IMP23ABSU Ultrasonic MEMS Microphone

[Ivan Bravo]

Hello,

I am attempting to build an ultrasonic microphone for predictive maintenance applications. I have produced a flex circuit per ST's Application Note: 5522. I have not, however been able to pick up ultrasonic signals. What sort of output(voltage range) am I supposed to expect if I want to pick up ultrasound? I have built the circuit recommended on their datasheet, but the signal levels remain real low. Is this expected? Is a high pass filter required? Can anyone provide additional guidance?

[Peter BENSCH]

What kind of opamp did you use?

What is the value of R at the non-inverting pin of the opamp?

What is the value of Vref?

Where did you measure the signal levels?

Regards

/Peter

[IBrav.1]

Hello,

What kind of opamp did you use?

• LM358AN

What is the value of R at the non-inverting pin of the opamp?

• 500 ohms

What is the value of Vref?

• Vref is about 1.64 (about half of 3.3V)

Where did you measure the signal levels?

• I measured the signal levels at the output of the Op Amp

• R1 is about 160 Ohms
• R2 is currently 172 kOhms

My current setup the flexboard connected to a breadboard op amp circuit. I am using an oscilloscope to measure the output of the op amp.

Regards,

[Peter BENSCH]

There are a few things you should do better:

• the LM358 is a bipolar opamp designed for higher supply voltages. Even if you can use it with a 5V single supply, it has about 2V voltage drop and fits only for small signals, with a supply of 3.3V it is no longer useful. Definitely a better choice would be a low voltage rail-to-rail CMOS opamp with enough margin in bandwidth, e.g. the TSV912 or TSV912A.
• The resistor R at the non-inverting input is much too small and creates a huge load for the output resistance of the IMP23ABSU, for which the datasheet gives min 15kohms.
• If a lower frequency limit of 100Hz is OK for you, then please use
  • for C1 a ceramic capacitor (MLCC) of about 0.1µF (can be even smaller for a frequency limit of e.g. >1kHz)
  • for R a resistor of 47...82k
• Please reduce the feedback from factor 1075 (way too high) to e.g. 25...30, i.e. to R1=1.8k, R2=47k

Good luck!

/Peter

[Ivan Bravo]

Hello,

I appreciate the feedback.

the LM358 is a bipolar opamp designed for higher supply voltages. Even if you can use it with a 5V single supply, it has about 2V voltage drop and fits only for small signals, with a supply of 3.3V it is no longer useful. Definitely a better choice would be a low voltage rail-to-rail CMOS opamp with enough margin in bandwidth, e.g. the TSV912 or TSV912A.

• I have ordered the TSV912 and should be receiving this in about a week.

The resistor R at the non-inverting input is much too small and creates a huge load for the output resistance of the IMP23ABSU, for which the datasheet gives min 15kohms.

If a lower frequency limit of 100Hz is OK for you, then please use

• for C1 a ceramic capacitor (MLCC) of about 0.1µF (can be even smaller for a frequency limit of e.g. >1kHz)
  • for R a resistor of 47...82k
• R has been changed to 56k.
• C1 has been changed to .1uF MLCC

Please reduce the feedback from factor 1075 (way too high) to e.g. 25...30, i.e. to R1=1.8k, R2=47k

• R1 has been changed to 3.3k, and R2 to 102k; feedback is now about 32

With the changes mentioned, and a "Dog Whistle" smartphone app outputting sound, I was able to measure a frequency of 19kHz at about 135mV p-p. Is this the type of behavior expected from the MEMs microphone?

Could I expect a larger p-p voltage measurement using the TSV912?

Up to what frequency could one realistically expect to accurately pick up using the IMP23ABSU?

Thanks,