1. Place a piece of tape and/or use a brightly-colored marker or dab of paint to mark a point on the belt.

2. Measure the length of the belt starting and ending at that point.

3. Turn on the treadmill and run on it while it is set at 5:46 pace.

4. Have someone else time multiple revolutions of the belt (the more, the better, but no need to go nuts) by watching that same spot go past a fixed point, e.g., the end of the back roller (make sure to start count at zero, not at one, when watch is started).

5. Calculate the actual speed by multiplying the length of the belt by the number of revolutions and dividing by the total time.

Note that step #3 can be critical, as depending on the quality of the treadmill, the size of the motor, the state of wear of the underside of the belt and the platform underneath it, etc., some treadmills run significantly slower under load than when no one is on them.

EDIT: Note also that the discrepancy between the pace at which the treadmill was set and the pace indicated by your Garmin could also be due to slight differences in your stride mechanics when running on a treadmill vs. overground (assuming you calibrated it when running overground, that is).

If your bike has a speed sensor thats been calibrated via a rollout meaurement, you hold it on the treadmill and compare readouts. I did that for my precor and found it was out by about 5%. No way of changing the calibration on the Precor, because it uses a wheel with black and white stripes and an optical sensor, so it's a digital encoder.

Have you got a calibrated computer on your bike ?

Hold it on your treadmill at whatever speed you set - make sure you hold it steady, then read speed of your computer - may need to convert between mph or kph and min/mile or min/km

Andy's method is the best. That is how I do it.

---

------
Scott McMillan, M.Sc
Twitter@Factor9Coaching | Factor9Coaching.com | Facebook

C'mon guys. Why would one method be "the best"? Seems to me, you do the one thats easiest, unless you plan on writing a PHD thesis on the variations of treadmill belt speed to room temperature, loading, etc. And typically, the treadmills tend to have a speed feedback that uses a digital encoder like the precor, so the variation is minimized. I'm sure either method is well within the amount of error bar that you need to know it for, just one involves precise marking and timing, the other just puts a bike on the darn thing to read a number.

Problem with the bike, is you have to calibrate the bike, mights as well just do the treadmill unless you are calibrating the bike anyways.

---

------
Scott McMillan, M.Sc
Twitter@Factor9Coaching | Factor9Coaching.com | Facebook

Lol. But then you know that the next forum topic will be "how to calibrate a bike computer", and someone will say " put it on your precisely calibrated treadmill", instead of just doing a rollout measurement!

I did a combination of the bicycle-on-treadmill and Andy's points:
1. Measure your bike's unloaded rollout and calibrate your unit accordingly
2. Have someone hold your bike while the measuring wheel rolls on the treadmill.
3. If possible, run next to or behind the wheel at your "critical" pace for >5 minutes.
4. Calculate correction factor from set speed/pace and the measured average speed/pace.

My treadmill is spot on unloaded but slows down 2.9% when I run around 3:45/km. How it handles 4:40/km I don't know but then again it doesn't really matter at easy pace, does it?

I'd be surprised if any home/gym treadmill has a functioning speed feedback, error integrating controller. I.e. eliminates covered belt distance error with load.

I have a precor so I guess there is no calibration to be done. What I really wanted to know was accuracy so I'll try the colored tape thing and maybe cross-check with the bike computer too.

I'm pretty sure I run differently on the treadmill than outside so maybe the footpod does get confused. What I really need to know if what speed to set my treadmill to get a real world speed so the tape should give me that.

Thanks

---

---

Warning: Username does not in any way apply to triathlon past and future performance.

Precors have a speed feedback mechanism wherein the speed from a digital encoder is compared to what the treadmill thinks the speed is set to on the display, and then the circuitry adjusts the motor speed to match. All this is a long way of saying, the treadmill is supposed to adjust itself to maintain a constant speed under varying load conditions. However, the actual constant speed could be wrong, and there's no way of adjusting this. So, using a bike should work because the speed is independent of the load if the circuitry is working correctly. Thats the way a Precor is, I can't speak for other machines.

Some high performance treadmills are very good at controlling accuracy. We used to use a Woodway in the lab. Except we had custom slats made to make it more road-like.

---

------
Scott McMillan, M.Sc
Twitter@Factor9Coaching | Factor9Coaching.com | Facebook

Digital encoder or simple reed switch doesn't really matter. The controller has to keep a running tally of actual pulses and expected pulses and set speed accordingly to maintain constant and accurate rate of distance. Doable but not childsplay.

The big question is wether the pulses come from the driven drum or the passive drum. To transmit power between a smooth drum and a smooth belt there has to be slip. This slip is influenced by foot impact/liftoff mechanics, belt tension, surface smoothness, temperature, humidity, etc.
So the driven drum is clearly not a good source for speed feedback. The passive drum should provide much better signal for the controller to measure belt speed/distance.
I suspect most treadmills do use the driven drum (i.e. motor shaft via toothed transmission) as feedback source.

Counting belt revolutions yourself trumps any accuracy claims by the vendor.

Never mind...Missed the "foot pod" the first time I read.

Thanks