Our bias supply consists of Q4, the tuned RF circuit of L1, L2, &C8, the positive rectifier-doubler D4, D5, C11, R12 &C20 and the negative rectifier-doubler D6, D7,C13,R13, &C19.


Because of the high resistance, drawing even a microampere at the capsule will drop the voltage several volts.

We need to take our measurement as early in the filter circuit as possible to avoid the drop across R12 and R13, so measure at C11 and C13. If our voltmeter is a typical digital type with a 1 megohm impedance on DC measurements, it will knock anywhere from 15 to 20 volts off even so. Increasing the meter impedance to 10 megohms as many better quality DMMs do will reduce the drop to 2 or 3 volts, well within the tolerance we can accept. A couple of volts one way or another will have little effect.

If you have a 1 meg meter, you can make a 10:1 probe for it, which is simply a 9 megohm resistor in a piece of shrink tube. I made one from a 5.1M and a 3.9M resistor in series from my parts box.

That will increase the input impedance to 10M, and reduce the voltage reading by 10X, so 1V on the meter is 10V at the probe tip.


Likewise for a really high impedance probe which will load a circuit very little, you can make a 1000:1 probe from a 1Gohm resistor (999M for the purists). I made one for use with my oscilloscope, and 1mV on the ‘scope equals 1V at the probe tip.

This is what I use to measure capsule supply voltages.


Why do we care about the capsule polarizing voltage? The charge on a capacitor, our capsule, is voltage X capacitance. Since the capacitance, especially the portion which varies with sound, is fixed by the manufacturer, the higher voltage we charge the capsule up to, the more signal we get out of it. Too much voltage, and the capsule diaphragm will collapse and stick to the backplate. For most modern capsules with 3 mil mylar diaphragms, 60V is safe. If the mylar is thicker, 6 mils, 75 or 80V may be OK. 100V is too much for most. A puff of breath will collapse them.


Doubling the voltage doubles the output (6dB) without increasing noise, all things being equal. Here’s a capsule measured at 26, 30, 40, and 60 volts. As always, no smoothing applied.

Notice that the vertical scale is a lot more sensitive than I usually use. Sure enough, the sensitivity is 6dB higher at 60V than at 30V. I gotta find some way of smoothing my graphs without too much work. A 1/3 octave filter makes mics look SO MUCH nicer, and a filter like some manufacturers use just gets rid of ALL those nasty-looking wiggles...


Most of the small wiggles here and in most other measurements I make are due to the cardioid pattern. The reference mic is omni and responds only to sound pressure. A cardioid or fig-8 responds to sound velocity as well. If I had a cardioid reference mic, the responses would match better, but cardioids are neither pure pressure nor pure velocity mics, and standards just don’t exist. I suppose I could declare some common mic to be the reference standard, but folks who don’t own the reference would be out of luck.


I’ve read that capsules sound different at various voltages, due to the tension in the diaphragm or variation in damping due to backplate spacing. If there is any such effect here, it doesn’t show in the graph, tho I think I’ve heard such an effect with other capsules. Probably my imagination. When you THINK something OUGHT to sound different, you usually WILL hear differences.