Fets, bias and phantom. What I think neumann designers had in mind with the km84.
First the FET is a high gain one IIRC x60. The circuit runs off phantom. The circuit does not use a DC to DC converter for the capsule. X60 is to much gain without feedback. The transformer is 7:1 I believe in the KM84.
Note the use of a 24V zener to a 10K to the 47K in the drain leg, with the drain at 10V and the "top" of the 47K at about 20V the current is about .21ma 10V / 47K. .21ma is low current so the 48V phantom does not drop alot at the microphone circuit allowing a voltage in the mid 40s to still be advailbe for the capsule. Now from the Neumann book "Microphones" the circuit gain should be the capsule cap in pf / feedback cap that in parallel with the FETs open loop gain, stock km84 34pf and 4pf. The fet being used at a low current and the use of feedback controls the gain so the signal level is not to great for the transformer.
The circuit is kind of like the KM54 circuit if you look closely the AC701 has feedback from the plate to capsule. From what I have read people like KM54s that often don't seem to like microphones circuits with feedback.
I like this circuit because it is clean and allows you to hear the capsule/grill more, it is a good test circuit IMO.