Over the past 7 days there’s been a slew of folks with certain GPS watches seeing accuracy issues, whereby the GPS track is offset from reality (usually in a specific direction). This often results in GPS tracks where the visual pattern might be correct, but it’s not in the correct location. Meaning, it’s as if someone shifted your route slightly to the side, ultimately letting you walk on water, or run through buildings. Alternatively, it just might have weird gaps, or be outright dumpster-fire wrong. In theory, the issue should be resolved as of last night.
This issue specifically impacted those with Sony GPS chipsets, which is essentially a repeat of an issue that occurred 18 months ago that was far greater reaching (for a variety of reasons). The first time, the issue happened over the New Year’s Holiday Weekend, and was the first time these companies saw it, which significantly slowed the response rate (and also, increased how many people it impacted).
This time around, the impact was certainly measurable (as my inbox can attest to), but the duration was more limited. It initially impacted people last Tuesday/Wednesday (June 28th/29th), then again over this past weekend and into the early week (July 1-July 5th). The number of impacted watches is vast here, and essentially any Garmin watch released in the last few years (but not most of those in 2022, as they use a different chipset), as well as all Suunto/Polar watches, and theoretically all COROS watches except the Vertix 2 (though, I haven’t seen any COROS reports, so perhaps they’ve put in place a different filter).
Roughly speaking, that’d be the following watches (plus the 29 million variants of these watches Garmin has):
– Garmin Forerunner 45/245/745/945/945 LTE/Fenix 6/MARQ/Vivoactive 3/4/Venu
– Garmin Edge 130 Plus, Edge 530/830/1030 Plus
– Garmin Enduro watch
– Garmin Instinct Series 1 & Series 2 watches
– Polar Vantage Series, Grit Series, and Pacer Series Watches
– Suunto 5/5 Peak & Suunto 9/9 Peak
Again, the above list is not all-inclusive. There are tons more watches impacted, especially from Garmin, which uses near-identical watches in various other divisions such marine, golf, and aviation. Additionally, this would theoretically impact the Wahoo RIVAL & COROS watches (except Vertix 2), though, I can’t find any instances of people with issues there. Which means either they were reported, or they’ve put in place mitigations.
Meanwhile, the Garmin and COROS watches not impacted are:
– Garmin Fenix 7 & Tactix 7 Series
– Garmin EPIX Series
– Garmin Forerunner 255 series
– Garmin Forerunner 955 Series
– Garmin Edge 1040 Series
– COROS Vertix 2
All of those watches use a different multiband chipset from MediaTek/Airoha, which isn’t impacted here.
The issue has to do with the ephemeris data file, also called the EPO file (Extended Prediction Orbit) or Connected Predictive Ephemeris (CPE). Or simply the satellite pre-cache file. That’s the file that’s delivered to your device on a frequent basis (usually every few days). This file is what makes your watch near-instantly find GPS satellites when you go outside. It’s basically a cheat-sheet of where the satellites are for the next few days, or up to a week or so.
Your watch or bike computer automatically gets this file via Bluetooth Smart from your phone, WiFi, or USB, depending on how you connect your watch. Most companies deliver it anytime your watch syncs and needs a new version. So from your side, you never do anything – it just quietly happens in the background. The data in this file was wrong, and thus the data your watch uses for those first few minutes is also wrong – leading to the offsets.
The good news is the fix is simple: Just sync your watch to get an updated copy. Garmin for example says that if your watch is connected to your phone, as of last night, it should have/be downloading a new EPO file, which will fix the issue. Garmin also says they’ve put in place blocks to ensure this doesn’t happen again. Albeit, they said that 18 months ago too, so my assumption is whatever Sony did this time broke it in a new and special way that wasn’t the same as last time.
With that, hopefully that helps folks understand what’s going on, and how to fix it.
Thanks for reading!
FOUND THIS POST USEFUL? SUPPORT THE SITE!
Hopefully, you found this post useful. The website is really a labor of love, so please consider becoming a DC RAINMAKER Supporter. This gets you an ad-free experience, and access to our (mostly) bi-monthly behind-the-scenes video series of “Shed Talkin’”.
Support DCRainMaker - Shop on Amazon
Otherwise, perhaps consider using the below link if shopping on Amazon. As an Amazon Associate, I earn from qualifying purchases. It doesn’t cost you anything extra, but your purchases help support this website a lot. It could simply be buying toilet paper, or this pizza oven we use and love.
Like clockwork (*boom tish*), the topic of EPO always comes up at Tour time! 😉
Happy to see your article that I was not the only one. I had the same strange GPS deviation this morning. I am using an Epix with Alpha 9.14 and synchronising regularly.
Before my run this morning and to compare with my run from yesterday, I changed from Multiband All Satelites to GPS Only. The picture of the run yesterday was spot on. The GPS Only picture this morning was horrable.
GPS only is indeed the ‘least’ awesome accuracy option, though, I’d agree that’s not idea. That said, the EPO issue wouldn’t affect your Epix. That sounds like just plain and simple bad GPS.
Was this a problem for all users or random, or depending on use case? Just checked my two tracks from the weekend on a new-ish Edge 530 and they seem bang on, right from the start. First one I was following a pre-made course, second was tracking a club ride – or did I just get lucky?
It essentially depended on whether or not you got a bad EPO file (satellite pre-cache file). Basically, the EPO file lasts a few days before it expires, so if yours hadn’t expired, you could have skipped along just fine.
I was thinking that Instinct 2 uses “old” Sony chipset but I cant see it on list of affected watches? It isn’t on the list of watches that aren’t affected either.
Yup, it does, good point!
Came here to make an EPO comment, GPLama beat me to it.
😉
The tracks on my run was all okay at the start, went bonkers in the middle, and was back to normal towards the the end. My phone was in the car. I am guessing the CPE file is not just used at the start of an activity?
Oh, I have still not figured out how to get training metrics like v02max to disregard the wonky GPS activity, even after deleting the activity from both my watch and Garmin Connect? They should put in place an option to disregard a badly recorded activity from training metrics, heat maps, challenges, etc.
Happened with my Fenix 6 Pro. The recording of my Monday recovery run was indeed completely screwed up with the start off by well over a km (and I had waited for a GPS fix before running). The position got more or less accurate 3 km into the run (which resulted in a spurious acceleration).
Luckily, the triathlon recorded the day before did not suffer such failure. I would have been somewhat upset ;-)
EXACTLY what I had. GPS seemed to be OK to start my run but the track and measurement were completely off for the first couple of k’s.
That was Monday, today (Wednesday) everything was back to usual.
What aggravated me the most during this episode was Garmin’s customer service, or lack thereof. No acknowledgment or worthwhile response to the seemingly hundreds of tweets they received about it.
So, what caused this issue? Why was the EPO file incorrect? So many questions. I don’t understand why the chipmaker would be the one providing the location data for satellites, but since only Sony chipsets were impacted, that seems like the only explanation. So is Sony messing up this file?
Just guessing but I assume Sony gets the location data and builds the file in the format for their chipsets.
“So is Sony messing up this file?”
Yes. :)
And then from there, the manufs just pass it along. When it went wrong last time, companies built in protections to guard against that failure from occurring. But it sounds like this was a different type of EPO file failure. Said with a bad analogy, last time it was overbaked, and this time it was undercooked. Both are wrong, but detecting one is different than detecting the other.
Hi, I did see this on my Garmin Enduro last week. Does the Wahoo ELEMNT ROAM have any issues with this, because I did see some offset position on my Roam during cycling on my map page.
Yup, ROAM uses same GPS chipset, so it’d likely be impacted.
Garmin Forerunner 955 was impacted too:
link to strava.com
So I asked Garmin about this. They haven’t heard of any of the Fenix 7/FR955/etc (new chipset) ones having any issues (logical, since it’s not Sony).
They looked at your specific linked activity, and believe this was more of a general GPS acquisition failure at the beginning, and then didn’t get a good lock for much of the duration. Also, since it happened this morning, it wouldn’t likely have been impacted if your phone was syncing behind the scenes.They said they’re keeping an eye on it of course, but at present, this issue shouldn’t be impacting any of the newer devices.
My Tactix 7 is affected as well. I did a run on tuesday and after nearly 2h into the run the signal drifted off to the northeast.
Maybe it’s simply a solar storm :)
Your article did not address the obvious question, how did Garmin got it wrong two times in a row this time. As all impacted users reported, it was bad, then good for a few days, then bad again. This can’t be brushed off as a simple repetition of the alleged EPO file story a few years ago.
Alleged? Take off the tinfoil hat.
As many have pointed out, this wasn’t just Garmin that was impacted, but Polar, Suunto, Wahoo and more. Or did that not fit your John Doe tin-foil hat conspiracy theory?
As explained above, the reason it was good for a few days was because a proper EPO file was sent again. Then broke again.
My epix gen 2 was affected too. Firstly In the run on 26/06 when suddenly my record for 1km was 0.00 sec (epix added 1km to my run), than on 01/07 run it appeared that i was running on the water like Jesus (quick run about 150m in the water than return to land), but I didn’t catch any fish. I hope the problem is gone now
Once is nothing, twice is a trend. So let’s meet in DC Rainmaker’s comment section in 18 months to discuss the same issue.
Yep, a lot of odd Strava activity.
Yeah, Strava needs to build into their algorithm the ability to ignore activities from records where the pace exceeds world record pace.
Totally agree. This would help eliminate Course Records where someone posts a bike ride as a run, or a car ride as a bike ride, etc.
Got hit with the “dumpster-fire” type with my Vivoactive 4 on July 3, but it tracked normal on July 4.
Same with my Forerunner 945. Open water swim on July 3rd was dumpster fire. Bike ride July 4th was back to normal.
I don’t get it. If EPO file makes finding satellites easier, it should eventually, during a run, correct itself and get back on track. After all it’s the satellites that inform you of theirs and eventually your position…
That explains why I walked on water, and an extra 1.2 miles over the weekend lol. Suunto 9 Baro here.
I also had this problem … in my Honda which has an in-dash GPS made by Garmin. Is it possible that the ephemera problem exists outside their sports and outdoors product line?
I know absolutely nothing about Garmin car GPS thingies, but, my assumption is that Garmin buys chipsets across multiple product lines from likely the same vendors.
The EPO file is being downloaded from the internet. Your in-dash unit probably isn’t connected to the internet.
“Your watch or bike computer automatically gets this file via Bluetooth Smart from your phone, WiFi, or USB, depending on how you connect your watch.”
GPS units can also download the “ephemera” data from the GPS satellites but it’s very, very slow (which is why units try to download it from the internet).
GPS units can use stale EPO data (it just takes longer to lock in your location). It appears that, in this case, the EPO file being downloaded from the internet was broken for a short period of time.
Additional to problems with the epo files, there was a little event calle G7 summit in Garmisch and in Orts also in Munich. The US shifts GPS positions for the summit area as long as the summit lasts.
This explains why I ran over buildings on 28-Jun-22. Didn’t know what was the problem with that GPS glitch. Thanks for the post.
Did these people steal your article?
[Removed stolen article link…thanks!]
word for word “clive” did.
Wow. Pure rip off.
Sigh…yeah, it happens. Trying to take them down is just playing whack-a-mole and unfortunately usually a giant waste of time.
(I’m going to zap the link up above, merely so it doesn’t give them more search/Google credibility. Thanks for the heads up!)
It’s nice to see my track featured on this blog :) Thanks Ray for showing that it’s not just me, now I can be a little bit more relaxed.
Haha… thanks for being a DCR Supporter Ciprian!!!
Funny! That route on the picture is passing my home in Munich… :)
Tengo un SUUNTO 7 y efectivamente revise el mapa en la ruta del sabado 2de julio y presento un desplazamiento fuera de la que realmente seguí afectando las demás métricas espero se resuelva el problema
Thanks for the artcicle DC.
It happened two days ago to my fenix 6 tracking my walk.
I thought was a problem of new watch.
Thanks for the artcicle DC.
It happened two days ago to my fenix 6 tracking my walk.
I thought was a problem of my new watch.
I’m glad to see this post as I experienced it on June 29th and was wondering if there was a global issue
Do you fix the Wahoo Roam the same as described for the Garmin? My Roam has definitely been effected.
I wear a FR955 and I experienced this on July 1st in the middle of a hike 3 hours into the Canadian rockies backcountry. Thankfully the day and path were clear. It was not an initial acquisition problem as the beginning and the end of the activity looked ok, it was after going through a tunnel when things went awry.
The funny thing is that only the watch map was affected, in the GC map looked correct and strava showed only a spike.
Would this also impact height data? Did tweo 100km Rides this weekend, both about 100km / 1000m elevation gain and my Polar Gritx reported 3000-4000m of elevation gain. Did a factory reset today because of this.
Most height data is also from the GPS Satellite positions, so yes, it would impact height data
Oddly, it shouldn’t have though. The Grit X has a barometric altimeter in it, which means the data is coming from there, and not GPS. Whereas a lower-end GPS watch like the Polar Vantage M or a Garmin FR55 would have a GPS-based altimeter, which would be impacted.
Yeah ist was weird. After the factory reset everything seems fine again
Haven’t connected my Garmin to the internet for any sync for over three years. I manually copy the FIT files to HDD and then upload to necessary locations. Never taken my Garmin more than 60 seconds to acquire signal. When uploaded to certain sites the accuracy of the GPS coordinates is also perfect. Don’t trust Garmin (or any manufacturer) not to put a kill-switch in a future update to purposely trash the watch so it forces you to buy a new one. The worst part is that even if they did do this, and it was reverse-engineered and proven – they can just claim a ‘coding error’ caused it and never face the consequences.
What about Apple Watch? I’m seeing this weird GPS anomalies in some workouts, such as the one in the attachment…
…taken from the NRC app…
They don’t use Sony GPS chipsets to my knowledge, so whatever is going on there would likely be something else.
Just fyi: Same issue with Epix 2 on last Friday. See screenshot below
is there a solution
It only happened that one day
Vivoactive 3? Doesn’t that use an older, non-Sony chipset?
The track on my 935 was initially wrong for a run I did on Friday, but that may have just been a coincidence for not running with it for a while.
So that’s what it was. On the 29th I was about to go for a ride, navigating so I could use ClimbPro. Recently I haven’t been turning my Edge 830 on until I was ready to go. Turned the 830 on, woke up and paired the sensors, found the course on the Edge, but when I hit start it said something like “do you want to navigate to the beginning of the course”, which I found odd since I was at the beginning of the course. Then I looked at the map and noticed that it showed me where I wasn’t. I waited until the position it showed was where I actually was, a minute or so, and then successfully started the course. As I understand it, the EPO.BIN file is used to quick start GPS lock, but it will get the current ephemeris from one of the satellites and use that. The track looked fine after I started. Since then I’ve reverted to turning on the Edge 5-10 minutes before I start, and it’s had a good lock every time since.
None of the posted tracks are as wonky as mine. The last couple of kms are OK but everything else is all over the map (literally :-)). Strangely, the distance and pace numbers tally exactly to my feel on the run, so I didn’t notice until much later.
Garmin 645 and I noticed almost the exact same pattern on my run today. It shifted a section over on a section of my run and made it look like I ran a 0.25 mile that I didn’t run. Noticed it that other time 18 months ago, too.
Also had this issue – about 4km into a run I noticed my current lap pace was down around 8 minutes when it should’ve been in the high 4’s so knew something was up. Then a few k’s later it went to warp speed and seemed to pretty much catch up. Checking the map later showed it went haywire during that period. I especially loved that but where I seemed to cut through about 10 blocks of houses 🤣
It’s a shame as I think your proper track would have looked just like a player from Among Us.
Have same problem with Samsung watch connected to the Samsung note 20 ultra. GPS showing location, wichs is almost rkm from real location
I noticed this on my Vantage M, which I use for runs. Took me two runs to figure out it was not simply wrong, but rather “precise but not accurate”.
In the same time span tho, my Garmin 530 was as good as ever. Do bike GPS use different chipsets?
It could have just been that your Edge 530 didn’t download the dodgy EPO data.
This kind of issue in fitness activity tracking is just a PITA, but I’m assuming Sony GPS chipsets must also be used in devices where the consequences could be a lot more serious (e.g. marine, aviation, emergency response). I’m sure an oil tanker or an Airbus is going to have backup systems, but I’d imagine there are still plenty of dinghies and Cessnas that primarily rely on these devices, I wonder what the fallout in those areas is, or if they have better safeguards?
Another reason to use Stryd for distance / pace.
Seen the issue 18 months ago (on 945), and had weird tracks last week (again on 945). But on both occasions I didn’t care a lot, since my pace and distance were coming both from Stryd.
Does anyone know how to show the laps on the map for Garmin 245? I’ve emailed Garmin but they will not respond.
I have Fenix 7X Solar Sapphire and I went on jog yesterday in eastern U.S. and had apparent issue of my route into buildings and generally shifted throughout entirety of route. This was set on All plus Multi-Band.
THANK YOU SO MUCH. I felt totally bonkers. I have a FR 245. I had a short run briefly impacted on June 28 (a skip in the GPS data between two points), and then a 15 mile run on 7/3 that had a totally bizarre GPS map — like you said, shifted over so I was running through parks and houses, but at a few locations in my last mile it just skipped from one spot to another (at one point I was apparently running a 2:30 min/mile pace, that’s cool).
Hasn’t been an issue since Sunday and I’m very happy to have an explanation. Thanks!
I noticed this with a few walking activities last week whilst on holiday 400 miles from home. All the walks ended back at the start location but four of seven activities show a different start and end. I think these were the ones where I started walking as soon as I saw a green “GPS acquired” message. The other three that weren’t affected were over the same period of days but I selected the Walk activity before I started changing into walking boots, etc., so the GPS had around 5-10 minutes to acquire signal and then settle down to an accurate location. So, I’d guess that even if the EPO had some errors these are only significant if the GPS doesn’t have time to completely acquire signal and calculate its position and uses the ephemeris data instead until it can get an accurate fix. Waiting a few minutes might be enough to circumvent the problem.
Do you reckon this track is an example of it? Recorded on a 645 Music on Sunday
link to strava.app.link
Yep – walkin’ on water! (or an astronomically low tide!)
Yes. On July 1, I took my Garmin 245 for a swim in open water. Although I walked out about 20 feet and headed due west, my watch decided I went about 300 yards due north and then back, before charting my true course to the west. My speed appeared as though I had been waterskiing.
I have a frequent issue. When I’m doing intervals with sprints, and a sprint is on a corner, the GPS route will cut off the corner. This makes it appear that my route was shorter than real life, and thus that my speed was slower too.
This is with my Fitbit watch recording, or with the Strava app on my phone recording. I think the Fitbit uses my phone’s GPS for location.
Any idea how to fix it?
Easy fix and future-proof: warm up in place for a few minutes. You should be doing it anyway. Your watch will pick the satellites properly meanwhile.
Not really future proof. GPS was fine for a lot of us at the start of our runs, but then went crazy during the run. Garmin/Sony needs to tests the CPE files before pushing it to devices.
My Spartan Suunto has just stopped recording via gps since this problem has been reported, runs are longer and bikes don’t record the distance at all, soft and hard reset didn’t work. Very frustrating. Their solution is to send off the watdh.
Rainmaker Please talk slower. We watch all your utbe stuff but you talk way to fast. Many comments here about it.
Thanks P B
Well I made sure to sync my forerunner 35 before my run today but the accuracy is still very poor. Not quite as bad as on Saturday when it was outstandingly poor. (See image of two identical parkruns).
Just a thought, but could any of these gps accuracy issues have anything to do with what’s happening in Ukraine? Is accuracy of the systems that we (and possibly the Russians) rely on being deliberately fudged? Just a thought.
My Forerunner 45 tracks still look like a drunken stagger…
955’s seem to be having terrible GNSS issues as of today. Lots of reports on the Facebook group saying the same. My watch completely lost its position twice for up to 10 minutes on my last run.
Thanks for the article DCR. This issue still seems to be happening though. My run this morning looked like I was flying over the water instead of my regular neighborhood run. The distance reported is somewhat accurate.
How scary it is that so much depends on one file. If sony stops supporting this chipset it will be almost useles?
Hey Ray. This is happening to me right now on the brand new Forerunner 255 with the new chip.
Just bought it yesterday, and out of 5 activites tracked, two had this problem. The map overlay was way off. Even weired, in one activity, for the first half the overlay was off, but then it started to follow the actual path.
These are the activities:
link to connect.garmin.com
link to connect.garmin.com
Any idea what to do? The watch was synced and received the latest update.
Got problems yesterday and today (November 05 and 6).
Does anybody noticed the same?
My Coros Pace 2 had rock solid GPS accuracy for a year but has suddenly gone haywire for 3 consecutive runs since November 5th. Really strange, I opened up a case with Coros.
Really the issue isn´t only due to EPO / CPE files delivery to the watch, in the FR 945 and maybe in all the watches with this SONY sensor, the real problem is the hardware.
In my personal expeperience (fr945) there isn´t any (official or 3rd party) solution, Garmin only said “your watch reach the end of its life, buy a new one”…
That´s why i decided to switch to other brand.
Garmin watches are fine. It was a bug. I’ve had a Forerunner 255 since launch (summer 2022). This error happened to me twice in the first two days – then it never happened again. The tracking is perfect, and I do it in tough conditions. Never had a problem since.
Yeah, it was really just a one-ish time bug, and impacted everyone at the time. Since then, almost everyone has changed chipsets to Airoha/Mediatek (Garmin included), so…not sure it really matters anymore.