Ionic GPS Inaccuracies

[MP4]

Hi, 

I'm a new Fitbit user, bought an Ionic on Monday and I'm having some issues with the GPS. Went on a couple of runs (at a track) and the gps seemed to be completely out both on the distance travelled and pace per kilometre.

I'm a long time strava user so yesterday I went for a run with both the Fitbit Ionic going and the strava on my iPhone running at the same time. Here are the results:

Strava - 7.9km   @ 4:15/km

Fitbit - 7.2km     @ 4:34/km

The difference of 0.7km is quite big and the difference in pace is also worrying. I waited till both had connected to gps before starting the run and I run in London where gps signal should be good. 

Is my Fitbit Ionic GPS not working correctly? Any help/advice would be appreciated. 

Thanks

Matt

Moderator Edit: Updated Subject for Clarity

[Tunah85]

Anyone who has posted in here stating their Ionic is working perfectly hasn't really provided any evidence...

If I just picked one up as my first GPS device and started running I would probably have assumed it was working fine as well.  Would love to see a side-by-side comparison of some 'correctly tracked' runs (i.e. tracked with two devices).

Also, the fact that Fitbit talks about a standard 4% variance is extremely concerning... no-wonder they didn't get it right... that's an 800m difference on a 20km run!  4% is completely unacceptable. 

On as positive note, I had my two most accurately tracked runs recently...

I had a 9.0km run that was tracked at 8.95 on the ionic... and a 7.6km run that was tracked at 7.56km...  Unfortunately in between this was an 8.0km run tracked at 7.85... still way too far off for my liking.

Almost negligible... but not.  IF it was sometimes over by .05 and sometimes under I would accept it... but its always under by 0.75-2% for me still.

I have a 30km race coming up... I am curious to see what the difference will be there.

But seriously, can someone who has there Ionic working 'correctly' please post some side-by-side comparisons to demonstrate just how correct their Ionic is?  (I am still skeptical)

[SunsetRunner]

I don't get the gps problems to the same extent as other posters it seems.
The first ever race I did with the ionic was on a hilly 5 mile course, so
my pace would have varied a lot, and the distance reading came in way
under. Like around 4.6miles. At the time I just assumed that the race had
been wrongly measured, and I had run a bit slowly because of the hills. I
say my Ionic pretty much always underestimates my Parkrun. I think the
worst it's come up with is 2.95miles. Recently, I've been using a fixed
stride length setting, which I estimated to be about right for my racing.
It still always underestimates, but it's more like 3.05 it seems. The only
time I can think that it overestimated was on a recent half marathon where
it estimated as 13.2, but most of the way around it seemed very close. I
think I slowed quite a lot in the last three miles, so if it was using a
longer stride length had been estimated earlier on then I guess that might
account for it. As far as other problems go, it always seems to get a GPS
connection reasonably quickly, and I think only twice has lost it mid run.
Once the trace on the map looked a bit wrong. I think I'll try running in
bike mode for a week and report back.

[Tex52]

@shipo.Just calling it as I see it; if we all have the same firmware, and we are are using it the same way, then the only potential difference is flawed hardware.  If these issues were software related, then everybody would have an issue.

Actually I believe your logic is incorrect. If the step algorithm the firmware uses actually matches your stride somehow then yours will always be correct. It isn't the hardware because I'm on the 3rd one and still incorrect GPS readings while using the Run App. And the distance is spot on using the Bike App. But I still refuse to accept that as a work around when I paid 300 dollars for this device.

[Giampi71]

---

@Tunah85wrote:

But seriously, can someone who has there Ionic working 'correctly' please post some side-by-side comparisons to demonstrate just how correct their Ionic is?  (I am still skeptical)

---

See my posts here:

https://community.fitbit.com/t5/Ionic/Ionic-GPS-Inaccuracies/m-p/2550858#M36433

https://community.fitbit.com/t5/Ionic/Ionic-GPS-Inaccuracies/m-p/2558265#M37269

I will have another run tomorrow morning, around 15k (if it doesn't rain). I will run with T0MT0M, Ionic and Polar and I will make another side by side comparison.

[SunsetRunner]

I wonder if that's because you run at a constant pace with constant stride
length and thus Fitbit algorithms work for you. Can you be bothered to do
one of the other experiments - for example setting your stridelength to
something very very long in your advanced settings, or going for a drive
with your running app on.

[Giampi71]

@SunsetRunner Sure, I will come back to you whenever possible.

[imeson79]

I went for a drive with the gps on set it to run.. distance was 0.0 

i then switched to bike.. funnily enough distance started to measure

[SunsetRunner]

I just tried bike mode for a run - here's what I found:

1. The shortest distance between cues is 5miles - so using cues isn't really practical.  However, the running app cues don't really work properly anyway - I usually have half mile cues, but at the end of the run, the half mile splits don't tally properly with the mile splits.

2. The pace is in mph (or km/h I imagine) and I'm used to min/mile - but I suppose I could get used to it.

3. I felt the pace on the watch didn't jump around so much when I was running, but it did still jump around a bit.  However, as I'm not used to the units of the pace, I might have just been imagining it was smoother.

4. The watch seemed to show I was getting progressively slower during the run.  Perhaps I was.

5. At the end of the run, when I synched, the steps had been counted and shown on my dashboard, the calories and heart rate seemed about right.  

6.The activity showed up as a bike ride, obviously, but I was able to change that with the app on my PC.  I couldn't see how to change it using my phone - there looked to be a way but it didn't seem to work.

7. Once I'd changed the activity to a run I noticed that the dashboard had recorded the right number of floors - I'm not sure if it had done that already, before I changed the activity.

8. The miles weren't shown on the dashboard, even after changing the activity.

9. It uploaded to Strava before I had a chance to edit it, but it was easy to edit on Strava too.

10. **MOST IMPORTANTLY** The run was recorded as 4.43 miles.  Recently it's been reasonably consistent around 4.63 using the running app.  Now that's a big difference, much bigger than I'd imagined, and I'd expected the watch to under-estimate in Running Mode, as it seems to do for Parkruns.  So now I'm confused as I don't know which, if either is accurate.

For my next trick, I will use bike mode for my Parkrun.

[Kierownik]

Can someone from Fitbit kindly explain how is it in their opinion justifiabe to calculate the peace and distance relying on steps instead of a built in GPS?

Seems like the most illogical thing I've ever came across. 

[SunsetRunner]

OK - further to my using the bike app - I entered the tcx file into https://www.sportdistancecalculator.com and it came out as 4.69 miles.  I entered a couple more runs on the same route, and that does seem to be about right.  I entered a 5k race into the website, and it came out as 5.02 - so I'm going to guess that the tcx file is right and that the website interprets it correctly.

So we've got this - I do a run that's 4.69 miles

I can use the stupid run app that uses steps and stride length and periodically recalibrates using gps and tends to give a distance of 4.62 or i can use the bike app that only uses gps and hilariously says I went 4.43 miles.   

So this gives us a clue as to why Fitbit like to use stride length.  In my case at least, it's more accurate than using GPS alone.  The question now is why? If the tcx files are accurate, which seems likely, why can't the bike app get things right?

[the_one]

[the_one]

Here is the run from last night again very very incorrect, I was running close to the river in open space. Had restarted my watch, waited for the GPS to connect before starting the run it still shows me running like a drunken sailor.. 

Can a Fitbit Moderator respond? 

[groberge]

I also had tried bike mode on a known distance run and it was incorrect.

However, like you saw, the tcx file seems to be correct when loaded into other apps.

Given this info, the work-arounds that others have mentioned have not worked for me.

[DavideFitbit]

Hi @the_one, @MurryIvanoff and @bmw54, thanks for taking the time to share your experiences with the GPS accuracy so far, and all the screenshots, I'm sure this will be helpful for other users as well. 

As you may already know, the GPS requires a direct path to the satellites in order to receive their radio transmissions, and there are several reasons why GPS data could be negatively affected.  The accuracy depends on both your environment and the weather; the GPS signal could be affected because you're underground, near many tall buildings, or as a result of atmospheric effects. 

I'm aware that you probably already tried all the basic troubleshooting steps, however, it won't harm to take a look at all the information here about the requirements, error messages and troubleshooting steps that could help to improve GPS accuracy. They could also be useful for other users checking this thread for the first time. Also, updating the firmware could be of help, as you mentioned. 

@the_one, thanks for letting us know that you've already contacted Support about this, I know you've probably tried all basic troubleshooting already. I do recommend that you continue the assistance process with them if all troubleshooting has been exhausted. They will let you know how to proceed. 

@bmw54, do let us know how it goes with the new device.  If you see improvement with the accuracy. 

Thanks for all the feedback @MurryIvanoff, Fitbit continuously works to improve the quality of products and services, and all this information is useful to our team of developers. Thanks again for taking the time to share your experience. 

[bbarrera]

---

@imeson79wrote:

I went for a drive with the gps on set it to run.. distance was 0.0 

i then switched to bike.. funnily enough distance started to measure

---

@shipo have you tried this?

[imeson79]

Make sure you do it when the GPS is connected.. And drive at a runable pace.. you will see that the distance DOES NOT get measured..

Had my Garmin Fenix 5x with Me and that recorded the distance.. the iconic did record the drive.. But distance was still 0.0

So I do think that the distance measured may come from steps calculation

[Ksphs]

That's good to know. My Fit bit has not crashed lately, but it did not
record my run properly today and some how lost approx .25 miles bout 3
miles into my 9 mile run. Not sure why that would have happened.

[Giampi71]

---

@SunsetRunnerwrote:
I wonder if that's because you run at a constant pace with constant stride
length and thus Fitbit algorithms work for you. Can you be bothered to do
one of the other experiments - for example setting your stridelength to
something very very long in your advanced settings, or going for a drive
with your running app on.

---

As promised I had a run this morning wearing the Ionic, a T0M T0M Runner 2 and a Polar m600.

I also set a 200 cm as stride for my run that is around double mine. However I confirm the distance tracked is still very good so my distance seems coming from the GPS. 

Fitbit Ionic: 10.32 km

Runner 2: 10.21 km

M600: 10.27 km

I also uploaded the TCX files on https://www.sportdistancecalculator.com/ as suggested by @SunsetRunner and go the following:

Fitbit Ionic: 10.42 km (+0.1 km)

Runner 2: 10.27 km (+ 0.06 km)

M600: 10.31 km (+0.04 km)

All the devices show some differences and my interpretation is that it depends on the filtering of the GPS raw signal that for the Fitbit seems slightly more aggressive. However we are speaking about a difference of 1% from the minimum to the maximum distance that is definitely small 

However I also did a test in my car and, with my surprise, indeed the distance tracked is 0. So I'm really puzzled. My best explanation is that the distance (at least for me, not sure why it should be valid only for me...) is that the Ionic checks if there is steps (measured by accelerometer) and only in this case it kicks in the recording and does not use the steps stride to measure the distance.

Still I'm confused about so many different results among the users (I'm very happy about mine) and we all are trying to interpret the resuls. It would be very easy if a Fitbit engineer clears the doubt and I can't understand why they keep us such in the dark.

[SunsetRunner]

Hi, thanks for doing the experiments. My guess is that after you changed
the settings of the stride length you didn't sync the watch up, so the
long stride setting info hadn't been recorded on your watch.

[Giampi71]

I'm sure it can't be because I changed my stride on Friday (hoping to run on Saturday morning) and run only this morning (Monday). And I have been synced so many time hoping for the notification of Fitbit OS 2 🙂