The thing that has had me wondering about
the Fethead is its amplification varies by load. That is a sign of high output impedance, which makes me wonder . . . in fact, the table they have published is consistent with an output impedance of 6.8K ohm! This makes me think that the circuit simply connects incoming pins 2 and 3 directly to the FETs' drains. This is the same way that consumer-grade camcorder-style mics work (albeit unbalanced). So if you use one, do not use it at the microphone side of the cable. This is the opposite of the diagram they have in their datasheet.
There are other things that make me wonder, such as the lack of a noise spec. They say that the noise according to the FET datasheet is too low for them to measure, but at 20dB gain it shouldn't be hard to measure equivalent input noise.
They also say that the noise experienced depends on the noise of the phantom power supply, which is another way of saying the device has poor power supply rejection (although the simple circuit above would reject supply noise *if* the gain at both FETs matched, which it would if they used source resistors or a differential configuration. If they grounded the sources, that would mismatch gain (without matching FETs) and degrade not only PSRR but also CMRR).
http://www.tritonaudio.com/tritonaudio/images/stories/fethead_user_guide.pdf
Personally, I would shop for a low input noise preamp instead, this device seems to have too many caveats for me.
