North Pole Engineering Runn Treadmill Sensor
Runn is a small sensor that you add to a treadmill to measure speed. The idea is that it transmits the speed using a protocol that can be received by various applications, including zwift.com, who actually sell the Runn sensor. Most treadmills won't transmit the right protocol, so until now you either need to use a Footpod or rely on the internal accelerometer in your sports watch. The sensor works by using 2 optical sensors that look for stickers on the treadmill belt. It measures the time it takes for the sticker to move from one sensor to the next, which should give a fairly accurate measure of the treadmill belt speed at the time the sticker is under the sensor. Unfortunately, that's where the problem lies, as the speed of most treadmill belts is not continuous, but varies with your foot strike. When your foot lands on the treadmill belt, the belt slows up quite a bit, then as the treadmill compensates it speeds up during the stance phase, and then when your airborne the belt will speed up even more. I exchanged emails with North Pole Engineering, and confirmed that their sensor will be taking the average belt speed, which will result in an error.
Consider my treadmill, where I've performed Treadmill Calibration. I've performed frame by frame measurements across multiple foot strikes to gather actual performance figures. When my treadmill is set to 8 MPH, the average speed of the belt when I'm running is actually 7.17 MPH, which is 10.4% slow. This average represents a minimum speed just after foot strike of 6.6 MPH, rising to a maximum speed of 7.3 MPH just before toe off. While I'm in flight, the speed rises further to 7.6 MPH just before foot strike. The average belt speed while my foot is in contact is 7.06 MPH, which is 1.5% slower than the average belt speed. Does that matter? Well, if you were marathon training and Runn said you'd run 20 miles at 9:00 min/mile pace, you'd have actually run 19.7 for 9:08. On the other hand, if you'd believed my treadmill, when you thought you'd run 20 @ 9:00 you'd have only run 18.1 at 9:56 pace! A bigger concern would be the measurement points coinciding with a particular point in your foot strike, which could result in readings anywhere between the min and max speeds. Adding more marker stickers may help with that. My plan is to buy one of the sensors and try it out.
So, is it worth it? Well, the sensor is fairly cheap, listing at $99 at the time of writing, and it's almost certainly more accurate than your treadmill display, and more accurate than the accelerometer in your running watch. I'm expecting it to be more accurate than a standard Footpod, and while Stryd is remarkably accurate outdoors, Stryd tends to struggle on a Treadmill due to the speed variations. Runn is also a convenient approach, and if you took the time to calibrate your treadmill, you could add the calibration factor into most running watches. The Runn also measures incline, though I'm not sure what software would support that measurement. I'd love to have incline data from my treadmill runs, but it may require a custom Connect IQ data field.
Below is the data from a few steps on my treadmill. The orange markers show when my foot is in contact with the belt. I've assumed that any contact, no matter how slight counts as contact, so flight time is low.
