Testing vibration sensors.

[kokkoplus]

I am building electromagnetic transducers for audio frequencies. The design is similar to a geophone, so basically a magnetic mass suspended inside a coil.
How should I measure the response? I want to see what happens when I change parameters like the mass of the magnet, the coil, the suspension, ect.

The transducer is pretty light, about 5 grams and is a cylinder of 13 mm diameter and 13mm tall.

I though of using some double tape to attach it to the cone of a speaker and sweep the whole range of audio sine waves and measuring the output with a software like spectraplus or such. Can you think of a better setup? How do I calibrate? I got some full range speakers with flat pistons thinking they would be better but I am not sure.
Thank you for your attention.

 

[Bo Deadly]

Your test rig should be as close as possible to how it would be used. So if you're capturing audio then presumably you just expose it to audio however it normally would be. For a microphone, the typical rig is one meter away from a monitor speaker. Then just amplify, feed that into a decent USB audio interface and analyze the results with Room EQ Wizard or some other FFT software (my personal favorite is GNU octave but that necessitates writing code which is not for everyone) and then analyze the results.

However, measuring transducers whether they be output (speaker) or input (microphone) is difficult because you generally don't know if the response of the stimulus device is flat. So there are all sorts of techniques for compensating for that. You also have room effects like standing waves / resonance that must be minimized for accurate results. The ideal procedure is to use an anechoic chamber (a room sound deadening on the walls so that it's echo-free) and a speaker or speakers that are known to produce a very flat frequency response. The practical procedure is to use a reference microphone (basically a cheap electret in an expensive aluminum tube) that is known to be very flat from 20-20k and use that to find a suitable speaker, capture the frequency response of that and then use that to adjust the stimulus and or recorded response to be as flat as possible. Now you put the transducer you want to measure in front of you speaker, record at different levels using various stimulus like stepped tones or impulse responses or whatever and then crunch the numbers in software to figure out what your transducer is capable of.

Also, I suspect an additional issue for designing a transducer is what kind of amplifier to use. Usually vibration sensors are piezo's which use a charge amplifier (basically a high gain amp bootstrapped to be very high impedance). But for a mass suspended by a coil, I have no idea what you would need to do for that. You would need to study the impedance of whatever the output is at different frequencies and work it out.

 

[kokkoplus]

Thank you for your response.

I should have been more clear that it is a contact mic. There is going to be a mechanical coupling between the vibrating object and the transducer. I used some mounting putty with success amplifying a classical guitar with the sensor placed on the top of the guitar.

Since the vibrations are so small, i don't think it would make sense to use an industrial shaker, beside the fact that they are super expensive.

The impedance is about 500 Ohms and I get a similar output voltage of a dynamic microphone.

 

[trobbins]

Compare it to a lightweight accelerometer.

 

[abbey road d enfer]

As trobbins suggests, your contraption senses the acceleration it is submitted to.
Actually it delivers a voltage that is proportional to the relative speed of the coil and the magnet.
In the mass-controlled region, the magnet is motionless.
The voltage is proportional to speed, which means it decreases at 6dB/octave with frequency.
A flat-response loudspeaker in its piston range produces flat sound pressure, which is also the result of flat acceleration, so the output voltage of your transducer should be a -6dB/octave line.
However, at low frequency, when the transducer shifts from mass-controlled to compliance controlled, the response shold gradually shift from -6dB/octave to flat to +6dB/octave asymptotically to minus infinity.
This will combine with the loudspeaker, which will shift from flat to +6 to +12 ultimately.
If the compliance of your transducer is high enough (resulting in a lower transition frequency than the loudspeakers') and the mass low enough, it should not interfere significantly with the louspeaker.
Regarding calibration, what uniy do you want it to be measured? I think the only valid data would be in V/m/s². The simplest would be to compare with an existing accelerometer. You may use this
https://www.amazon.fr/DollaTek-anal...keywords=accelerometre&qid=1639306838&sr=8-16Limited to 1600Hz, but cheap.

 

[moamps]

.....
A flat-response loudspeaker in its piston range produces flat sound pressure, which is also the result of flat acceleration, so the output voltage of your transducer should be a -6dB/octave line.
However, at low frequency, when the transducer shifts from mass-controlled to compliance controlled, the response shold gradually shift from -6dB/octave to flat to +6dB/octave asymptotically to minus infinity.
This will combine with the loudspeaker, which will shift from flat to +6 to +12 ultimately.
If the compliance of your transducer is high enough (resulting in a lower transition frequency than the loudspeakers') and the mass low enough, it should not interfere significantly with the louspeaker.

May I add a picture?

The picture shows the acceleration, speed and displacement of the diaphragm of a woofer (normalized values) in relation to the produced sound pressure.
(v(x) is correct up to 150Hz, purple curve mj.mic (x) represents near field sound pressure)

 

[abbey road d enfer]

I forgot to mention that my comments were valid for a loudspeaker in infinite baffle, and wrong for a bass-reflex or any kind of tuned enclosure.