MEMS microphones and contamination

jja · ‎2014-11-25

Posted on November 25, 2014 at 19:01

Hello!

I have a very specific question about MEMS microphone and noise induced by contamination. Refering to http://www.st.com/st-web-ui/static/active/jp/resource/technical/document/application_note/DM00103202.pdf, it is stated in chapter 2.1.1 that microphones can exhibit a high SNR or high noise output if contaminated. The high SNR I can understand if for example the MEMS membrane get sticky with say soldering flux. But what is the mechanism for just high noise at the output?

The background to my question is that I'm developping an acoustic camera using 32 MEMS microphones. I got as far as ordering my first prototype PCB and everything worked fine at first but after while I started to see very high intermittent noise at the outputs of some of the mics. The noise comes and goes with time but has gotten worse over time. I confirmed that the phenomena is actually high random noise by recording all the mic signals in a quiet environment and compared the recordings to normally working mics. Then I ran an FFT on all the signals and for the high noise one's I could see an strong otherwise quite white noise shape but also very strong near DC components (around 3Hz).

After this I looked at the sound-inlets through a microscope and saw that some of the inlets to the noisy mics were completely clogged with soldering flux. Could this then alone be the reason for the intermittent noise?

My logic says that if the sound inlets and the mic is contaminated with flux I would only get a very insensitive mic but not higher ''standby'' noise levels?

BTW, I also checked the integrity of all the clock signals etc. and they looked fine.

I really appreciate if any one would know if this is my problem alone.

Thanks in advance,

Jonas Nyberg

#snr #noise #mems #contamination

Alessandro MORCELLI · ‎2014-12-01

Posted on December 01, 2014 at 11:42

Hi Sir,

your understanding is correct. If the sound inlet is clogged with some material, the failure mechanism would be the lower sensitivity of the microphone. But, as a matter of fact, the microphone inlet has been contaminated with a flux and so is not possible to exclude that this foreign material is also contaminating the MEMS sensor. The contamination of the MEMS can cause very unpredictable failure modes since depends on the dimension, the weight and the position of the particles. According to my experience i saw some contaminants causing high noise only in presence of sound or in an intermittent way. This is the reason why ST strongly recommends to avoid that contaminants enter into the microphone package.

These kind of issues are very common when the soldering of the microphones is done by inexpert PCB assemblers or not skilled with this devices.

Regards,

Alessandro

jja · ‎2014-12-01

Posted on December 01, 2014 at 18:22

Thanks for your answer Alessandro.

Do you know if there's any chemical that I can use to clean the contaminated mics that I have? Like some alcohol-based that would dissolve the flux but wouldn't hurt the MEMS-membrane or ASIC inside?

Jonas

Alessandro MORCELLI · ‎2014-12-01

Posted on December 01, 2014 at 18:44

Hi Jonas,

actually i would not recommend to apply any chemical. Unfortunately i regret to inform you that once the MEMS is contaminated with a substance like the flux, the device is definitely compromised.

Regards,

Alessandro

jja · ‎2014-12-11

Posted on December 11, 2014 at 11:56

Ok, thanks again for your reply.

One last question, I'm just wondering why the microphone would be damaged if I try any chemical. Isn't the MEMS membrane made of some kind of silica which would be quite chemical resistant right? Or perhaps it will damage some part of the ASIC?

Thanks,

Jonas