Making a mic, part 3

So last time I posted pictures of the nearly finished preamp. Several modifications were made since those pictures were taken, namely, the addition of some offset adjustment circuitry, as well as the DC supply regulators. Because the board wasn't large enough, these components were attached to a second board, and stacked together with the first. Connected to the second was also the contact (a banana plug) that fits into a slot in the back of the back-plate.

Top side of the preamp. Kind of ugly, what with the wires going everywhere and resistors splaying out all oover. When I settle on a final design, I might get to make a printed circuit-board design, which should clean things up considerably.

Stacking the second board with the voltage regulators etc. beneath the first

Banana pug for fitting into the backplate. The top is sparsely populated.

Final, assembled microphone - the circuit board is inside.

The mirror like film on the front is actually a 6-micron thick sheet of mylar with aluminum deposited on it. It is very thin but very tough stuff.

Unfortunately, ambient pressure changes fairly fast, sometimes on the order of a pascal per minute. If you have a sealed canister, it does not take long for a very large (from an acoustic standpoint) pressure imbalance to form across the diaphragm. If the mic is sensitive to DC, this can cause major problems. For this reason, an air leak is put behind the diaphragm of microphone. The leak allows the pressure across the diaphragm to equalize. The size of the air-leak determines the low-frequency cutoff of the microphone. In this case, it forms the dominant cutoff. I want this mic to be good down to about 0.001 Hz, I built this little canister here, to determine the appropriate leak:

Inside the canister is a capillary tube wrapped around a brass spool, that is long enough to determine the appropriate size of resistance.
This leak screws into the back of the condenser microphone. The reason for this is to make the back-leaks conveniently interchangeable. (I.e. suppose I want to make the back-leak cut-off at 0.01 Hz instead of 0.001 Hz, I can unscrew the old back-leak, and screw in a canister with a different length of capillary tubing). There will be problems with this related to heat-expansion of the tubing. The acoustic resistance of the tubing is very sensitive to change in diameter, and, though I have not tested it, the thermal-expansion of Teflon should be quite a bit larger than steel. Oh well, it's the best I could do with what i had on me. The 0.01 - 10 Hz region should remain reasonably unaffected, and that is the region of most interest.

(Back leak, screwed to the microphone)

So, here is the whole microphone, sitting upside-down on my desk, with a big syringe sticking out of the back.


The syringe allows me to raise the pressure inside the microphone and watch the decay. This lets me find a time-constant for the leak, and allows me to characterize the resistivity of the back leak. It also allows me to see how thermo-viscous losses will be affecting the compliance of the chamber behind the diaphragm (not an insignificant thing).

Not quite done however... the signal is still pretty low-level. Also, as of right now there is still some junk in the signal left over from processes done in the preamp. This little box allows me to battery power it if I want, as well as apply some selective filters to the output.

(The inside of the box. The big empty space is for if I want to make the thing battery-powered)

Here's the box from the outside. Two outputs and power input +-gnd inputs (if it's not being run by batter).
Here's Matt, posing while I test whether or not I busted the SD card in the camera. I guess it still works. Matt is research faculty in the lab, and he's helped me with innumerable mechanical problems in the past. He also helped me get my bed from storage at the beginning of the year, so I didn't have to sleep on the carpet anymore.

Anyway, as far as the mic goes, I could keep working on the thing for a long time but I can't. There's a billion improvements I would like to make. However, the mic itself isn't my research project, it was just built to help with my project. So even if I see a dozen ways to make it better, if it's good enough I sort of need to say "it's good enough" [rrrrgh] and get on with the rest of my research. Today though, in the shower again (where, like normal, all my best ideas come to me. I don't understand this), I though of a way to possibly suppress 1/f noise in the electronics. Even Dr. Gabrielson thought it was cool and said I could have a shot at cooking up the system I described to him. So who knows... there might be a part 4 coming up sometime.