KS16S FW 2.01 speed/distance deviation

[Chriull]

Since there are already some topics around wondering/guessing about speed/distance deviations, i decided to look at the numbers for my KS16S FW 2.01 and realized that Libre Office (most presumably excel, too) has a perfect tool to get the numbers easily from an EUC Log!

So how to make this:

- Log a ride

- Make a XY Scatter plot of speed vs gps_speed (distance vs. gps_distance)

- Add a trendline for the dataset. (Linear/force y-axis cut at 0/show equation)

and voila: +11% for the speed shown (f(x)=1,1104... x) and -2% for distance!

Edited February 27, 2020 by Chriull

[Daley1]

So the speed is overestimated by 11% by the ks app and the distance is overestimated by 2 %????

[Chriull]

1 hour ago, Daley1 said:

So the speed is overestimated by 11% by the ks app and the distance is overestimated by 2 %????

Exactly. For the KS16S with FW 2.01. For my tire diameter(air pressure, ...) 

So i have in reality something about a 35/1.11=31.5 km/h wheel...

PS.: And yes - they do calculate the speed and the distance different! 

Edited February 27, 2020 by Chriull

[Chriull]

After the discussions in the KS18X topic about driving straight tracks, carving, random riding with many corners, etc and the influence on the gps speed i drove a ~1km straight track:

My first impression is that the gps speed is at least to be looked at more closely (if not just questionable?). Btw - i like the gps accuracy loggend now with EUC World

Edit: From my gut feeling i'd say the trend of the speed graph is more as i drove - not the trend as shown with the gps speed graph. ...but feelings can easily be different from reality...

Seems for more detailed detailed considerations i'll go for this track for longer time (~10 km straight line). Time measurement from EUC World should be more accurate then any other method i could consider, distance is available from google maps/goolge earth and then compare these values with the speed and gps speed averages.

Edit: Maybe i'd also choose some other track/part of the track - here were too many big trees on over the road so gps accuracy was not overwhelming...

Ps.: For the record - KS16S, FW 2.01

Edited May 6, 2020 by Chriull

[travsformation]

On 2/27/2020 at 11:01 AM, Chriull said:

On 2/27/2020 at 10:01 AM, Daley1 said:

So the speed is overestimated by 11% by the ks app and the distance is overestimated by 2 %????

Exactly. For the KS16S with FW 2.01. For my tire diameter(air pressure, ...

2% distance difference....which is 20m/km...the exact same number I'm getting on my 16X on FW 2.02 (and previously, 1.0.7)

Now for the 1km carving test, I expect bigger differences in GPS distance AND speed.

[Chriull]

Made some graphs out of the log from my last post:

Here is beside the normal from the wheel reported speed als Dd/Dt - delta distance (from wheel reported) / delta time, so speed calculated from the distance values. (with 5 past and 5 future values used for averaging)

Same with GPS Speed and DGPSd/Dt - delta GPS Distance / delta time. Here 15 past and future values were used for averaging. This much points needed for averaging the line unfortionately "cover" the dips shown by gps speed  Maybe i'll look at this once with the logged gps timestamp.

Overall Distance vs GPS Distance deviation of this part of the track was 917m vs. 885m - so ~3,5%.

Here the three different speeds in comparison:

Strangely GPS Distance and wheel distance deviate overall by ~3.5%, average deviation from GPS Speed to speed calculated from wheel distance still deviates by 12%!

My wild speculation for now would be that the "GPS speed drops" shown are "faulty values" ?due to bad gps reception?"

Another wild speculation - as Distance deviates from GPS Distance by ~3.5% and wheel reported speed is about 3% higher as "speed calulated from wheel distance" the "real deviation" of speed from reality could be about 6%. (As GPS speed deviates about the same amount in the other direction?)

[Seba]

3 hours ago, Chriull said:

Strangely GPS Distance and wheel distance deviate overall by ~3.5%, average deviation from GPS Speed to speed calculated from wheel distance still deviates by 12%!

There is nothing strange  2.x firmware added correction to distance, but still inflates speed:

 

[Chriull]

35 minutes ago, Seba said:

There is nothing strange  2.x firmware added correction to distance, but still inflates speed:

In this case i calculated speed from the reported distance values - so the deviation is the same as the distance values (3,5%). But there is still a deviation of 12% to GPS Speed! Seems GPS Speed gives no average since the last measurement, but some momentary value "additionally faulted by bad reception"?

Edit:

If one looks at GPS Speed versus the Speed calculated from GPS Distance for this track:

The area below the blue GPS Speed line (==distance) is "much lower" as the area below the green line (Speed derived from GPS Distance). So (at least) in this case reported GPS Speed values are tendentially lower than "reality". In any case they are in the best case just samples from one point in time, eventually deviated by GPS reception.

Edited May 7, 2020 by Chriull

[Daley1]

the Ks app is really a guide and not a good indicator of speed ,distance or battery left.

I went for my normal 47 k ride with a wheel that was fully charged 36 hrs prior. I forgot to top-up the wheel just prior to the ride .Before I took it out I checked the app and indicator lights and they registered 100%.I rode to the halfway point and checked the app which said 74% left and 57ks.On the way back I got about 12 k got "please decelerate" and then rode it to it became inoperable. Had to get my wife to pick me up.The person who invents an app that reads truthfully and presents realistic data will get a big plus from me!

[.....]

I am now taking donation to buy a speed gun to put an end to me wondering. I PROMISE it'll help ME with my own testing and I'll post results that mean nearly nothing to you.. I'm seeing bushnell sells a radar gun with +-1 mph accuracy for about $80. I havent looked very hard, but perhaps a few people are so interested, it would seem that SOMEONE with a lot of wheels or at the factory, would spring for one of these damn things. Be fun to have at the NYC drag races I bet. +-1mph is not very accurate however. Theoretically, you could ride with the gun, point it at a stationary object and easily get YOUR speed?

@Daley1Ive found the same thing happen to me. Charge it 'fully', come back 2 days later and it be quite a bit lower. The tolerances for % of charge reporting is so damn large, I KNOW it isnt full, even when it says it is.

Edited May 10, 2020 by ShanesPlanet

[johnyz89]

3 hours ago, Daley1 said:

the Ks app is really a guide and not a good indicator of speed ,distance or battery left.

I went for my normal 47 k ride with a wheel that was fully charged 36 hrs prior. I forgot to top-up the wheel just prior to the ride .Before I took it out I checked the app and indicator lights and they registered 100%.I rode to the halfway point and checked the app which said 74% left and 57ks.On the way back I got about 12 k got "please decelerate" and then rode it to it became inoperable. Had to get my wife to pick me up.The person who invents an app that reads truthfully and presents realistic data will get a big plus from me!

If You plan Your ride to leave less than 10% battery in the end the You need bigger battery wheel. You can't predict exact battery consumption and always need some everhead. Also don't trust in app distance prediction algorithms, they show distance based on ideal conditions which never occur.