How we built a smart sensor for a sport that had never had one

The idea from Ve2Max sounded simple: put a sensor on a padel racket and turn its movement into useful feedback for the player. The engineering was not.

The hardware had to survive repeated impacts, capture motion fast enough to distinguish different shots, run on a small battery, stay connected to a phone on a busy court, and keep the session intact when Bluetooth inevitably dropped.

The first challenge was making the hardware disappear

The sensor had to fit inside the racket's butt cap, survive repeated impacts through a match, and run long enough on battery power that the player did not have to think about the electronics.

We built around a Nordic nRF52840 and a 6-axis motion sensor sampling at 230 times per second. Zephyr RTOS handled the embedded firmware. One less visible problem was power-state management: factory mode, shipping mode, and active play each needed different behaviour, without keeping the device awake when it should be sleeping.

Bluetooth behaved differently on court than it did on the bench

A sensor swinging at the end of a racket, a phone in someone's pocket, and a busy court make for a very different Bluetooth environment than a lab bench.

The goal was to stream the full raw sensor feed live so the heavier analysis could run beyond the device. We added detailed logging, integrated Sentry crash reporting, and iterated on reconnection behaviour across Android and iOS.

We designed for the connection to fail

Rather than assume Bluetooth would stay connected for an entire match, we built the data path around interruption. Short drops are buffered in RAM, longer ones can fall back to flash storage, and stored data can be replayed after the phone reconnects.

That architecture was designed to preserve complete sessions even when the wireless link wasn't continuous, so the analytics pipeline didn't have to depend on a perfect Bluetooth connection.

Motion data only matters if you can turn it into a shot

A six-axis sensor gives you acceleration and rotation. It does not tell you whether the player hit a lob, a flat drive, or a slice. That classification had to be learned from real movement.

We worked with the PadelPlay and Ve2Max teams on data collection and calibration, including sessions with professional players.

That is the useful part of keeping the raw-data pipeline intact: new match data can continue feeding the analysis rather than locking the product into whatever the first firmware release could recognise.

The team is small and international, with developers in the Netherlands and Brazil working closely with the PadelPlay and Ve2Max teams on product and go-to-market. Gen2 hardware is already prototyped, with a new sensor chip that reads more accurately. There is a lot more to build.