Hints for a low noise measuring microphone for test

you need to decide if you want an OMNIdirectional mic or cardioid.

usually acoustical rather than electrical noise is dominant

m

Not an audio guy, have you tried a little cheap piezo mic? We use one in an acoustic resonator, works fine. What sort of frequency range are you interested in?

George H.

Low resistance has low noise. Use a 2 ohm loudspeaker, in reverse, as a microphone.

How about summing an array of cheap electrets?

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

It also has very low signal voltage, which makes the preamp a serious challenge unless you use a transformer. H&H have an extended example on a transformerless ribbon mic preamp.

To get low noise, you need to average over a decent diaphragm area, so John's notion of averaging several electrets sounds promising. At high or low frequency there are phase shifts to worry about.

Cheers

Phil Hobbs

So I see the phase shift worry at high frequency. (wavelength ~10 cm at 3kHz, assuming SOS is 300m/s) But why is there a low freq. phase shift? Are you just thinking about the sensitivity fall off at low frequency?

George H.

That and the associated phase lead due to the AC coupling. One wouldn't wor ry about it too much for voice, but this is an instrumentation device.

Cheers

Phil Hobbs

While make-your-own is appealing, if you measure, there must be a calibration. So, what you want is also gonna be recalibrated at some interval...

When one really CARES how it keeps a calibration, buy (or rent) something that has been field-tested for years. Preferably with manufacturer calibration support.

I don't know if this helpful, take a look.

Mikek

I haven't tried a piezo. I need linearity across frequency, and I think the piezo is limited in that respect, and due to the construction more sensitive to vibration

I am working on acustic emissions from a product, range from 100Hz to 5kHz is the normal range we need to observe

Cheers

Klaus

se floor) measuring microphone

tion.

g that has

upport.

We have a precision setup, but I cannot access the test equipment since it has full booking for many weeks

So, I am making my own setup. The first measurements were done in the big t est lab, so that's my baseline. I "just" need to recreate about the same no ise spectrum in my setup, and then I can fiddle to reduce the noise and whe n done bring the unit back to the big test lab for verification

Cheers

Klaus

** Mics with an equivalent 6dB SPL noise floor are rare beasts - are you SURE you need that ?? The environment must be as quiet as a good anechoic chamber. ** Reported to have a self noise level of about 24dB SPL - typical of most small, electret capsules.

The secret to very low noise is a combination of large diaphragm size and high polarising voltage - this one from B&K using a 1 inch diaphragm and 200V.

The Neumann TLM103 also has large diaphragm and low self noise ( equivalent to 7dBA SPL) for a lot less money. It's a cardiod so totally flat but maybe close enough.

.... Phil

** Should be:

" It's a cardiod so NOT totally flat but maybe close enough."

.... Phil

Here is an interesting project using digital-output MEMS microphones:

Yeah, I was envisioning a lot of small microphones arranged as a parabolic surface or something. I guess a flat array can work like an actively-scanned radar, with some signal processing. If the array can be highly directional, s/n will get better.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

--
~ Adrian Tuddenham ~ 
(Remove the ".invalid"s and add ".co.uk" to reply) 
www.poppyrecords.co.uk

That's pretty high

In my application I am measuring noise on a motor. The motor has a control box which is coupled to the motor so all sorts of frequency modes are present

I will try with the cheap microphone, the ECM 8000

I think I can get very good measurements, since the acoustic noise will be linked to the RPM of the motor, or rather the pole frequency in the motor control.

I can bring out a signal which is syncronized to the pole frequency and then do synchronous detection on that signal and the microphone signal, so the white/pink noise of the microphone "should" disappear :-)

Cheers

Klaus

I note here the regulator labeled 78L05 is probably supposed to read

78L09.

I thought so, too. However, notice that there's a resistor between the 78L05 "ground" leg, and the actual ground. The resistor value isn't specified. This resistor would "see" the 78L05's quiescent current, and lift the 78L05 reference point above ground. Choose the right value and you'd get 9V above ground at the 78L05 output.

Seems like a strange approach to me - possibly done so that the design could use a less expensive part (78L09 is less common than 78L05, I think)?

Dunno that I like it, if indeed I'm understanding it correctly... I think the regulation stability and noise performance might be compromised.