its still basically zero

[trontrytel]

F[3] = -1.8681f19
H[3] = -1.24394f20

For ρᵢ = 200 kg/m3 exp(-ρᵢ) = 0.0f0. Or in Float64 exp(-ρᵢ) = 1.3838965267367376e-87. So for all intents and purposes those terms are zero no matter how we write them. We would have to go down to apparent densities of 50 kg/m3 for those terms to start showing up.

But just in case, I think this implementation will behave better numerically.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 95.81%. Comparing base (852c662) to head (3e56f04).
Report is 3 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #478   +/-   ##
=======================================
  Coverage   95.81%   95.81%           
=======================================
  Files          41       41           
  Lines        1601     1601           
=======================================
  Hits         1534     1534           
  Misses         67       67           

Components	Coverage Δ
src	97.47% <100.00%> (ø)
ext	69.79% <ø> (ø)