Indeed. And D3 shouldn't have been there in the first place. That's why I proposed to connect it to the collectors of the output BJTs. OK, original Schoeps circuit has the zener also connected to the drain resistor directly, but that is a less noisy 6.2V zener, shunted by 220uF instead of just 22uF. In my opinion, this is a design flaw. Quickly simulated a slightly different Schoeps circuit that I had on my PC: moving the zener from the drain resistor node to the collectors node decreased output noise density at 1kHz from 30nV/sqrtHz to 20nV/sqrtHz...
Moving the zener to the collectors will inject some noise into the 150k resistors, but with sufficiently large coupling caps and low value drain and source resistors, will be strongly attenuated. And the noise will to a large extent appear as common mode noise on the outputs, so will largely be rejected by the mic preamp. Alternatively, take over the power supply scheme from the pimped Alice design, but forget about the useless 10nF and 100nF caps shunting the decoupling electrolytic capacitors. The multistage RC filtering will completely prevent zener noise from entering the audio path.
Jan
