Rob Danielson wrote some really good questions:
(1) is this (AT3032) an electret mic capsule, not a "true" condenser?
It's an electret capsule. I make no distinction between electret and non-e=
lectret microphones in terms of their being "true" condensers. They're bot=
h condenser microphones, and they both have air gaps between the diaphragm =
and the back plate. One type may have an advantage over the other, but as =
you write below:
> I think we may be concentrating on capsule design "explanations" too much=
I'm quite sure that you are right. Microphone manufacturers try to make th=
eir designs resistant to humidity, but the results seem to be all to variab=
le in practice. Sometimes mfgrs. use a conformal coating on the PCB in the=
area right at the capsule input. But I wouldn't be surprised if this does=
n't make problems worse under some circumstances. The impedance at this po=
int is a gigOhm or higher in most designs (excluding RF microphones) and th=
e conformal coating may act as a sort of matrix to accumulate various sorts=
(2) Could it be that small dirt particles are collecting or some other phys=
ical change is happening with these exposed capsules that is affecting the =
nature of the gap?
Possibly, but it should be nearly impossible for dust to get in the gap (un=
less it was there when the mic was assembled) with an omni microphone. A l=
ittle easier for dust to get in the rear of a cardioid microphone, but most=
mfgrs. take good care to make that difficult.
(3) one could actually study the duration-induced phenomenon, maybe with a =
microscope of some capsules, to find out what is producing the physical cha=
nge and build this "change" into the mic at the outset.
An interesting experiment would be to substitute a capacitor with the same =
capacitance as the capsule and test to see if the discharge phenomena occur=
under those circumstances. Assuming that the problem occurs reliably (is =
that a contradiction in terms?), it could be determined whether the problem=
is occuring within the capsule, or afterwards in the circuit.
Let me think out loud here for a moment. These kinds of problems (popping,=
fizzing, moisture induced failure) occur because moisture prompts a discha=
rge of current in the input circuit of the microphone. This could occur in=
the gap of the capsule, in the circuit from the capsule to the input to th=
e FET, across the bias resistors, across the input coupling capacitors (if =
any). My gut feeling is that most of these problems occur in the path from=
the capsule to the gate of the FET. But I'm really talking from my nether=
orifice. I don't know without doing some experiments.
I recall Carl Countryman, manufacturer of Countryman lavalier microphones (=
http://www.countryman.com/ ), demonstrating his products at AES conventions=
by immersing his microphones in a glass of Cola and then withdrawing them =
to show that they worked instantly after being withdrawn from the beverage.=
This is one of the reasons that I've occasionally recommended using laval=
iers for nature rcording in the past, the others being durability and relat=
ively low cost. But the quietest of the Countryman designs have a self noi=
se of about 24 dBA. That's not really in the same category as a microphone=
like the Rode NT1-A or Audio Technica AT-3032. But it does lead one to th=
ink about a design that uses lavalier construction principles around a very=
quiet capsule and circuit. The basic idea here is to keep _all_ the moist=
ure out of the sensitive parts of the microphone.
Any one want to work on that?