# Calibrate RUNN

I’m hoping someone can help me understand calibrating the NPE Runn device.

I thought the point of this device was to give an accurate speed reading, instead of relying on the speed displayed by the treadmill.

It comes factory calibrated at 5 mph, yet the calibration methods says to set the treadmill at 5 mph, so does the device then calibrate to your treadmill, so that 5mph on the treadmill display is what shows in Zwift, even though you might actually be doing a different speed?

For instance, I installed my Runn device and when I set the treadmill to 10 km/h, Zwift showed 10.3 km/h, and to get 10 km/h in Zwift I had to set the treadmill to 9.7.

So if the RUNN is properly calibrated at the factory, then I know I should set the treadmill to 9.7 km/h to be accurate.

But I see that to calibrate it, you do it with the treadmill set at 5 mp/h. Does this calibrate it to the belt movement, so that in my case, the RUNN would measure 10 km/h when the treadmill shows 10 km/h?

Surely this just makes the RUNN device inaccurate? (I’m not worried about incline)

Correct, you are calibrating the Runn module to your treadmill’s speed. The assumption is that your treadmill is reasonably accurate with the speed it displays. If you’re concerned about the treadmill’s speed you will need to work out a way to confirm the actual speed of the belt vs the speed being displayed.

Thank-you Aaron.

So is the RUNN device accurate? That is how I am determining the actual speed of the belt verses speed displayed on the treadmill and coming to the conclusion that the treadmill is displaying .3 km/h slower than it actually is.

But at 4 km/h there wasn’t any difference.

Let me point out that I am being pedantic, I guess 0.3k difference at 10k is only 3%, is that considered reasonable? It is a 15yr old treadmill so cant really complain.

But is that considered fair if I entered a race? (thats probably irrelevant as I’m likely to be at the back anyway

This is from DC Rainmaker: https://www.dcrainmaker.com/2020/01/npe-runn-treadmill-smart-sensor-everything-you-need-to-know.html

More or less you don’t need to calibrate the NPE RUNN, it measures the treadmill speed (actually the time the sticker or mark passes between sensors) and transmits that to Zwift or other receiving device/app.

Belt speed can and probably will change with someone running on it and that can be the difference in what the treadmill is displaying and what the RUNN is broadcasting.

2 Likes

Your explanation might explain why there was a difference at 10k but not 4k (as I wasn’t pounding it at 4k).

1 Like

Hi,
I received my RUNN few days ago and i applied this procedure to calibrate it :

I confirm the issue at 4Kmh, nothing are transmetting to the receiver (Garmin Watch)
Anothers issue, i am french and we used KMH not MPH on treadmill.
There are some differences with speed between both because 1MPH = 1,609344KMH

mph kmh
1 1,609344
2 3,218688
3 4,828032
4 6,437376
5 8,04672
6 9,656064
7 11,265408
8 12,874752
9 14,484096
10 16,09344
11 17,702784
12 19,312128
13 20,921472
14 22,530816
15 24,14016

So is it possible to calibrate it in KMH ?

Thanks

1 Like

Hi, I don’t have my Runn device any more so I can’t confirm, but you could just set the treadmill to the equivalent for the calibration?

For example, when it asks for 5mph, just set treadmill to as close as possible in kph etc.

You can set Zwift to use km instead of miles (and km gets you more points).

I would just leave the runn at the factory default for speed as that is calibrated to 5mph so is a known good speed. Set Zwift to kph and it’ll convert the runn value to kph. Might not match exactly what the treadmill display shows but that means the treadmill might not be going the speed it thinks it is.

Frank. If you download the Configurez app from NPE you can configure the Runn in KM. It configures at 3 different speeds so will be accurate.

Kevin,

What did you move to instead of the Runn?

I killed my treadmill so wanted something really sturdy. Ended up getting a refurbished commercial treadmill, Precor 835 which is about 3yrs old. It’s big but worth it.

This has a “csafe” port so I got the Gem module (from NPE who make the runn). This effectively turns it into a smart treadmill that sends speed and incline straight to Zwift on my iPhone via Bluetooth.

Before I sold the runn I used it on this machine and the speed was almost exactly the same, perhaps 0.1 km difference, so I’m happy with that.

My experience of the Runn is the better the treadmill the more accurate it is. My speed bounces all over that place at home on my treadmill.
On the work commercial ones it’s solid.

That’s possibly due to the power of the motor in the treadmill. A lower power motor will slow with each foot strike and speed up when you’re in the air, so the runn is picking up those variations. A higher power motor isn’t affected as much.

My old treadmill was constantly going up and down a few .10’s of a km on every run, but the new one it didn’t fluctuate (or no where near as much).

I guess bodyweight also plays a part, and let just say, I’m not a light runner!

If I ever got another treadmill that needed an external speed sensor I would probably get the runn again as I didn’t have any real problems with mine, although some people have loads of problems.

I concur with your findings. My home setup is 3hp but still struggles. Think the work one is 5.

@Stuart
Thank you
I tried with the Android app “configure” and I can calibrate it in KMH on 3 different speeds (walk, run and speed).
After extensive testing, the gear ratio is not accurate.
North Pole Engineering support informed me: +/- 0.5 km / h from the treadmill.
I will therefore only use it for Zwift and I will keep my Garmin watch for my real speed of my training

My little feedback
When you are running with Zwift, you can see your cadence at the bottom of the screen.
And you can see some yellow ranges in the orange graphic.
I think the captor doesn’t detect correctly the stickers everytime

PI : My treadmill is a ProForm ZLT700

Thanks at all for your help

Your Garmin watch might be even worse depending on the model.

It hopefully has the option to calibrate and save at the end of your run. If so it will get more accurate the more you run.
If you have a newer watch you can link via Virtual Run which removes the need for the RUNN. You can also calibrate and save this to match the treadmill.

I’ve experiments with numerous different watches and all offer varying accuracy to the treadmill speed/distance. Most are within 100m of 1km so about 10% innacuracy.

The best way I have found to calibrate the entire setup on my old school treadmill, irrespective of whether the treadmill is accurate is as follows.

• Put a marker on it.
• Set it at a range of speeds (slow, medium,fast) and count how long it takes for the marker to past 100 times.
• Do the calculation to work out exactly what speed the treadmill is actually doing at any of the speed settings. Now you know how acurate or not the treadmill is. e.g. 2.5m belt at 10kmh setting, passes 100 times in however many seconds.
• Calibrate the NPE Runn to the actual measured treadmill speeds. e.g. when I press 10kmh on my treadmill, it is actually doing 9.4kmh. So when I press 10kmh on the treadmill I would calibrate the NPE to measure that as 9.4. ConfigureZ asks for 3 speeds (slow, medium, fast) so do it for the three speeds you calculated in step 4.
• All devices are now calibrated to each other.

Hope that helps somebody

Scott
Well done with this. This is what I have done to prove that treadmill belt is not running at the advertised treadmill set speed.

I’m going to apologise for the length of my reply but I hope, like you have done, that if I can help others to set up a Runn it is worth the time writing this.

I have also contacted Runn to discuss why they have a ‘calibrate your Runn’ option if their Runn very accurately measures the belt speed. They are very confident that their Runn is accurately calibrated and does not need calibrating. They say the option to calibrate your Runn to your treadmill (or any speed you desire) is for those people who would like their Runn and their treadmill to report the same speed, irrespective of whether that is actually the real belt speed.

I have measured my belt length and carried out very similar calculations to your own to check belt speed at three speeds, slow, medium and slightly faster (but in no way considered fast!). When I, eventually, calculated the belt speed correctly i actually found that it very closely matched what the Runn was reporting.

So as you have written some nice easy to follow instructions for new users or anyone struggling to set up their Runn may I suggest the addition of a couple of extra steps.

5a. Set your belt speed to the Slow speed you have calculated it to be and check what the Runn ConfigureZ app says the treadmill belt is running at. Up the treadmill speed to what you know the medium speed to be and check Runn ConfigureZ app speed reading then do again same for fast speed. (This is similar to what you are describing in your step 6 but at step 6 you are already starting to talk about calibrating your Runn, I believe my steps 5a and 5b might save you having to calibrate at all)

5b. If your calculated belt speeds are not closely (0.1-0.2 kmh or 0.1 mph) matched by the Runn may I suggest a recheck of the belt speed calculations again before moving to step 6. (Now move to step 6 if you really believe you need to calibrate your Runn)

I believe these two additional steps may persuade users that there is no need to move to your step 6 to recalibrate their Runn as it is accurate.

I add these two points from experience as I initially got my set up procedure and belt speed calculations wrong.

I may also add that counting more than 100 revolutions, say 200 or 300, leads to a slightly more accurate belt speed although obviously takes longer.

For those wanting a slightly different belt speed calculation option I decided to find out how many revolutions of the belt was 1km and time how long it took to do exactly 1km at various speeds.

So in my case

1. belt length was 322.5cm or 3.225m
2. 1km / belt length —- 1000 / 3.225 = 310.08 revolutions = 1km
3. 310 revolutions took (say for easy maths sake and my medium run speed) 360 seconds —- 3600 seconds / 360 = 10.0kmh

I am very happy with my Runn, I believe it accurately reports the true belt speed at all belt speeds (without the need to calibrate at all). It tracks my cadence well although doesn’t register cadence at walking speeds but think I have set it up too far towards the back of the treadmill and might at some point try moving it closer to where my feet hit the belt.

There are questions on other forum threads as to whether testing belt speeds and counting revolutions should take place with you running on the treadmill or not. In my experience the treadmill belt does move at very slightly different speeds depending whether you are on it or not.

Does this mean you should calculate belt speed whilst running on the treadmill. Absolutely no need to. Why - because the Runn doesn’t know whether you are on it or not. All it is required to do is measure the belt speed. Does the belt speed slow down when you run on it -Yes(*1 see below). Does the Runn identify the drop in speed - Yes(*1)

If you want proof here is a test you can take. For H&S sake I recommend only doing this test if you are experienced with stepping off a treadmill on to the side ‘running boards’ whilst the treadmill is moving.

Set up and enter Zwift Run.
Raise the treadmill to a comfortable running speed in my case 6.0mph.
Run for 30 seconds - does Zwift/your avatar continue to report 6.0mph, hopefully yes.
Without touching the treadmill speed setting button, hold onto the treadmill handles and step off the treadmill onto the side ‘running boards’.
Does your Zwift Avatar start to run at a faster speed.
Yes in my case it now runs at 6.1mph.
Proof that the Runn has identified that the belt is moving faster now you are not on it. IMHO proof that the Runn is accurately engineered to calculate belt speed irrespective of whether you are on it or not.

Note 1* - I suspect, but have no proof, that the better quality top end treadmills with more powerful motors will have a smaller deviation in speed when you are on it or not. I also suspect, with no proof, that a much lighter runner will have less impact on belt speed deviation than a very heavy runner.

Once again I hope this helps someone else get their head around the NPE Runn setup.

1 Like

Good grief Ian!

I was being a bit ‘lazy’ in writing mine up, just hoped it would give some people a starting point, but great that you have added to it.

Your steps 5a and 5b may well save a few people, I assumed that treadmills are never accurate, so opted to jump straght to calibration

When I did mine I actually took belt speed readings every 2kmh then interpolated. Once I did step 6 I went through my speeds 1kmh at a time and the Runn follwed them perfectly.

I agree, if your eyes can take it, especially at faster speeds, as high a count as possible for the revolutions will result in a more accurate reading. Another point is to make sure you have allowed the machine to warm up, run on it for a wee bit too to loosen the belt, your point about variations in motor speed are valid I think.