Connected GPS Pace and Distance is Inaccurate

[DanielMena9]

Went on a run today with my Charge 2 on one wrist and Charge HR on my other wrist. Also had RunKeeper to measure my distance. I started my run with about the same number of steps on both FIT devices. 

After the run, my Charge HR had almost 2K more steps than my Charge 2. In addition, my GPS map on FitBit matched my Runkeeper map which is about a 3.5 mile run BUT FitBit pegged the same route as 1.73 miles >.> 

Why doesn't this device work as advertised? I really like the design and feel of Charge 2 but if it can't do its job accurately, all is for not and I will be shorting FIT as it appears to be a widespread issue (I have maps and photos if someone from FIT wants the data).

Moderator Edit: Edited thread title.

[Bibster]

Or it could be the FitBit app.

The reason others are quoting more "variables" are simply because they know they are right! They are demonstrating that other apps and trackers work fine.
I do the same cycle to work each time, I know the distance. Strava and Goggle Maps record the distance accurately, FitBit using connected GPS does not.

Get Outlook for Android

[Rollerskier]

All tests presented are based on manual recording. Start and stop by pushing the buttom on the tracker or the Fitbit App. This is done at the same street sign. If I miss by 1 m at the streetsign thats abaout it.

I would asume that the quality of the GPS in the Apple iphone7 is not worse than in earlier versions of iphones? I got an iphone 6+ as a spare. Have the same iOS versions on both of them.

Regarding the Fitbit App. In order to "Track an excersise" you need to Set up an Fitbit device to your account and pair it with your spesific phone. If I remove the tracker from the App on a spesific phone Fitbit removes the menues for tracking from the App too. The App will not allow you to do much except logging your weight and food.

So you can not say lets just use the App for GPS map tracking and forget about the tracker.
The App is the same regardless of what Fitbit tracker you have set up but there must be some configurations that differs that affects the information flow/programs since I get different results.

I think that Fitbit should consider the GPS sampling rate versus the expected speed of the exercise.
If I have a speed of 30 km/h thats 8,3 m/s. If sampling is every 5 s I moved 41,7 m between 2 datapoints.

If I have a speed of 5 km/h thats 1,4 m/s. If sampling is every 5 s I moved 6,9 m between 2 datapoints.

If I get very good GPS signal Fix. The accuracy is 5-8 m at best.
Signal Fix will not be 100% stable so somethimes its 8-12 m or 12-15 m
So your real positions is in an area with a radius of 15 m but the GPS datapoint tells that you always are in the center.

So when walking I am not outside the radius of accuracy for the first datapoint when I get a new datapoint. The Fitbit draws a strait line between the 2 datapoints centre positions and culculates the incremental distance I walked in 5s. And adds up the accumuleted distance every 5 s and plots how I moved every 5 s. And we can all see how that looks like when walking.

So assuming the accuracy of the distance at driving 30 km/h is good enough calculated by Fitbit with a sample every 5 s.

When walking 5 km/h we need 5s x 6 = 30s betwwen every datapoint. Then I moved 41,7 m
And the track on the map and the distance should be the same as when driving 30 km/h

I dont know the sampling rate Fitbit is using but I would guess every 5 s is not far off and it looks like independent of the speed.

[Rollerskier]

Hei,

Today I will do another test.

I have configered up Charge 2 to both my iphone 7(active link) and iphone 6+(passiv just to have the option to start a RUN in the App but no communication).  Then I will disconnect the SIM so I not have internet and GPS connection on any phone.  And I will leave my Charge 2 home.  I will walk the same route with the iphone 7 in one pocket and iphone 6+ in the other pocket recording manually from the app on both phones.  So I can compare if there are any GPS signal quality differences on the 2 phones, on the same track as before, recorded at the same time.  Since GSM is disconnected I will rule out that GSM signals are used to determine location in case of GPS signals drop out somethimes during the walk.  It will be recorded in the RUN mode as before on the Fitbit App.

Test to find out:

1) Are there any major differences in distance calculated based on GPS signals from 2 different models of iphones 7 and 6+?

2) Compare the results with earlier test on the iphone 7 when I also left charge 2 home but had GSM cover on the phone7.

[Rollerskier]

Here are the results from today.

iPhone 7 ( was active link to Charge 2 but left Charge 2 home and took out SIM - start from App manually)

iphone 6+ ( Passiv link charge 2, took out SIM . Start from App manually)

True distance is about 3,2 km so iphone 6+ was on target and iphone 7 far off.

Then compare the previous test leaving Charge 2 home for iphone 7.

So we got 4,41 km last time and 4,31 km this time.

Conclusion:  The tracking of distance on iphone 6+ on the map was accurate while iphone 7 was far off again and but only 0,1 km different from last time it was tested on iphone 7.

So it looks like it can be something with the installation on the iphone 7 that´s corrupted or the iphone 7 itself. I will reinstall the App and then reset the tracker and reconnect the Charge 2 to the iphone 7 to see if this helps.

Or I can can make iphone 6+ active versus Charge 2 and see if this works better in all tests.

[Rollerskier]

Reset everything did not help on iphone 7.

So next step is to move everything over to iphone 6+ to see how thats goes with Charge 2 paired.

[Glennbest]

Hi Rollerskier,

Not sure why you want to do that test, but go ahead if you like. There aren't any issues with the gps signal to the phone because distance is calculated correctly when using the mobile run feature of the Fitbit app. The errors come about when the charge 2 calculates the distance from the gps signal shared to it from the phone via Bluetooth. If we could test the gps data direct to the phone versus the gps data to the charge 2 via Bluetooth it would be interesting to know if they are the same.

[Glennbest]

I'm using iPhone 6 and still get huge errors with distance using the charge 2.

[Rollerskier]

Thanks, I used iPhone 6 with Charge HR for 1,5 year and it worked well with pace and distance.  Then I got the iphone7 and shortly after the Charge 2 and I only have had unreliable information.  I don't know if there are some known GPS issues with iPhone 7 also? I have reinstalled iOS on iPhone7 in case something was corrupted and will check again for the last time. Unbelievable that Fitbit can launch a new model without checking that basic features like distance is correct on different platforms it supports....

[RyanJ5o]

Rollerskier - thank you for your diligence in doing all those tests. 

If one uses a phone where you know the GPS is accurate and compare it to the Charge 2's distance calculation and the numbers are consistently off (way off?) - even when - it is connected to an accurate phone with GPS... it is a problem with the way the Fitbit unit is calculating distance. Right? 

It seems obvious that the Fitbit Charge2 disregards/ignores a connected GPS phone when calculating distance. Isn't it is safe to assume the Charge 2's distance calculations are determined by the unit's calculation of steps?

Would anyone dispute that theory/assumption?

iPhone 7: There are reports of problems with the iPhone 7 GPS accuracy:

http://forums.macrumors.com/threads/gps-problems-iphone-7-or-ios10.1998933/

http://www.idownloadblog.com/2016/10/14/gps-not-working-iphone/ 

Android Devices: With most Android phones you can "Dial In" the accuracy of GPS in order to save battery life. See this post on RideWithGPS .com: https://ridewithgps.com/help/improving-gps-accuracy and AndroidPit .com: https://www.androidpit.com/how-to-improve-your-gps-signal-on-your-android

Seems FitBit doesn't trust data from a connected GPS device. Maybe now we know why. As others have posted, there are phones out there that don't provide accurate GPS readings and depending on your phones settings the GPS readings may be less than accurate.

I have been reading the posts on this thread for months. Not sure why I care. I returned the Charge 2 months ago because of the apparent insistence on using distance caluclated from the unit's STEPS rather than an accurate connected GPS device.

The Carge 2 has a lot going for it other than this "Connected GPS" anomoly. It is a shame Fitbit just can't let the wearer choose if they want distance calculated from the Fitbit's estimated Steps or from the Connected GPS unit.

Wouldn't that solve the issue?

[Rollerskier]

To sum up todays tests.  Same walk as before.  True distance 3,2 km.  All tests started manually on the Fitbit App.  Start and stop at the streetsign.

This afternoon.

3,2 km  Iphone 6+,  Fitbit Charge 2 disconnected blue tooth link to phone.  Start/stop on App.  Mode Run. No SIM

4,31 km Iphone 7.  Start/stop on App.  Charge 2 left home but Charge 2 was paired to the phone (blue tooth ) No SIM

This evening.

3,2 km Iphone 6+. Fitbit Charge 2 paired to this phone.  Active bluetooth link. Charge2 on my arm.  Start/stop manually from the App. No SIM.  Charge 2 reset.

3,85 km Iphone 7.  Fitbit Charge 2 disconnected blue tooth link.  Start/stop manually from the App. SIM.  iOS10.2 reinstalled before test.

Conclusions so far:

1) For iphone 6+. Looks like the Fitbit App reported distance accurate both when there 1) was an active bluetooth link between Charge2 and the iphone 6+ and 2) when the link was disconnected.  Same results 3,2 km both times.  True distance is 3,2 km.

Have not tested if I will get the same result when starting Run from Fitbit Charge 2 tracker yet?

2) For iphone 7.

Disconnecting the Charge 2 from the iphone7 reduced the distance reported from 4,31 km to 3,85 km.  However, still far off from 3,2 km True distance.

There could be an issue with the GPS accuracy for iphone7.  To be discussed with Apple.

Looks like the error in distance is larger when the bloe tooth link between tracker and phone is active for iphone7.

[Rollerskier]

Hi,

I did a test comparing GPS Accuracy for iphone7 and iphone6+

Conclusion: My iphone7 is not as steady as my iphone6+.  Signal quality is fluctating much more.

Best signals are about the same but iphone6+ can keep them much more stable while my iphone7 loose and regain accuracy in a fluctating matter.  Best way to see this is in a video but I attached a photo. Probably my iphone 7 performs somewhat better when the antenna is pointing up in the car window with the best GPS satelite signals than when it is in the pocket of my jacket, jeans or bag on my back. I will bring this up as a warranty case with Apple to see if they can give me a better iphone7 than this.  However, I still think there are a lot of issues with accuracy of Fitbit Charge2 as well and will continue testing start/stop on the tracker itself and the blue tooth link.

 

[Rollerskier]

I have sent a case to Apple technical support that will look at the inaccuracy of iphone7 compared to iphone6+ that I have compared with.  The position in App Maps ( Apple program ) is both too inaccurate and unstable for iphone7.  Apple is going to investigate if this is a firmware or hardware problem to be fixed. 

[rafsanchez]

I doubt this is an issue with the iPhone per se...both my iPhone 6s and iPhone 7 have been recording my distance perfectly accurately on the Strava iOS app...

[Rollerskier]

Well it could be just the GPS hardware on my iphone7...

Apple senior technical support in Norway will revert to me Friday on their findings based on the documentation/screen shots I have provided from Maps, GPS status and Fitbit tracks comparing iphone 7 vs iphone 6+.  They will Investigate what has been reported globally on iphone7 and issues with GPS so far.....I am probably not the first.  And likely upload some files from iTunes too. At least there is some action from Apple when issues are reported....

Now I will go and do another distance test with Charge 2 paired to iphone6+ starting the RUN from the Charge2 tracker to tests the blue tooth links.  And compare with iphone7 ,reinstalled as new today to be sure its cleaned up, with no bluetooth connection to Fitbit - started from the Fitbit App.

[Glennbest]

Hi Rollerskier,

I appreciate the tests etc you are doing to compare iphone 6 and 7 gps accuracy, however I think this is off the topic of this forum.  This forum is concerned with the problem of connected gps accuracy, whereby the charge 2 is not calculating the correct distance and pace during tracked excercise when using the gps signal of the connected phone.  The discrepancy appears to occur on several phone types.

For the record, I have an iphone 6 and I'm confident distances and pace are correct when excercise is started from the iphone app.

I did discover something interesting in the excercise data though; 

I downloaded the fitbit data (tcx file) of two almost identical runs.  I then calculated distance travelled between sets of latitude and longitude by using simple trigonometry and compared this to the distance between the points calculated by the app and the charge 2.  For the fitbit app, over the course of an 18km run I found that there was an average discrepancy (about 0.01m) between the distances calculated by myself using the latitude and longitude co-ordinates, and distances calculated by the fitbit app.  However, for the charge 2 the average discrepancy between distances calculated using latitude and longitude co-ordinates and distances calculated by the charge 2 was 30 times larger (about 0.3m).  I'm comparing the average distance discrepancy between every set of lat and long co-ordinates, every second for almost 6000 seconds.  For an average discrepancy of 0.3m that works out to be 1.8 km error!  I did the same comparison for another run and got similar results.  I've made no allowance for the curvature of the earth etc but the comparison between the iphone app and the charge 2 shows the two are not giving the same results, but clearly they should.  I've provided an example of the spreadsheet below to illustrate how I made the comparison, ignore the figures in the diffrence column, I averaged them all out at the bottom of the talbe to get the results above.

ns1:LatitudeDegrees	ns1:LongitudeDegrees	Distance between Lat & Long Co-Ords	ns1: Total DistanceMeters2	Distance travelled (calculated by iphone App)	Difference	ns1:Name
-37.59396756	143.7233062		0	0	0.000000	Fitbit iPhone App
-37.59396756	143.7233062		0	0	0.000000	Fitbit iPhone App
-37.59394789	143.7233104	2.011	2.217833279	2.217833	0.207115	Fitbit iPhone App
-37.5939244	143.723314	2.377	4.849384723	2.631551	0.254230	Fitbit iPhone App

ns1:LatitudeDegrees	ns1:LongitudeDegrees	Distance between Lat & Long Co-Ords	ns1: Total DistanceMeters2	Distance travelled (calculated by Charge 2)	Difference
-37.59375155	143.7233199	0	0	0.000	0.000
-37.59375155	143.7233199	0	5.69	0.000	0.000
-37.59373164	143.7233199	1.991	7.12	1.430	-0.561
-37.59367657	143.7233282	5.570	8.67	1.550	-4.020
-37.59364665	143.7233316	3.011	10.37	1.700	-1.311
-37.5936166	143.7233299	3.009	14.76	4.390	1.381
-37.59358656	143.7233232	3.077	17.37	2.610	-0.467

Fitbit take note, the charge 2 is way outside acceptable accuracy.  Stop selling these things and issue a recall.

[hawkeye1952]

yesterday mine was .13 off. from what I can tell on the 2 maps. they look the same from start to finish.

[Rollerskier]

Today I completed the last test of iphone6+ and Charge2.

Charge 2 was paired up with iphone6+ and the RUN was started manually from Charge 2.

Same route as before and the phone was in the same pocket as before. Total distance 3,16 km and no wobbling as before.  True distance is about 3,2 km.

Yesterday I did the test on iphone7 after reinstall iOS10.2 as new cleaning up all setting as a request from Apple Support. The RUN was started from the App and I had cut Bluetooth to Charge2 to avoid interference.

Unfortunately it did not help. See the results below. 3,81 km and wobbling....

Attached is the total route walked comparing the 2 tests. True distance is about 3,2 km

 

To sum up the testing on iphone6+ and iphone7

Same routes all tests. True distance about 3,2 km.

IPhone 6+

3,2 km Fitbit Charge 2 disconnected the blue tooth link to iPhone. Start/stop on App. Mode Run. No SIM - No wobbling 15. Jan 2017

3,16 km Fitbit Charge2 paired to Iphone6+. Bluetooth link active. Start/stop manually on Charge2. SIM inside. - No wobbling 17. Jan 2017

IPhone 7

3,81 km Started on App. No active link to Charge2. (After reinstall IOS10.2 as new) 16. Jan 2017 Mode Run

4,31 km Active link to Charge 2 but left Charge 2 home and took out SIM - start from App manually (Prior to reinstall as iOS10.2) Mode Run 15. Jan 2017

Charge 2: 3,87 km (start/stop on App - tracker on arm) 13. Jan 2017 Mode Run

Charge 2: 3,52 km (start/stop on tracker) 11. Jan 2017 Mode Run

Conclusions:

For some this might come as a surprise but:

1) I found the Fitbit Charge 2 distance accurate enough when using iPhone 6+, both when starting the Run manually from the App and from the Charge2 tracker itself. No major wobbling effect found. Deviations of 20-40m from a true distance of about 3,2 km is not a lot. It is easy to cut a few corners as well during walking and you can loose 10-20 meters.

2) For iphone7 (at least my phone) the distances are always too long. Routes are wobbling and  have a lot of variations on the same track as well. I have found more fluctuations in GPS accuracy on iphone7 compared to iphone6+ when testing on App Maps and App GPS Status. If the satellite signals are not ideal as when leaving the phone in the window of your home, this is easy to notice. The components in the Apple phones vary between models so even with the same iOS operating systems the BIOS will differ between phone models so there might be so diffrences on values.

3) I tested outside yesterday with both phones in the palm of my hand walking a few hundred meters. IPhone 6+ had excellent fix of consistent horizontal Accuracy of 5 Meters and vertical Accuracy of 3 Meters.

Iphone7 had Excellent fix in the range of 6-8M and vertical fix of 8-23 Meters on the same walk. This is ok and explains probably why I get fairly good distance results when driving the car with the phone is in the front window or going for a walk with the phone in the top pocket of my bag. Not too many wobbling effects on the track then. Probably the GPS signals to the phone in my pants or jacket are not that great and stable, This affected my iphone7 more than iphone6+.

4) My intention has not been to focus on limitations of iPhones and take away the focus from Fitbit Charge2. Rather illustrate that good and stable GPS signals are required to calculate distances correct as the Fitbit Charge2 currently works. Fitbit has some homework to do calculating distances and pace more correct when the GPS raw data quality is not superb.

[rafsanchez]

I think the limitation of your tests is that in both cases (iPhone 6 and iPhone 7) you are only using the FitBit app to record the distance. What your tests do show however is that within the FitBit app there is a difference in how FitBit communicates with an iPhone so is that the source of the problem?

I have regularly run and biked exactly the same route for the past 2 years and with both phones (I used the iPhone 6s previously but now use the iPhone 7) the Strava app for iOS reports the distance accurately on a very consistent basis.

The FitBit app on iPhone 7 and using a Charge 2 is very inconsistent - although the GPS tracks seem to be relatively accurate I get wildly inconsistent distances. I mean crazy inconsistencies...look at this:

[Rollerskier]

Yes I agree, iPhone7 and the combination Fitbit Charge2 is a disaster....your track looks the same on the map until you zoom in very close.  Large variation in distances comparing the same exercise over time.  

By the way I had some problems too with some navigation maps to like NAVIGON Europe on iPhone 7 with the first release of iOS 10....have not checked out 10.2 yet.  The car position was not moving in a continuous motion on the map while driving as it used to do... more moving stepwise and dragging behind my position on the road.

By the way my test was on the iPhone 6 Plus ( large screen ).

[bobby2478]

I just went for a walk and performed the following test :

Device 1: Galaxy Note 5 paired with Charge Hr.  Started and stopped recording walk in the Fitbit mobile app.  Distance was 3.28 miles with a fairly smooth map line. 

Device 2: Galaxy S4 not paired with any tracker.  Started and stopped recording walk in the Fitbit mobile app.  Distance was 3.27 miles with a map line nearly identical to device 1.

I took both devices with at the same time and recorded both activities simultaneously.  Since they were both almost identical, I'll next perform the same test except swap out the Charge Hr for the Charge 2, recording from the app.  Then I'll perform another test where I'll record from the Charge 2. Between those tests I should be able to tell how close the charge 2 is to the distance from the Fitbit app.