MEMS Seismometer Network
What it measures
Ground acceleration in three axes. Peak ground acceleration (PGA), spectral response, and site amplification characteristics.
A low-cost seismic node for dense urban monitoring
Professional seismometers are accurate but expensive — which is why the Kathmandu Valley has so few. A MEMS (micro-electro-mechanical systems) accelerometer puts a tiny, robust motion sensor on a chip for a fraction of the cost, accurate enough that a dense network of them can map how shaking varies street by street. Each node is GPS-timed and Raspberry Shake–compatible, so its data joins the global seismic network.
From a chip to the seismic network
1 · ADXL355 MEMS sensor ▸
A tiny suspended mass inside the chip shifts under ground acceleration; the change is read out as a low-noise digital signal on three axes.
2 · GPS timing ▸
Earthquake location depends on comparing arrival times across stations, so every sample is stamped against GPS time to sub-millisecond accuracy.
3 · Raspberry Pi streaming ▸
A Raspberry Pi digitises, buffers, and streams the data continuously — offline-first, syncing when a connection is available.
4 · Joining the network ▸
Raspberry Shake compatibility means the data flows into the global seismic network and standard analysis tools out of the box.
Why the sample rate sets what you can see
A digital recorder can only faithfully capture motion up to half its sampling rate — the Nyquist frequency. Choosing the sample rate is choosing the band of ground motion the node can resolve:
This node is the building block of HICS's Seismology research programme — a dense MEMS network to map how the Kathmandu Valley's soft sediments amplify shaking, where fewer than five continuous stations exist today.
Planned. The node design and prototype come first; deployment follows. The goal is the first dense, openly published seismic dataset for the valley.