not really lowpass filtering, simply exploiting the feature of diodes to keep a constant voltage across them regardless of the current through them (for reasonable current values, that is).
3v3--+
|
/
R1 \
/
| R2
Vreg+--/\/\--MICBIAS
|
Z
Δ
|
GND
You'd need a 2.5V Zener diode (other similar values are also fine: the actual value is not so important, as long as it's below 3.3V and above about 2.2V), not much power is needed, 1/2W should be enough. If you don't have a zener around but you have some regular silicon diodes (e.g.: 1n4148, 1n4007, mostly any will do), you could have 3 or 4 of them in series in place of the zener, and pointing the other way:
3v3--+
|
/
R1 \
/
| R2
Vreg+--/\/\--MICBIAS
|
V
T
|
V
T
|
V
T
|
GND
If you want to power more than one microphone, just add more copies of R2 from the 2.5V point, e.g.:
3v3--+
|
/
R1 \
/
| R2
Vreg+---/\/\--MICBIAS-1
+---/\/\--MICBIAS-2
|
Z
Δ
|
GND
As for resistor values, R2 is 1k, and R1 will depend on the characteristic of the diode and how much current is needed in total by all your microphones. 470 ohm is probably a decent starting point, but a bit of experimentation may be needed. The nominal value at Vreg
is whatever the nominal Vz
of the zener, or n * Vf
, where n
is the number of forward-polarised silicon diodes you have in there, and Vf
is about 0.7V, but may depend on what diodes you are actually using. R2 is 1k, and R1 will depend on the characteristic of the diode and how much current is needed by your microphone. Start with R1 = 470 ohm, and connect your microphone(s). Measure the voltage at Vref
. If it is significantly larger than the nominal value, try decreasing the value of R1. If it becomes significantly larger than the nominal value, try increasing the value of R1. There are more scientific approaches to this, where you'd need to know the current absorbed by the load (your mics), the "on" current of the zener (or diodes), and then you can compute the value of R1. See, e.g.: here for more details on this process and Zener regulators in general.
jayrope at the normal 3.3V sources were more or even too noisy
Anyhow, they may actually be just fine for many uses. The 5V source is definitely too noisy, but 3v3 could be fine. Maybe before doing any of the above, try using 3v3 as it is.