Hans, in your sims have you looked at how noise changes depending on where you put the first pole? I know it's low enough to actually not be worth worrying about, but the expectation would be noise would go down as the pole moves up in frequency.
Whith poles moving up in frequency, either Lcoil is smaller or Rcoil is larger.
The first one results in less noise, the second one in more noise when Rcoil is the only resistor in series with Lcoil.
When for instance looking at the 580mH and 460 Ohm Cart that I used, I had to insert a 700 Ohm resistor in series to fulfil R= 2*Lcart-Rcart.
This gave a pole at 318 Hz.
Had I left the 700 Ohm, the pole would have been shifted down to 126 Hz and noise generated from the 700 Ohm resistor would have been avoided.
After some recalculation of resistors around the first stage, this would have given a slight improvement in dBA.
However a 700 Ohm resistor generates 3.5 nV/rtHz as against 5.1nV/rtHz for the OPA 1642, so I do not expect anything significant.
R=2*Lcoil - Rcart is good enough and no reason to worry.
Hans