Here's a schemo of an AT capsule (at least the one I opened up, I forget the model). It had an internal source resistor, which is a good thing, it lowers distortion. There are three connections; ground (bottom), FET drain (top), and FET source (middle). There are a number of ways to connect to it:
"2-wire": this would involve wiring bias voltage across a resistor to drain, and not using source. Signal is taken from the drain wire as well. In this configuration, there will be gain (within limits) at the FET according to the ratio of bias (drain) resistor to source resistor. If the resistors match, there is no gain.
There are a few possible ways to do 3-wire; the one I suspect most people here are thinking of is to wire the bias voltage straight to drain, and take signal from the source. This is a unity gain configuration.
Another way to do it is supply the bias voltage across a matching resistor, and take signal off both drain and source. This uses the FET as a phase splitter and provides an electrically balanced output. To take advantage of that, you would need a differential amplifier (or a transformer) somewhere down the line. That would also be unity gain, but you get a free bump from the balanced signal.
I don't know what AT does in their bits . . . I have a wireless pack on my desk, but I don't feel like opening it up and tracing its circuit. They do appear to use matching resistors in the capsule and bias supply, so I suppose they might provide a balanced connection, but I don't really know. I suspect this may be the cause of the claims that AT's adaptors "ideally power" the microphones, but again that is wild speculation. The adaptors probably also provide a relatively high-impedance input, which helps the capsules out a bit . . .
Thanks for the info. What Chris said: some, but not all AT capsules have that source resistor!
I've noticed it in *a very few* AT831 (connects to wireless transmitters, I think).
But most AT853 I've seen don't have this. The source resistor (5.1k usually) is in the phantom power unit. That is, the package that is either hardwired to the cable, or connected with a 3 pin miniXLR.
To determine if there is a source resistor *inside* the capsule, take off the connector. There are three wires: red (positive), yellow (audio), and shield. Simply measure the resistance between yellow and shield. Open circuit = no source resistor.
By the way, both Chris and I have done mods where we *add* that source resistor (usually a 4.7k or 2.2k) right inside the capsule. AT853 caps are easy to disassemble, so this is a nice fix. Then you can run off a "standard" 9v battery box. Note: I'm not offering to do this. Chris might, but I'm not doing mods right now.
Richard