Crazy microphone calibration ideea

[mr_push_pull]

Hello everybody

I've had the ECM8000 for a while, but I really distrust its performance as a measurement mic.

While searching for mic calibration services, I remembered about a discussion on diyaudio concerning DIY methods for calibrating.

Here's my derived ideea. If one could come up with an easily replicable naturally-generated sound, like... 2 cm iron ball falling on flat iron surface or something... and the time response of that, measured in well known conditions would be known...

What do you think?

[adason]

you are absolutely right
your ideea is crazy

[joe carrow]

There's already a method, I believe it's the exploding wire method. A known voltage and current are discharged through a known small-gage wire using a well known circuit. This ensures that a precise quantity of energy is dumped into the system in a well known fashion. Wires can hold very fine size tolerances, and can have very well known compositions.

In order to be valid for calibration purposes, it would probably need to be performed in a large anechoic room. For you or me, that would mean suspending the test rig high in the air, away from any walls or ceilings. It also means that there would be a low frequency where data would become invalid.

Having said that, the ECM8000 really is fairly reliable for most uses. It has good quality control, low variability, and measures very closely to far more expensive microphones. The main place it falls short is at extremely high SPL.

What don't you trust about the mic?

[adason]

Having said that, the ECM8000 really is fairly reliable for most uses.

exactly!
I have ecm8000 and find it more than adequate for any speaker design work

[CeramicMan]

How about random noise from a pressurized CO2 canister? The continuous nature of the sound, which is akin to a step function, would supply a lot more bass energy compared to an impulse.

[soongsc]

Originally posted by mr_push_pull
Hello everybody

I've had the ECM8000 for a while, but I really distrust its performance as a measurement mic.

While searching for mic calibration services, I remembered about a discussion on diyaudio concerning DIY methods for calibrating.

Here's my derived ideea. If one could come up with an easily replicable naturally-generated sound, like... 2 cm iron ball falling on flat iron surface or something... and the time response of that, measured in well known conditions would be known...

What do you think?

I think it's not evident how accurate you are planning to calibrate to.

I have not come across naturally generated sounds that are fully repeatable.

Iron ball falling on an iron surface has lots of problems.
1. If the ball fall on differen points of a surface, the different vibration modes are excited, and material purity also becomes an issue.

2. If the ball fall on the exact same point every time. The surface is diformed differently each time.

3. The ball itself is deformed each time, so regardless whether the contact point is the same ceneter location or not, the actual contact area changes.

...

[joe carrow]

Likewise, the CO2 cartridge will sound different based on the orifice through which the gas is escaping, it will sound different as a function of time (because of the changing gas pressure), and it will be dependent on atmospheric conditions (although this is true for any acoustical measurement).

I think that the exploding wire has a lot going for it- but it's pretty irrelevant because the ECM8000 really isn't that bad, and the sound received by the microphone is so heavily influenced by the surrounding environment that it's probably not going to give any kind of high quality results without an anechoic chamber.

[Ron E]

Just send it to Kim Girardin. He can calibrate it for you. Do a search for Wadenhome sound or his name and you will find contact info.

For s's and g's, I once tried to do the reciprocity method using a small speaker as a "reversible" transducer, but I couldn't get enough signal to noise to make it work. Perhaps some more tweeking and I could have got it to work.

For the reciprocity method, you measure the output of a speaker with the microphone you wish to calibrate and also a reversible transducer. Then you measure the response of the reversible transducer, dump all this data into a formula for each data point and calculate the response of the microphone.

[Salas]

If we take a well known human organism like a member of family and feed him a precise quantity of certain beens or something, will he be repeatable?

[Geoff H]

I have been having problems with repetitive results. My problem turned out to be the photographic tripod I was using. Often showed up as a deep hole at about 200 Hz.

"If we take a well known human organism like a member of family and feed him a precise quantity of certain beens or something, will he be repeatable?"

No, unless you can calibrate the resultant flatulence. Scary idea.

Geoff.