Wahoo Speed Sensor Calibration Issues
 

I have a calibration issue with the Wahoo Bolt when paired with the Wahoo Speed Sensor that I can't seem to solve.

When the Bolt is computing speed and distance from the speed sensor, the results are consisently 5% optimistic compared to Strava on my iPhone. However, if I unpair the Wahoo speed sensor and allow the Bolt to pull data from its own GPS sensor, the results closely match Strava.

So Wahoo techinical support suggested that I take precise measurements of my wheel circumference (the speed sensor is attached to the hub and only knows revoluations) and calibrate based on those results. It turns out that my wheel is 2133.6mm round and the calibration I had already set the Bolt to was 2136mm. This is a difference of <0.2%, well below the 5% difference I'm measuring on the road.

I want correct speed and distance readouts, but my two methods for checking calibration errors differ. Does anyone know what might be going on here?

Thank you,
Robert

You need to travel over a known distance and compare the results from both the sensor and the phone's GPS-- there's absolutely no guarantee that the GPS is more accurate, as GPS is best at placing a stationary object, but can be off by 30+ meters while in motion. This is why people running the phone app can KOM Strava segments, as their segment that day was a hundred meters shorter than everyone else's.

I know that the SART has mile markers on the ground, the LART probably does as well. My Wahoo speed sensor reads 1.03 miles for every mile marker, yet if I do a ride with two computers running, one paired to the sensor and one on GPS, the GPS can be as much as 5% shorter on distance. Also, the auto-calibration with Wahoo may be as bad as Garmin's. My Edge 520 was set to auto-calibration, and shorted me around 4% of my total distance for over 2 years.

Strava segments are done by time over distance, so no matter what the speed sensor says, we're at the mercy of the GPS. I think my speed sensor over-reports about 2%-- so I get a "bonus" 200 miles out of every 10,000. I'm not really worried about it.

What are SART and LART?

Santa Ana River Trail and Los Angeles River Trail.

Is Ridewithgps considered an accurate way to measure distance along a route?

It will be as accurate as the GPS receiver in the device you're using to record-- they are not all created equally. The oldest ways are the best-- measure time over a known distance, use that to calibrate your speed sensor.

No, I meant if you create a route using RWGPS, does the mapping software contain precise enough measurements to establish a known distance? If not, then what would you use?

Using a pre-programmed route is going to introduce as large a margin of error as an improperly set wheel circumference, if not more. After all, if you rode your bike on Lane 1 of a track four times, you'd cover 1,600 meters. If you did 4 laps in lane 8, you'd cover 1,815 meters. Every time you go around a corner, narrower or wider, you'd be changing the distance. Traveling a mile in as straight a line as possible (like on a river trail) should be plenty.

Understood. So how do you measure the actual distance traveled for a given route? Or, how do you find a route with a known distance?

Can you find a low traffic highway with mile markers?

Some apps do a better job than others with gps data even on the same device because of software design issues, polling frequency, etc. Some errors can be systematic such that you almost always measure short or long.

Did you do a roll-out to measure the wheel circumference?

Just for kicks I tried wrapping a (fabric) tape around the tire, vs rolling out & it came out at least 1cm different.

I would get that dialed in, & then not worry about it. GPS does stuff like measure to the riverbed when you go over a bridge,

& Strava on a phone can include your trip into the bathroom- it's all an approximation.

Yes, that's how I measured the circumference of exactly 84" or 2133.6mm. The Bolt is calibrated to 2136mm, so that explains 0.2% of the 5% difference. The overread on the Bolt is rather consistent compared to GPS and by a fairly wide margin -- about 1 mph. Makes me question if the speed sensor is correctly calibrated.

I can do a ride with my wife, where we ride side by side on bike with literally identical tires-- 700x25 Giant Gavias. We both have speed sensors calibrated to the same rollout, because... it's the same tire.

Without fail, the auto-lap chime goes off for her at 5 miles, and I will be at 4.94. When hers chimes for 10, I will be at 9.88. And so on and so on, her always recording an additional 0.06 miles every 5 miles. Of course, this is 0.012 miles, or a "bonus" 63.36 feet per mile traveled.

This falls soundly into margin of error, even though it is repeatable. Best you can do is to get your computer to record as close to a mile as possible, and go from there. Or just throw caution to the wind and let it auto-calibrate. Might work in your favor, might not.

I have one health app that records exercise activity in the background, using GPS only, and auto-starting/stopping itself presumably based on velocity. It is generally about 5-7% short on distance.

Do you both also weigh the same?

You have to measure roll-out with the rider on the bike since the tires deform.

You want the circumference of the circle with the vertical radius from the hub to the road with the bike loaded. This radius is shorter than with the tire without any load.

GPS distance will tend to be lower than the distance measured by a correctly-calibrated wheel sensor. The measurement "error" (variability) is larger for GPS.

I did measure rollout unloaded. Is measuring the wheel loaded likely to make anything close to a 5% difference in circumference?

So I spoke to Wahoo and they're replacing the speed sensor. The tech support agent hadn't heard of this kind of thing happening, but was willing to swap out the hardware just to be sure. He seemed to take particular note of the fact that average distance reported by the Bolt with GPS was essentially identical to the iPhone with Strava.