I agree the JFET output should be connected to the start of the winding. That's also how Neumann does it, for whatever reason(s). But I'm not sure the positive effects can be attributed to better shielding: having the ground connected to the outer layer does indeed work as a shield against electric fields, both against emission and pickup, but the mic body already acts as a Faraday shield against electric fields. The transformer shield would do little to nothing to that end. And the shielding effect also doesn't explain the interaction with some downstream devices that @mrgrooves666 describes, at least I don't see how that can be the case.
I therefore think you can only explain the observed noise with different preamps by the different loads that the devices in question and the XLR cable pose to the microphone circuit. Transformer, cable and input circuit of the preamp form a complex set of parasitic capacitances, (coupled) inductances at RF frequencies that can affect the stability of a connected amplifier. This complex set of parasitic circuits changes as soon as you swap the transformer start and end of the windings of the primary and secondary. With the JFET side attached to the start of the winding, it has the highest capacitive load to ground through its high capacitance to the core, which on its turn has a high capacitance to the mic tube and other conductors connected to ground. Especially when the transformer core is galvanically connected to ground, the capacitance will be even higher.
Suppose all the circuit parasitics do indeed cause the JFET to become unstable, inserting a resistor between drain and coupling capacitor will introduce a dissipative element in the resonant circuit and prevent instability. Noise contribution of this resistor is negligible, because of the transformer ratio.
Jan
I therefore think you can only explain the observed noise with different preamps by the different loads that the devices in question and the XLR cable pose to the microphone circuit. Transformer, cable and input circuit of the preamp form a complex set of parasitic capacitances, (coupled) inductances at RF frequencies that can affect the stability of a connected amplifier. This complex set of parasitic circuits changes as soon as you swap the transformer start and end of the windings of the primary and secondary. With the JFET side attached to the start of the winding, it has the highest capacitive load to ground through its high capacitance to the core, which on its turn has a high capacitance to the mic tube and other conductors connected to ground. Especially when the transformer core is galvanically connected to ground, the capacitance will be even higher.
Suppose all the circuit parasitics do indeed cause the JFET to become unstable, inserting a resistor between drain and coupling capacitor will introduce a dissipative element in the resonant circuit and prevent instability. Noise contribution of this resistor is negligible, because of the transformer ratio.
Jan
