I've been looking around for a sensor that is capable of detecting the presence of a person's breath. It needs to fit inside a tube that a person will blow on. Detecting the force of the person's breath is not important at all, I just want to detect movement of air caused by someone blowing on the tube.

Is there a sensor that can help with this that will run on a 3.3V circuit? I can't seem to find something online, that's small enough to fit inside a tube around just below 1" in diameter.

Any recommendations. I've been searching catalogs for our local vendors till I'm blue in the face, but just can't seem to find something suitable.


2 Answers 2


You could try using a thermistor that is powered sufficiently to raise its temperature to significantly above "breath" temperature - the air passing will have a cooling effect and this will be seen as a change in resistance.

Use a constant current feed and a positive temperature coefficient thermistor or RTD. If it's a 100 ohm device it'll be about 169 ohms at 100degC and you'll need to experiment a little bit with the current feed to get it in the vicinity of (maybe 80degC).

As colder air crosses the thermistor it will cool and lower its resistance - this will result in a change of voltage across the device's terminals and this could be detected by a simple comparator circuit with a little hysteresis and a low pass filter on the reference input. Here's an idea: -

enter image description here

Current into the thermistor is controlled by R and Vx - current is \$\dfrac{V_X}{R}\$

Vx can be tweaked by altering the potential divider on the op-amp on the left.

  • \$\begingroup\$ This is a great answer, haven't thought of that. It would increase the power consumption budget though, and my apologies, I didn't mention the circuit is running on a 9V or some AA batteries. \$\endgroup\$ Feb 13, 2014 at 20:23
  • \$\begingroup\$ @josef.van.niekerk you'd probably get away with 20mA and that is nearly do-able on a 3V3 supply (probably 60mW to the thermistor) - it'll warm a bit above ambient and that might be enough. What's your power budget? \$\endgroup\$
    – Andy aka
    Feb 13, 2014 at 21:54
  • \$\begingroup\$ We're building a novelty device which measures alcohol breath concentration, and the device only needs to last at least a couple of days. We're planning on using 1x 9V or possibly 2 to 3 AA batteries, nothing more, as weight is also a factor. The toy is going to get thrown around by delegates... \$\endgroup\$ Feb 13, 2014 at 22:51
  • 1
    \$\begingroup\$ @ScottSeidman spot on with your analysis. Without hysteresis it's also known as a data slicer as used on the back end of an FM demodulator (or any data that has passed thru a "channel" that doesn't have a dc value i.e. scrambled or manchester encoded). The filtered signal tracks the average and the hysteresis is just to stop it going unstable when breathing ceases. \$\endgroup\$
    – Andy aka
    Feb 13, 2014 at 23:18
  • 1
    \$\begingroup\$ @Scott - yeah it might be a bit hit and miss with no particular positive and negative temperature to drag it to a value that definitely isn't body temperature. I did mean a bit above body ambient - maybe run it at 50C. \$\endgroup\$
    – Andy aka
    Feb 13, 2014 at 23:20

A wide range of pressure sensors exist that are suitable for this purpose. The sensor may be larger than you specify but can monitor a smaller tube via a hole or a pressure sensing pipe.

You say it needs to be a certain size but do not explain why - knowing such things can help with solutions. eg it may be that absolute size matters. But it may be that it can be larger as long as the tube itself is small enough.

Here is an example of a barometric sensor from Sparkfun and another here.

Here is the datasheet for the Bosch BMP180 barometric sebsor used in the first unit above.This is in an LGA package with a metal lid and measures 3.6mm x 3.8mm x 0.93mm tall. Small enough for your application, I imagine :-)

Pressure range is 300 to 1100 hPa (hecto-Pascal), corresponding to +9000m to -500m relative to sea level pressure. Absolute accuracy is about 1 hPa typical and 2 hPA worst case if your breath is not freezing. (About 4.5 hPa worst case at -20C).

They say:

  • The BMP180 consists of a piezo-resistive sensor, an analog to digital converter and a control unit with E2PROM and a serial I2C interface. The BMP180 delivers the uncompensated value of pressure and temperature.
    The E2PROM has stored 176 bit of individual calibration data. This is used to compensate offset, temperature dependence and other parameters of the sensor.

  • UP = pressure data (16 to 19 bit)

  • UT = temperature data (16 bit)

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.