A cheap sensor network: can a MEMS barometer catch an event?

Library → A cheap network

A professional station is expensive. HERD's idea is to turn the approach on its head: instead of a few golden instruments, deploy thousands of penny ones and win through density. The million-dollar question: will a cheap sensor catch anything meaningful at all?

What cheap sensors can already do

Modern MEMS barometers — the same chips used in smartphones and drones to determine altitude — measure pressure with a resolution of a few pascals and cost only a few dollars.1 Amateur and citizen networks already catch serious events: the wave from the Tonga 2022 eruption was recorded by thousands of household barometers worldwide,2 and the citizen seismo-acoustic nodes Raspberry Shake & Boom have long been picking up distant phenomena.3

And "cheap" doesn't mean "imprecise". The open-source Gem logger and the infraBSU sensor cost a small fraction of a professional station, yet they are independently calibrated against a reference at the Sandia laboratory and record data in the field for months on ordinary batteries.5 6 And when Tonga erupted in 2022, the amateur Raspberry Boom network independently captured the event — a peer-reviewed study confirmed that penny-priced citizen nodes catch a global phenomenon.7 What's more, low-cost small-aperture arrays measurably improve infrasound monitoring (Azores study, Jesus et al., 2024)8, while inexpensive mobile MEMS platforms like the KNMI INFRA-EAR / mini-MB make field geophysics affordable (Den Ouden et al., 2021).9

A single cheap sensor is a toy. A thousand connected cheap sensors are an instrument.

Why "numbers" is a strength

Density beats precision. With many nodes, a real event shows up at once on dozens — this sharply raises confidence.
Geometry. A multitude of spaced-out points = a huge "array" that determines the wave's direction and speed.
Noise rejection. Local wind at one sensor is uncorrelated; a real wave arrives at all of them in step.

In fairness: this is the project's main risk

The whole project rests on the assumption that a dense network of cheap barometers really catches meaningful events — and tells them apart from weather fronts, which also produce a coherent pressure change at many stations. This is the riskiest and most important stage. Proving it before a mass launch is our priority №1, otherwise we are selling a promise, not an instrument.

What could go wrong and how we test it: false triggers from atmospheric fronts, insufficient sensitivity of the cheap chip to weak events, calibration drift. The answer is not faith but data: a pilot network, cross-checking against a reference monitor, and open statistics of "caught / missed / false".

In fairness: the limits of cheap sensors

Cheap sensors have limits: a Raspberry Shake & Boom test on elephants showed that quiet vocalizations can be lost in the sensor's self-noise (Lamb et al., 2021).11 That is exactly why calibration and network density matter.

What the network hears — and what it doesn't

Our nodes reliably catch the infrasound of large earthquakes and eruptions (0.02–5 Hz) and global pressure waves of the Tonga-2022 class — the latter arrive hours ahead of the meteotsunami they generate, and are invisible to seismic systems. What a cheap barometer does not catch is the "pure" tsunami wave in the sub-0.01 Hz (millihertz) band: that needs a calibrated reference instrument (SAYA/Hyperion), which we place at the center of the network rather than pretending every $25 node can do it. Density and the neural network raise sensitivity and reject false alarms from weather fronts — but they do not replace the reference in the mHz band, and they do not invent a signal below the physical noise floor.

Did you know?

The barometer in your smartphone (it counts floors for navigation) is a close relative of the sensor that hears infrasound.
Google assembled the world's largest earthquake detector from millions of Android phones — exactly the "win by numbers" logic.
A network of smartphones already works as a planetary seismograph: Android phones detect and warn of earthquakes worldwide (Allen, Stogaitis et al., 2025) — a direct analogy to our cheap-sensor bet.10
The noise of a sensor array falls as √N: ~100 cheap sensors give roughly a tenfold improvement in sensitivity.

Why this matters for HERD

This is the heart of our engineering. We are building a node on an affordable MEMS barometer and testing it in the field against a reference. See our sensor →

Sources for this article

These sources are part of the full HERD library — 272 vetted sources, with meaning search and topic filters.

organization Bosch Sensortec. BMP388 — high-accuracy MEMS barometric pressure sensor. bosch-sensortec.com
peer-reviewed Matoza R.S. et al. (2022). Global seismoacoustic observations of the January 2022 Hunga eruption, Tonga. Science 377. science.org
organization Raspberry Shake & Boom — citizen seismo-acoustic sensors. raspberryshake.org
peer-reviewed Mayer S. et al. (2020). Performance of an operational infrasound avalanche detection system. SLF. slf.ch
peer-reviewed Anderson J.F., Johnson J.B., Bowman D.C., Ronan T.J. (2018). The Gem infrasound logger and custom-built instrumentation. Seismol. Res. Lett. 89(1). doi.org
peer-reviewed Marcillo O., Johnson J.B., Hart D. (2012). An inexpensive low-power low-noise infrasound sensor (infraBSU). J. Atmos. Ocean. Technol. 29(9). doi.org
peer-reviewed Clive M.A. et al. (2024). Crowdsourcing human observations expands volcano monitoring (Raspberry Shake & Boom, Hunga 2022). Commun. Earth Environ. 5. doi.org
peer-reviewed Jesus M.C. et al. (2024). Low-cost small-aperture array improves infrasound monitoring in the Azores. Pure Appl. Geophys. 181. doi.org
peer-reviewed Den Ouden O.F.C. et al. (2021). The INFRA-EAR: low-cost mobile platform for geophysical monitoring (KNMI mini-MB). Atmos. Meas. Tech. 14. doi.org
peer-reviewed Allen R.M., Stogaitis M. et al. (2025). Global earthquake detection and warning using Android phones. Science 389(6757). doi.org
peer-reviewed Lamb O.D. et al. (2021). Assessing Raspberry Shake & Boom sensors for recording African elephant vocalizations. Front. Conserv. Sci. 1:630967. doi.org
peer-reviewed Grangeon J., Lesage P. (2019). A robust, low-cost and well-calibrated infrasound sensor for volcano monitoring. Journal of Volcanology and Geothermal Research 387. doi.org
organization Slad G., Merchant B.J. (2021). Evaluation of Low Cost Infrasound Sensor Packages. Sandia National Laboratories technical report (OSTI). doi.org
peer-reviewed Bowman D.C., Lees J.M. (2015). Infrasound in the middle stratosphere measured with a free-flying acoustic array. Geophysical Research Letters 42(24). doi.org
peer-reviewed Poler G., Garcia R.F., Bowman D.C., Martire L. (2020). Infrasound and Gravity Waves Over the Andes Observed by a Pressure Sensor on Board a Stratospheric Balloon. Journal of Geophysical Research: Atmospheres 125(16). doi.org
organization InvenSense / TDK (2024). ICP-10111 - high-accuracy low-power MEMS barometric pressure sensor. invensense.tdk.com (datasheet / product page). invensense.tdk.com