MEMS speakers are one of the biggest innovations in speaker technology in recent years. Offering full range performance with compact size and low power, they are rapidly being adopted for use in devices such as earbuds, hearing aids, smart glasses and more. With SoundCheck you can make exactly the same MEMs speaker measurements as you can with conventional mechanical speakers. Watch this short video where we demonstrate frequency response, impedance, and distortion measurements on the xMEMS Montara MEMS speaker.
MEMS Speaker Measurements
We’d like to thank Michael Ricci, Sr. Director of Electroacoustic Engineering at xMEMS for the technical guidance on Piezo-MEMS transduction.
Learn More About SoundCheck’s Advanced Features
Read more about more measurement features in SoundCheck.
More information is also available in the SoundCheck Manual.
Video Script: MEMS Speaker Measurements
SoundCheck is one of the most widely used loudspeaker and microspeaker measurement systems in the world, but did you know that it can also measure MEMS micro-speakers? MEMS micro-speakers are rapidly becoming popular for devices such as hearing aids, earbuds, smart glasses and more as they offer full range performance with compact size and low power, and they are also SMT reflowable. They’re constructed in an entirely different way to conventional miniature speakers – rather than using inductive coils and magnets, they use a voltage driven capacitive actuator to provide full range performance.
I’m going to demonstrate a MEMS micro-speaker test using the xMEMS ‘Montara Plus’ full-range Piezo-MEMS microspeaker, that uses a monolithic solid state fabrication. These devices are entirely manufactured with MEMS processes in a semiconductor wafer foundry. When you’re testing these devices, the xMEMS provided driver circuit delivers Voltage bias and boost converter to step up the voltage as piezo-MEMS devices have a very high input impedance and draw very low current.
Here, I’m going to use xMEMS’ own charge amplifier. You’re also going to need to build the speaker into an earbud or make your own test jig in order to test it. I’m going to demonstrate using this test jig, which is actually the one that xMEMS uses for their own measurements, and we’re going to put an ear simulator coupler on it to simulate an in-ear measurement. Aside from that, the test setup’s very similar to what we would use for any other speaker. We have an AudioConnect 2 interface which will power the coupler, and that’s connected to SoundCheck for analysis.
So we have a test sequence that will play the stimulus and analyze the response. You won’t hear it as it’s all in the coupler. And here we can see the results.
Let’s start with the frequency response. You can see it has a very flat response at low frequencies, and then in the higher frequencies you have a resonance due to the piezoelectric material and the resonance of the coupler.
We can also look at the impedance. You can see here that it’s a very different shape from a conventional loudspeaker impedance. The values are much higher but it’s very linear, which makes it easy to compensate for.
We can also look at distortion. The total harmonic distortion is also very linear right up to where we get into the ear canal response.
And while we’re on the subject of distortion, I just want to use the measurements on this device to highlight the importance of using frequency normalized distortion measurement.
With this conventional distortion measurement, you can see the second and third harmonics plotted at their actual measured frequencies, along with the fundamental.
Frequency Normalized distortion measurement compares the harmonic levels to the fundamental level at their measured frequency before their ratio is plotted, rather than the fundamental level at the excitation frequency. This removes the effect of the non-flat frequency response from the distortion and makes it easier to see the peaks in the distortion response independent of the peaks and dips in the fundamental response. Here, you can see both regular THD, the orange line, and normalized THD, the blue line. And as you can see, you have a high Q here at resonance, but apart from that there is very little distortion, so you can focus your efforts on planning around this peak. If you were going by conventional distortion, you could be wasting your time trying to solve resonances you don’t have with this second bump on the graph here.
So that’s piezo-MEMS speaker measurements in a nutshell. Check out our website for more information on testing MEMS speakers, or if you want to learn more about normalized distortion measurement.