Skip to content
←All work

Industrial & IoT

A soil sensor that had to last a season on one charge

A seed-stage agritech startup

3 wks → 9 mths

Battery life

14 weeks

Design to pilot

80%

BOM lines second-sourced

The problem

The founders had a working prototype built around a development board. It measured soil moisture accurately, but it drained a battery in under a month — and their entire pricing model assumed a device that field staff would visit once per growing season.

What we did

  1. 01

    Measured actual current draw across a full duty cycle rather than trusting datasheet figures, and found the radio was holding the rail up far longer than the firmware suggested.

  2. 02

    Reworked the architecture around a low-power microcontroller with a properly gated sensor rail, so the analogue front end draws nothing between readings.

  3. 03

    Rewrote the transmit path to batch readings and wake the radio on a schedule instead of per-sample.

  4. 04

    Selected the radio module for availability as much as performance — the first-choice part had a lead time that would have delayed the pilot by a full season.

Outcome

The redesigned unit met the eight-month target with margin, and the pilot deployment went out on the original schedule. The second-source radio identified during design was used for the production run when the primary went on allocation.

A worked example: a representative problem of this type, and the reasoning we would apply to it.

Next project

A diagnostic front end with a noise floor problem

→