How to test contact vibration sensors.

I forgot to say that the mass of the transducer is 6 grams.

Maybe an online frequency generator...

https://meettechniek.info/additional/additive-synthesis.html

Then you need an accelerometer...

https://revibeenergy.com/measuring-vibrations-with-accuracy/

You could buy one...

https://www.spectraplus.com/?gclid=Cj0KCQiAzMGNBhCyARIsANpUkzMQAG6ci7ab8ouoErSTFClDo4SsXBP2kaVHleutPz77kGjjSpxsYjwaAqiIEALw_wcB

...or build one....

A laser interferometer can be used to measure vibration amplitude which you can use to calibrate your vibration sensor.

https://www.researchgate.net/publication/267971389_Interferometric_vibration_displacement_measurement

Good idea, but if you plot the sensor response vs speaker driving voltage on the XY scope (if I am understanding you correctly), to calibrate the sensor, you must assume that the response of the speaker is flat across frequencies.

If you are interested in "DIY" testing then read the other responses and ignore this one.

If you are interested in professional level testing with NIST cert traceability in the test report then google "MIL-STD-810 vibration test lab". You probably have one or two in Austin and a dozen or more in Texas.

You can download MIL-STD-810 for free (legally) from http://quicksearch.dla.mil. This isn't a very exciting read but it has lots of words and phrases that you can you can use to trick Mr Google into giving you useful information.

I'd be curious to know what the sensor manufacturer says about how they calibrate their product.

How many axes of motion does this sensor measure? Is the speaker motion in line with the axis of detection?

I don't like gluing any sensor to anything but its final operational location. Gluing it to a test jig means you might damage the sensor when trying to remove it from the test jig.

Speakers are designed to move just air and not a 6+ gram mass attached to the speaker cone. I'm very nervous this added mass may cause the cone or dust cap to flex or distort making calibration difficult if not impossible without some sophisticated, scientific instrumentation. Then there's the possible dampening effect an easy-to-remove glue might add to any theoretical kinematic analysis.

I am manufacturing them, one by one.

Yes, the cone's motion is in line with the axis of detection. Actually it is more a flat surface then a cone.

https://www.parts-express.com/Tectonic-TEBM46C20N-4B-BMR-3-Full-Range-Speaker-4-297-2157

The surface appear to be very light and stiff and I believe it has an honeycomb structure inside.

I said glue to simplify. In reality I am using a tiny amount of mounting putty by the Gorilla brand.

The speaker is not inside an enclosure, just mounted on a vertical piece of plywood. Good? Bad? I am driving the speaker with one channel of a NAD 2020A integrated amp.

Okay. This gives us some data to work with.

I'm still apprehensive you may get piston distortion and/or speaker damage from adding this mass. The sensor is just over 2.6 times more mass than the moving mass of this speaker. This will lower the resonant frequency and make the speaker less efficient. Early testing should regularly monitor how hot things get.

The key problem will be determining how far the piston moves with frequency and power level. If the top of the mounted sensor extends above the rubber surround then some simple optical measurements can detect how far things have moved. I envision an optical laser (cat toy) mounted on a sliding caliper with the beam parallel to the piston surface. Simply looking for the elevation when the light just misses the sensor can give one the difference between no motion and motion. The rest is kinematic calculations.

Thank you all for your ideas. I am going to investigate the laser interferometer.

I have an idea which is simple to test but I an afraid some of you will laugh..

I was thinking to place a condenser mic in front of the speaker and amplify its signal to the point of lighting an LED. Then, in front of the LED, put a photocell. All inside a tube.

The changing of resistance of the photocell could be used to change the amount of signal going to the amplifier. This servo feedback is very crude, maybe could be realized more elegantly but I didn't find much on the web.

BTW, during my searches i came across something called Vactrol which is the ancestor of our modern photocoupler. Still used by some for analogue sound effects

The condenser mic might not be a bad idea if you have a calibration curve for the mic.

What you describe might give you a relative response curve, but you still need to figure the speaker displacement given sound level for calibration. Your top frequency (12kHz) has a wavelength of about 1 inch, so you also need to worry about reflections.

I'd go with the interferometer, especially if you can find a laser diode with a built in photodiode as described in SE post #4. The number of fringes per half cycle gives pk-to-pk displacement, Xpp, given the laser wavelength λ.

The displacement amplitude D = Xpp*λ / 4.

I'm assuming your sensor measures acceleration.

If f is the test frequency,

Velocity V = 2*pi*f*D, and acceleration A = 2*pi*f*V.

So, for any frequency f, you can calculate acceleration from fringes and then associate acceleration with sensor output.