Please enable JavaScript to view the comments powered by Disqus.


Tornado Probe


Design a probe capable of a direct hit from a tornado that is based on commonly available hardware

Learn from past attempts at designing probes and improve in previously discovered weaknesses

Project Lead

Ray Bohac & Kevin Barton

Project Status

In development & testing

Phase 1 completed

Phase 2 expected early March 2015


We began experimenting with an Arduino Uno platform as a new base system. Several sensors were tested including:

  • Bosch BMP085 Pressure Sensor
  • DHT22 Pressure & Temperature Sensor
  • ADXL326 triple axis gyro
  • SIM908 Quadband SIM
  • Adafruit Ultimate GPS Logger Shield
It was a mess but it worked. This was the first wire-up on the new probe. Don't worry, the spaghetti is gone now

It was a mess but it worked. This was the first wire-up on the new probe. Don't worry, the spaghetti is gone now

With this platform we were able to successfully log Pressure, Temperature, Humidity, Altitude (via pressure.. probably not useful),  and GPS Location to an SD card. We added a SMS module to the system and via text message to a Twilio service we were able to track this data and overlay it on a Google map.

While purchasing additional hardware we ran across a miniaturized hardware platform called the TinyDuino which is manufactured by Tiny Circuits. It happened to be that they are only about 2 hours away so we reached out to them.

After discussing the goals they were able to quickly come up with ideas for us that will harden the setup and hopefully reduce the likelihood of failure due to damage or lost units.

My background is in software, not hardware, so through Tiny Circuits we've engaged with Roy Stevens at Henway Technologies to design the release product. 

Once completed, the entire project will be open sourced on this site. Any feedback is welcomed on this initial design, Field testing will be in the April/May 2015 time-frame.

New Plan - The new design will be based on the Arduino framework, and specifically we will be using the commonly available TinyDuino platform. 

The following features are planned:

  • Small form factor: Goal is the size of a ping pong ball or slightly larger
  • On board storage instead of using a SD card 
    • Reason: High likelihood that a SD card would come loose causing data corruption or loss
  • Low power Bluetooth data transfer to an IPhone application
    • Will allow for immediate data transfer
    • Also if a probe is in a tree or down a cliff we should be able to retrieve data without fully retrieving
  • Backup data transfer via USB
  • RF based location. Will transmit fast and slow beacon down over time to save battery life. Plan is 2-3 days worth of battery availability
  • GPS capable of 10Hz
  • 9 Axis Gyro/Accelerometer
  • Pressure Sensor
  • Temperature Sensor

Phase 1 (Complete) - Probe Recovery: Experiment with COTS long range transceivers in different environmental scenarios - on the ground, in the mud, floating in water, in a tree, under debris, etc. The goal here is to prove that we can successfully retrieve data from a tornado probe from long distances in a variety of post-tornado scenarios.

Test: We'll simply place an arduino, the transceiver, and a large battery in a waterproof case and set the device to transmit once per second. We can receive the transmissions with another radio and an FTDI cable into a laptop. The transmission data can be the RSSI signal strength.

Result: Using the setup described above, with a Yagi antenna, 2 mile range achieved. Transmitter was placed in a ravine and receiver received a data signal at a 2 mile distance through several neighborhoods and across a power distribution area.

This is the phase 1 prototype (more images are below)

This is the phase 1 prototype (more images are below)

Phase 2 (March 2015) - Enclosure Research: GPS, 9-axis gyro, temperature, etc. are simple sensors that can be mounted and potted in any sealed enclosure. More thought needs to go into mounting the pressure sensor, humidity and the antennas, which may need access the outside of the enclosure and compromise our perfect waterproof seal (If anyone has ideas please comment below). We can experiment with enclosures using actual sensor parts and software on a TinyCircuit platform. Any software written for this experiment can be reused in the final version of the product; saving time and money.

Test: Build mock-up enclosures and mount sensors. Write software to continuously log sensor data and expose enclosure to poor environmental conditions. Extract data (in real time?) from the device to discover the moments of failure. 

Phase 3 (April 2015) - Custom electronics: Based on the results of Part 1 and Part 2, common hardware comfugured on custom board for a more durable and convenient system. This includes adding things like bluetooth and inventing multiple ways to extract data in the case of an emergency.