The following tests have been triggered for fd1a90b: build (Build queue - API unavailable)

About FNALbuild. Code review on Mu2e/Offline.

[michaelmackenzie]

This looks good to me, bringing in the degrader positioning fixes

[FNALbuild]

☀️ The build tests passed at fd1a90b.

Test	Result	Details
test with	✅	Command did not list any other PRs to include
merge	✅	Merged fd1a90b at 1891b32
build (prof)	✅	Log file. Build time: 04 min 19 sec
ceSimReco	✅	Log file.
g4test_03MT	✅	Log file.
transportOnly	✅	Log file.
POT	✅	Log file.
g4study	✅	Log file.
cosmicSimReco	✅	Log file.
cosmicOffSpill	✅	Log file.
ceSteps	✅	Log file.
ceDigi	✅	Log file.
muDauSteps	✅	Log file.
ceMix	✅	Log file.
rootOverlaps	✅	Log file.
g4surfaceCheck	✅	Log file.
trigger	✅	Log file.
check_cmake	✅	Log file.
FIXME, TODO	➡️	TODO (2) FIXME (2) in 1 files
clang-tidy	➡️	2 errors 267 warnings
whitespace check	✅	no whitespace errors found

N.B. These results were obtained from a build of this Pull Request at fd1a90b after being merged into the base branch at 1891b32.

For more information, please check the job page here.
Build artifacts are deleted after 5 days. If this is not desired, select Keep this build forever on the job page.

[sdifalco]

The suggested fix solves the problem