Does impedance match matter for mic to preamp?

t

signal

you'd need to follow it with an unstiction filter :)

NT

Mechanical impedance is all about air coupling. Electrical impedance is determined by the winding, ribbon, charged sheets etc.

Sound is only what can move an eardrum. Brownian motion of a vast number of molecules over the surface of an eardrum is not able to move it or make sound. Yes, our hearing systems have noise too.

NT

Thermal motion is the equivalent of Johnson noise in a gas. I'd expect it to produce white noise, convolved with the frequency response of the microphone. At first sight, the physics should be exactly the same as for the resolution limits of the most sensitive galvanometers. That problem has been studied extensively.

A microphone is just a transformer between the movement of air and an electrical signal. It should be possible to come up with a model and an argument involving the equipartition theorem.

Jeroen Belleman

Is that entirely true? The frequency response in normal use will be modified by the characteristics of the housing and these effects will probably not be visible to the brownian motion generated noise as this is local to the diaphragm.

John

Brownian noise,

is said to go as 1/freq^2. Which makes sense if you think of the random walk of some particle... but I don't know how to think about that in a microphone. (Which has to have some restoring force to bring the element back to 'zero'... and so no DC response.)

OK. That points out another issue in my understanding. The microphone element should have 1/2 kT of energy (motional noise) even in vacuum without any molecules hitting it.

George H.

It's just his time of the month, Mike.

There must be some sort of inflection point in the variation of noise with air pressure when the mean free path of the air molecules exceeds the distance between the diaphragm and the backing plate (for electret mics at least where there is a closely spaced electrode directly behind the diaphragm).

John

Factoid: I remember reading, years ago, that the human "threshold of human hearing" (for a human with ideal hearing - e.g. none of us over

20 who have ever lived in a city or used a Walkman) is only a few dB above the SPL which is created by the Brownian-motion noise of air molecules randomly hitting the eardrum.

I can't find a cite for that at the moment... always wondered if it was accurate.

Hmm I'm not sure. So the 1/2 kT would be for the microphone element alone sitting in vacuum. The Brownian motion thing is not as clear to me. But the noise (variance) would go something like the sqrt(N) where N is the number of particles hitting the thing in some time period. And that would say the noise goes up with N. (and the pressure... PV=NkT) But I don't know if that's right or not.

George H.

Well that's being kind! His time of the month lasts all month every month. I'm curious what he is like in real life. Does he treat people around him this way? Does this behavior, just pop through occasionally or is it his way of life. Mikek

seems unlikely, else we'd hear "seashell noise" all the time, not just when the ear is surrounded by a refelctor.

--
  When I tried casting out nines I made a hash of it.

I tried to locate information about the absolute values of Brownian noise, but it was surprisingly hard to find anything. Some claim more than 20 dB below the threshold of hearing (re 0 dB SPL or 20 uPa). This, when designing really sensitive microphone/amplifier systems, the aim should be well below the threshold of hearing.

The Brownian noise appears to be constant power/Hz, but other sources

consider linear vs. logarithmic frequency scale, which is correct ?

I

elevant to other microphones which do NOT face the same restrictions on siz e and location.

Nearly all microphones have their noise specified as dB(A) which reduces the measurement bandwidth and particularly attenuates low frequencies. Some of the B&K microphones have information about the noise spectrum.

When I visited the Knowles microphone factory I was shown the "quiet cavity" used for noise testing. This was a small sealed cavity in a very large block of steel mounted on a vibration isolating table.

I have been in a large anechoic chamber at a facility in Malvern, UK which claimed to have a noise level of around -10dB(A). I don't know how this was measured. The whole anechoic chamber was mounted on springs inside a large concrete bunker, well away from busy roads.

John

On Saturday, March 23, 2019 at 12:19:58 PM UTC-4, snipped-for-privacy@downunder.com wr ote:

I

elevant to other microphones which do NOT face the same restrictions on siz e and location.

So I found that this RV. Jones article on galvanometers is NOT behind a paywell. Bless the Royal Society.

It's for a torsional oscillator, but treat a mic like a linear harmonic oscillator.

I'm coloring myself confused about a lot of this. I thought a thing(mic.) constrained to move in one dimension would have thermal energy equal to 1/2 kT. (In a vacuum the temperature would couple by radiation with the walls of the chamber.) But reading Jones and others they imply that the gas molecules are responsible for the random motion. (and not the thermal motion of the atoms that make up the mic.) I guess I'm mostly confused (now) by why the noise goes away when the little mics are put in vacuum? (The air is the damping resistance?)

Right I think I figured that out. The 1/f^2 is for an unconstrained particle doing a random walk. (Drunkards walk from a lamp pole.) It's the position (squared) of the particle. (the average position is zero.) In the mic the motion is constrained so 1/2 * m_mic * v^2 = 1/2 * k *T

Hand wavy number-wise we need to differentiate twice to go from x^2 to v^2 so the thermal motion of mic is white.

George H.

That's because he is a different species.

If by "real life" you mean in person (face to face), I recommend you do not travel to Australia and look him up. He might cut your throat, give you AIDS, or inject you with a cancer-causing agent.

Otherwise, just don't respond to his demeaning comments and accept the technological parts of his posts.

Kinda like having a Panther. Admire the beauty, but never let him out of his cage.

Rick C.

** That is not a general truism.

The worlds of RF and audio are very different, only fools try to equate th em.

** Two reasons, it is not necessary and it wouldn't work.

In the world of RF, it is important to match the characteristic impedance o f connecting cables in use to the load - this avoids reflections and uneven response. No such need exists in audio.

When equal value matching is done between a source and load, half the power generated is lost - so even with RF it is rarely done. Throwing away half t he power of a 50kW transmitter is not sensible and receiver antennas are of ten NOT matched to the connecting cable impedance.

Low noise design is most important when interfacing a signal source ( eg a transducer ) with a pre-amplifier. With no need to match a cable, the input impedance of the preamplifier can be made high as you like allowing the so urce to deliver its maximum signal voltage to that input. The result is the best possible s/n ratio.

** The fact is they deliver max output voltage into an open circuit, the te n time load rule is close enough for practical purposes and maker's output ratings. Response is rarely much affected by load impedance value. ** It would be far noisier than the "voltage matching" method normally used .

.... Phil

them.

of connecting cables in use to the load - this avoids reflections and unev en response. No such need exists in audio.

er

the power of a 50kW transmitter is not sensible and receiver antennas are often NOT matched to the connecting cable impedance.

a transducer ) with a pre-amplifier. With no need to match a cable, the inp ut impedance of the preamplifier can be made high as you like allowing the source to deliver its maximum signal voltage to that input. The result is t he best possible s/n ratio.

ten time load rule is close enough for practical purposes and maker's outpu t ratings. Response is rarely much affected by load impedance value.

ed.

Wow! A reasonable reply with no vulgarities. I'm impressed.

Rick C.

OK, how about 200mV.

--
 Thanks, 
    - Win