Well, here's page one of my homebrew stuff. I haven't taken photos of my gear yet but here's a few projects I've done.

• Reverse-engineered the TLM103 amplifier. It's clever.
• Upgraded an Apex 430 microphone's amplifier, transformer, and capsule. Really, rebuilt the whole mic...
• Rebuilt an Apex 460 microphone. New capsule, transformer, and removal of the cathode follower stage. It doesn't need it if the transformer is correct.
• Built from scratch a CK61-ULS based tube mic. It needs a quieter tube but it works. I'm using a 5744 tube.
• Designed an upgrade for the Oktava mic's. It emulates the C61 system where capsule bias is obtained via the cathode load. Check out the photo...

R1 feeds bias voltage to the capsule and to the JFET Q1. R2 sets the quiescent current of Q1. R5 is set to adjust the voltage drop across Q1, but is typically about 10k. R4 feeds additional current to Q2. Note that Q1's current is also reused by Q2 just like in the original Oktava circuit. R7 and R6 set the nominal DC output voltage. Note that the buffered output signal is applied to pin 2 (the ouput) but also via C2 to the power supply of Q1. This tends to reduce gate-channel voltage modulation by forcing an (approximate) DC voltage across Q1, as well as lowering Q1's output impedance. C5 is the familiar pseudo-balanced output grounding capacitor. To improve cable noise rejection, series resistors of 49 ohms in line with pins 2 and 3 of the XLR connector may be a wise idea. To do even better, the resistor in series with pin 3 should be matched to the output impedance of the amplifier. I have not tested this circuit yet, but simulations show it works well. The next run of PC boards I do will have this design on it as a replacement PC board for the Oktava MC012.

Dale's philosophies on condensor mic preamps:

•  the simpler circuits don't perform as well technically but they tend to give a certain character to the sound. If we wanted microphones that perform technically well, B&K (DPA) or Earthworks measurement mics would be all we would ever use. The U47 has a lot of distortion and anything but a flat frequency response. But they sound good.
• Frequency response isn't everything. Frequency and phase response and distortion characteristics need to be all considered. Remember that the output of most mic amplifiers (the one or two transistor kind, or glass, for that matter) is fed back capacitively to the input. Any distortion or frequency response anomalies in the mic amplifier are re-applied to the input of the amplifier but also to the electro-mechanics of the capsule. It seems pretty obvious but I never see a phase plot of a microphone in a data sheet, let alone real distortion characters. But separating out each component and duplicating each component (frequency, phase, and distortion) may not result in 'the sound'. This sort of transfer function may also be non-reversible. If you know the output you may not be able to get the input that accurately.
• 'The magic' isn't magical at all. Take the U47, for example. Its sound has a magical quality to it, yes. But that magic is predictable. Start with the shape and size of the body, mesh size and air volume around the capsule, the capsule itself, and all of the real amplifier characteristics. Add the magnetic characteristics of the output transformer and the inevitable coupling between the magnetic leakage of the output transformer and the amplifier tube located directly above it.... if you really analyze the system carefully you'll find the 'magic'. It ain't voodoo, folks. If you take the time to really analyze what the system is doing you find the answer. That's not to say you need to but you can't just dismiss things as magic or unpredictable or unrepeatable. If you really wanted to, you could clone the sound of a U47.