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

[tonny123]

---

wrote:

---

wrote:

@Frank00 I wanted to add that you can use apps and wireless payments but then I remembered that most of the apps are not available outside US and wireless payment system doesn't work with most of the banks in my country. So yeah, step counting and sleep tracking are the only features. And it's an attractive accessory of course. Especially with optional 60$ wristband.

---

And... If everyone simply measured their running stride daily and kept this constant while running, this whole GPS mumbo jumbo would be a non-issue. 😉

---

**ahem**! You are so right! How did I miss this!? So I indeed don't need any GPS watches - just a step counter! Now I feel myself stupid buying a 300$ step counter.

[tonny123]

---

wrote:

---

@Frank00 wrote:
Just for the record... The Vivoactive 3 also works with the bike sensors, foot pod and the chest strap HRM.

---

and even my Varia rear radar, but sadly, no power meter support. 

---

Not having ANT+ support is a real bummer. Nor can watch transmit HR data via Bluetooth so that you can see it on the screen of your training machine in the gym. Having Strava app that is mostly used by people who cycle and not having ANT+ support (which most of the power meters use) is beyond my understanding. I don't know who is responsible for product design there. Definitely not a person that do train. At least not the one that take is seriously.

[samgillespie]

@Tim76 Thanks for sharing sadly and coming from an it and customer support background myself this only helps confirm my fears that this is not being taken seriously and leads to a feeling that I/we are being fobbed off. I raised my issue with Fitbit on the 20th Nov  I bought my ionic in early Oct. I have been a fan and supporter of Fitbit since May 2015, I regularly buy about 6 fitbits each year to give to my customers as prizes. They say they have engineers working on this very hard. Well the question is if they have especially with all the data we have supplied. Why can they not after 2 months even set a timeframe, this seems to be a standard go away and wait response. Sadly for me if I don’t get a reasonable update soon I will be taking my business elsewhere. To be honest I am sure Fitbit are not bothered about me doing that, which is why I will likely end up having to do it. 

[SantiR]

Hello everyone! Thanks for reporting this. Your feedback is truly appreciated!

Fitbit is still investigating this, so your feedback is very useful.

Now, if you haven't done so already, perform a restart on your Ionic. You can also try turning off the GPS option on the online dashboard, syncing your Ionic, and then turning it back on. The option should be found on your device settings and should look like this:

GPS Settings

Also, check this helpful article about why GPS isn't working on your Fitbit device for more details regarding this.

Thanks for your patience. Please stay tuned for updates! 

[tonny123]

@SantiR Nice to see someone from Fitbit actually showing up in this thread. It's about time. At the same time, telling us "have you tried to turn it off and on again" - are you serious!? Have you tried it yourself!? Did it work for you?

We would really appreciate someone from the design team actually admitting the problem. "Yes, we were playing too much WoW so we had very little time to implement GPS support properly. Now boss banned Blizards servers so we will probably have to do out job properly and fix this problem. Expect a new firmware in the next 2 weeks."

How about that!?

[Colinm39]

I only use the GPS occasionally and I don't run  but it works for me.. but a BIG but... I always check my satellites with this App GPS Status to ensure I have a good chance of connection and last night the Ionic connected within.. 3m (10')

Here is last night's satellite snapshot  before I used the GPS. another but.. The GPS path recorded was accurate for length but a bit wavy...

[MattyW]

The trouble with threads like this is that everyone uses their watches a
bit differently, and everyone is reporting slightly different problems, so
it's very hard to work out if the problem is, for example, just our
particular watch. Personally, I think the problem, or at least the one I'm
experiencing, is a bug in the running app and not an actual problem with
the watch picking up a good GPS signal.

[samgillespie]

@SantiR I appreciate you sticking you head above the parapet and I am not going to shoot you down but we have done all of this we really need and update from the team working on it and any eta for a fix. I appreciate that this may not be available at the moment and there is no update. It’s just you have a bunch of unhappy customers due to this and believe me I know that isn’t good 

[MattyW]

OK everyone - here's a possible unsatisfactory workaround - at least to the problems I was experiencing.  Now personally I don't care about distance and pace all that much if I'm just out for an easy run, but I do care when I'm racing or doing a hard training session (I'm rubbish at it by the way, but quite keen).  The running app has definitely been recording slower pace/less distance at the beginning of my races, but this seems to sort itself out by the end of the first mile.  The algorithm definitely does seem to use stride length and cadence at the beginning of a run, for some reason, even though GPS is connected according to the watch.  If a runner tends to always run at a similar pace, then I don't think they'll notice anything strange.

I looked at stride length in the settings I notice you can enter your own value, or have it set to auto, which apparently recalculates your stride length after each run.  So if you do a very slow warm-up or a slow training run before doing a race then this data is used, and presumably it's a short stride.  However, we do have the option to manually enter your stride length so perhaps we can work out our racing stride length from a previous race and manually set it.  It sounds like an unsatisfactory solution, but it might just work.  Unfortunately, I'm rehabbing a slight injury, so haven't tried it out properly, but I did do a slow run this morning to try out my theory.  I set the stride length to a ludicrously long 200 inches, and as I jogged up the road I was apparently much faster than Usain Bolt at full tilt - 2min/mile (initial readings might have been 1 min/mile actually).  The first "half mile split" was in just over four minute mile pace - which will impress my mates on Strava (if only it would sync) .  I think I was actually trundling in 9 or 10 minutes/mile.  Anyhow, the pace gradually corrected itself and seemed about right after five minutes or so - the second half mile split was at 8:30 pace. I know the distance of the route I ran to be about 1.7 miles, but it was clocked at 2.05.

Anyhow, it seems to me that the algorithm starts with a pace dictated by stride length and cadence then immediately starts approaching the real pace as dictated by GPS, but does so gradually.  It seems possible that if I estimate my racing stride length pretty accurately at the beginning then the GPS measured speeds might actually be approached reasonably quickly.

I have also been having a problem when I've been doing fartlek runs - when I speed up, the watch takes too long to catch up the pace I'm actually running.  I also think it's slow to react when I slow down after a fast section - I suspect it might be the case that every time there is a big change of pace then the watch reverts to using stride length/cadence, but I can't really check this out until my foot is better.  Anyone else fancy joining in these experiments?

[bmw54]

MattyW's experiment confirms my own observations and what I've learned chatting with Fitbit reps:

That the Run app uses stride-length multiplied by step-count to measure distance and pace.

Additionally, every 10 minutes the software uses GPS data to re-calculate pace. 

This strategy should smooth out the jerkiness we'd see if ionic was using pure GPS data. But the strategy is flawed when racing - especially the 5K or shorter distances where your pace is significantly faster than training pace and the event is over in less than 20 minutes. 

Even when racing longer distances, like the half-marathon, I never see improvements to the pace my ionic displays. It's only when I see the official results after the race that I get any idea of my average pace.

I haven't performed the update we were advised to do yesterday. Hopefully it will correct problems with their already-flawed strategy.

[MattyW]

I'm pretty sure my Ionic is factoring in my GPS more frequently than once
every ten minutes. On the run I did this morning, with my ludicrous
giraffe stride length entered, even though it started by showing I was
running faster than Usain Bolt initially, it soon started slowing until it
seemed pretty plausible after five minutes or so. I reckon they've got some
relatively complicated algorithm in there that tries to take GPS, current
stride length/cadence, and immediately preceding estimated pace all into
consideration, and they do that because they think it will minimise
problems if people are getting an occasional anomalous GPS reading and it's
an attempt to stop current pace jumping about too much(which it sometimes
does anyway) . It probably seemed OK in testing because they only used it
at one constant pace per tester.

That's my guess.

I reckon my surge might well have done something similar, but was much
quicker to slot into the correct current pace, so it might have been harder
for me to spot any anomalies.

I'm pretty sure I read that they were having trouble with the GPS on
prototypes of the ionic. Perhaps they started developing an algorithm that
was less dependent on GPS because of that.

[bmw54]

It's encouraging that you're finding Ionic re-calculates stride-length earlier in the run and more frequently than every 10 minutes. Maybe Fitbit is correcting some problems. 

We should consider that using a blend of GPS and stride-length has benefits. An example:

In this training run Ionic lost GPS midway. The route mapping was straight-lined, but distance is correct (verified later on using mapmyrun) and overall pace was correct. (note the loss of mile markers between 6 and 12 - that's where GPS was lost.) 

Training run where GPS was lost midway 

[BahGahWah]

@SantiR I have experienced the ionic recording short distances on runs.  I have raised this via support emails, but it is quite exasperating to keep receiving stock responses, saying power cycle the ionic or its environmental issues.

I have taken some example runs with known measured distance of 5km and imported the .TCX file into a 3rd party site.  This demonstrates the issue with the ionic;

Run 27th Jan - ionic measure 4.73km - .TCX import measured 4.9km.

Run 9th Dec - ChargeHR / iPhone measure 5.04km - .TCX import measured 5.04km.

From this I assume the GPS reception is not the issue but the Fitbit algorithm for calculating distance implemented in the ionic. 

[bbarrera]

---

wrote:

We should consider that using a blend of GPS and stride-length has benefits. An example:

In this training run Ionic lost GPS midway.

---

Really? Stride-length estimation of distance is a bug, its not a feature. 

From your screenshot you lost GPS for roughly 5 miles (between mile 6 and 12). At a 9 min/mile pace your Ionic lost GPS for 45 minutes!!! Even at fast 7 min/mile pace that is losing GPS for 35 minutes!!! Something is seriously wrong, either with your Ionic, or the Ionic's firmware in general.

I've been GPS tracking walks, hikes, runs, and bike rides since 2009, starting with Runkeeper on my iPhone. Never have I seen a 5 mile loss of GPS in thousands of tracked activities. The last 2+ years I've got over 800 bike rides on Garmin GPS. Sure I've seen a little "wandering" for a minute or two, especially on older phones from 2009-2011, where the GPS track was off by a couple hundred feet. A few times in the mountains, riding in a deep canyon or steep section of switchbacks, maybe a quarter mile off the road but again only for a few minutes.  

[Smackems]

At the end of the day the way the running app is interpreting the data is just incorrect.  I the same route in my neighborhood and the Ionic has done everything from overstate how much I run by .2 to drastically understate how much I run by .2 of a mile.  It's a 5 mile route and typically the Ionic will overstate on recovery runs and understate on faster paced runs.  It's super frustrating to have such wildly inaccurate data when training.    

Now - if I look at the GPS map or export the file everything is actually correct so I have no doubt that the GPS hardware works fine.  At this point it seems that Fitbit isn't in any hurry to address the issue nor do they care enough to provide an actual update.  Our best hope might be that someone else develops a running app in the Fitbit app store that only utilizes the GPS for tracking.   

[MattyW]

Do you find the inaccuracies to be at the beginning of the run too? You
could try altering fitbit's stride length to suit the run you have
planned. I know it's a cludge, but it might work.

[BahGahWah]

Not sure, comparing the 1km marker from previous runs with the ionic runs, there is an obvious difference in the position.  Roughly measure it looked like it had lost about 50m in the 1st Km.  Average this across 5km and would be consistent with the 0.3Km loss.

Was also thinking to implement your suggested fix, however looking at past runs, the number of steps I take on the same course varies.  The faster my time, the fewer steps I take.  

I have seen the auto stride length set different lengths.  If the preceding run was at a slower pace then I can assume the stride length will be a shorter distance.   This could correlate with the beginning of the run being less accurate if running at a different pace.

Using stride length and number of steps as a distance measure when there is no gps signal makes sense.  As a primary measure of distance can’t see how it would accurately work due to the variance of stride length with pace.   I guess users with consistent pace and known stride length would not notice the deviation.  Sadly my running is not that consistent.

[Sfharris]

The GPS on this device seems to be very seriously inaccurate, and struggles to maintain signal.  I just ran the London Winter 10k - but FitBit thinks I ran 11k.  This is 10% out and this level of inaccuracy makes the device pretty much worthless for judging how my pace is improving, and if I'm getting in the miles I need as I train for a marathon.  I'm really disappointed that I bought this - at a high price point - and it seems not to do as well as the TomTom device I replaced it with (which cost about half the price).  Can someone at FitBit comment? Unless there are plans to fix this, I will be returning it as unfit for purpose.

[BahGahWah]

Hi, other users have reported similar issues. This was first reported in October.

https://community.fitbit.com/t5/Ionic/Ionic-GPS-Inaccuracies/td-p/2229217

From investigating the issue it appears the ionic Is not using the GPS signal to measure distance.  In order to verify if it is a GPS issue, I tried the following;

From the Fitbit Dashboard, download the .TCX export file for the activity.  

The following website allows you to import the .TCX file and will give you the GPS measured plot of the run.

https://www.sportdistancecalculator.com/import-gpx.php

If the plot of the run looks wrong, or the measured distance is not a you expected then it may be a GPS issue.   If the GPS data is as expected  and the distance accurate within a few 100meters then its the ionic calculation of the distance.  

If if you try for other known distances then you will soon realise it’s not working as intended.  

[SantiR]

@tonny123 @Colinm39 @MattyW @samgillespie @bmw54 @BahGahWah @bbarrera @Smackems @Sfharris I hope you're doing well! Thanks for sharing your experiences on this thread. I'm sure this will help other users that may be experiencing something similar with GPS.

Remember that our team has received the feedback from everyone in this thread and is working to investigate these reports further. I will share updates from the team when I hear them.

Thanks to everyone for your reports and understanding. In the meantime, I recommend checking the best practices on this post to help GPS accuracy (If you haven't done so already).

Thanks to everyone for your reports and understanding. 

See you around!