In the comments section of my recent review of the Garmin 910XT a friend mentioned that he had seen reports of GPS accuracy issues for the Forerunner 620. Fellrnr also has done some extensive testing of various GPS devices and his 620 did not fare well. Garmin has acknowledged this problem, and has a sticky post on their Forum announcing a software update to address the issue (reports of the effectiveness of this update seem mixed).
I’ve had the FR 620 since Christmas, but have not spent much time using it yet since I fell in love with the Magellan Echo + iSmoothRun App combo. Now that my Echo review is done, I’ve transitioned to using the 620 again (with the Echo on the opposite wrist) and thought I’d take a peek at my recorded runs to see if any GPS track anomalies might be found. Furthermore, I also wanted to hear if anyone else who reads this blog has had issues with accuracy on their device.
In looking at my run maps on Garmin Connect, I will say that most of the time the 620 seems to record my tracks accurately. However, there are times when I have noted that it seems to lose track of itself for a bit and the tracks look a bit funny. For example, a look at the track below would suggest that I took a detour up a hill through the woods to the right, then crossed back over the road for a dip in (frozen) Snow Pond, and then drunkenly made my way back to the road near the intersection of Shaker Road and Snow Pond Road:
The entire stretch of road here is tree lined on both sides, though the part where the track deviates from the road is at a lower elevation and is bordered by a large forested hill to the right. Perhaps being stuck between a pond and a big hill has some effect?
In fairness, I ran the same route the next day with the 620 and it recorded accurately, so maybe it was just having a bad Wednesday:
Below are two iSmoothRun (iPhone app) track recordings for the same stretch (on different days) – both are rock solid:
Below are two maps from a run I did last weekend where I had both the 620 and iSmoothRun recording at the same time. The 620 recording is on the left, the iPhone recording is on the right.
Note that shortly after the 1-mile mark the track recorded by the 620 deviates from the road until about the 2 mile mark. Interestingly, the track appears to be the right shape, but it’s as if that entire segment was shifted south a bit. The rest of the recording is solid. Once again, the track recorded by the iPhone is dead-on the entire time. This particular route is a flat neighborhood, not a lot of thick forest or tall buildings or anything like that. The only thing in common with the Snow Pond run above is that the area just above the mile 4 marker is a wetland with a stream. I read somewhere that water bodies can mess with GPS tracks, can’t recall where that was, and no idea if that might have anything to do with the anomaly.
Here’s a blow-up of the 620 track showing me running through backyards and detouring through living rooms:
Again, in fairness, I should point out that I have run this neighborhood on other occasions with the 620 and the recording has been accurate:
And here’s another where there was a more moderate southward shift of the track in the same spot:
So what can I conclude from this? I’d say not much yet given the small sample of runs that I have to work with so far. The 620 seems to be accurate most of the time, but it also has moments where it seems to hiccup a bit. This is probably something that happens with most GPS devices on occasion depending on satellite connections, environmental interference, etc. The occasional lapse does not concern me too much since I expect that a watch on my wrist talking to a satellite orbiting the planet in outer space (just think about that for a second!) will occasionally have issues (as my Garmin 205 did during the Hartford Marathon, really bad timing there…). The question ultimately be how frequently these issues occur and how bad they are.
Given that Garmin has acknowledged an issue with the 620 and has pushed a software update (I just updated my unit yesterday) there is some reason for concern given that this watch costs $399. I’d say it’s worth keeping an eye on your 620 if you have one, and report accuracy issues to Garmin if you feel they are happening too frequently. I plan to keep an eye on my tracks going forward as I put the 620 to everyday usage, and will report if I frequently see issues like the ones posted above (I’m especially curious if the update makes a difference, off to run that same neighborhood right now!).
How about you – if you have a Garmin 620 have you noticed any accuracy issues?
You might also like:
Garmin Forerunner 10 (FR10) Review: Great Performance in a Small, Low-Priced Package
Running Gear Review: Garmin Forerunner 205/305 GPS Wristwatch
Garmin Forerunner 610 (FR610) GPS Watch Review
Garmin Forerunner 910XT GPS Watch Review
Garmin Forerunner 15 (FR15) Review: Activity Tracking and GPS in One Watch
About Peter Larson
Thanks for this post, Pete – keep us updated as to how the software push changes the 620’s accuracy.
As for GPS not playing nice with water, all that I’ve heard on that front is that it doesn’t do well when actually submerged underwater (even as little as an inch), so it doesn’t perform well when swimming. DC Rainmaker has written about this a bunch (his solution = put it in the back of your swim cap). But simply going near a water body shouldn’t produce any interference.
Cool, thanks for the info. Can’t remember where I saw that, was scouring the web the other night for info on the 620 after I got Brad’s comment on another post.
Just a thought as to why the FR620 (or a bunch of GPS devices, for that matter) don’t work too hot underwater:
GPS signals are radio frequency signals with an embedded time code, which is to say they’re a form of light. Water quite significantly refracts and distorts light, which would throw off any signals being received.
The same thing happens from the atmosphere – which is why WAAS and LAAS systems were developed for airplanes to significantly increase the accuracy of standalone GPS. It’s also why our best pictures of the universe come from the Hubble space telescope – it doesn’t have to worry about blurring from atmospheric distortion.
overall I love the Garmin 620, but it is off by a little bit most runs, and that adds up when doing long runs. I am training for a marathon and ran 15 miles today, but the Garmin seemed to get stuck during part of my run in the woods, and just didn’t record accurately. I use both an iPhone app and the Garmin, but for something this expensive, I expect it to perform better. And I did install the update- no difference. The heart rate monitor is the best I have used, and the ease of use if much better than with prior Garmin’s. Still, being off by a mile on a 15 mile run is disappointing. I purposely chose a path where I knew the exact distance so that I could see if the Garmin was inaccurate, and it was.
No GPS issues, but have had an annoying reboot issue. Run a workout, hit save, hit done, then about 2-3 minutes it reboots. It has lost one activity in history, but oddly the FIT file was still on the watch, but watch couldn’t see it. I have an open ticket with Garmin, they said I could exchange, but 1) I’d be out of $198 bucks until I got my watch back to them as they would send me a replacement or 2) I send in my watch and 10-14 days after they get it I get my new watch.
Reboot issue is not every run nor on every programmed workout. Just odd.
I have the exact same issue Alan! It didn’t happen this morning, but it’s happened often enough to bug me. I’m not sure how big our exchange window is, but I’d just as soon wait and see if it gets worked out in a future update.
Hi Pete,
I have had the 620 for about 2 months now. I live in Philadelphia in the downtown with many large buildings. The accuracy is quite poor, even with the most recent update . It obtains a lock fairly quickly but it is not trustworthy. If I run away from the buildings, along the water it gets much better.
As an example, I ran one day and it mapped one street over (about 0.08 miles) a good bit into my run. When I reached my turn around point, (it was a short run, don’t judge) at 2.2 miles ,it was off by 0.2 miles.
The problem is huge in marathons. Somebody posted in DC Rainmaker that in the Tokyo marathon their watch was way off. Pretty much any city race, this thing is of questionable value.
My old 205 went haywire in the Hartford Marathon due to the buildings, so I don’t think it’s a problem isolated to this watch. Unfortunately in NH I have little ability to test watches in city conditions unless I drive to Boston :)
I ran the Hartford half lady year , was off 0.2. ……total 13.3. Which I think is pretty accurate …. I used the 910
Most of the time it’s OK.
But sometimes : link to pasaprespas.files.wordpress.com
Very self explanatory… :)
It’s very fast to catch the signal. But I wouldn’t say that it’s more accurate than the 610.
Wow, that’s impressive!
The 610 had the same kind of pb from time to time. And it’s on a stadium without any building around ! Quite surprising. I guess the size of the watch is not helping. But it’s quite rare anyway.
I’ve used my 620 nearly every day since my pre-order showed up, and never had a problem with GPS. I never had a problem with my 610 either. The 620 is definitely superior to the 610 in heart rate monitoring though! Or maybe it’s just the strap.
Ran with a friend at work using Map My Run on his IPhone and we were completely .25 miles out of sync by the end with my Garmin being too long. I’ve had really inconsistent readings with mine.
the run was 8 miles…
Don’t iphone apps use cell triangulation to smooth paths out, whereas GPS watches only have the satellites? I think that’s one reason you will see “smoother” results.
I have the 620. I ran 33 miles Saturday on trails where I’d expect to have had a dropout but not one occurred for the entire race. Don’t know if the update has helped or not, but haven’t had a dropout since the update. Before that, it occurred once or twice.
From ismoothrun website:
We usually give better results than other apps, when there is no GPS signal and other apps give positions from WiFi spots or cell towers, while we switch to pedometer till we get a satellite fix again.
this is also Suunto’s method, they call it Fusespeed and supplement with the wrist accelerometer on the Ambit2 models.
Pedantic/OCD engineer speaking here – GPS watches don’t ‘talk’ to satellites, they are ‘listen-only’ devices, essentially a fancy radio. Whew! I feel better now.
As ED (don’t mean to yell) points out, a smartphone comes with several different types of sensors, including GPS radio, 3-D accelerometers, cell radio, and wifi radio. Given all of these tools, plus significantly more powerful computing power, the theoretical accuracy of a phone ought to be better than any GPS watch available to the consumer.
Rant accepted :)
I’ve been using a 405 since 2008, and as of about a year ago I’ve started noticing the same thing, where my path (specifically at the start of the run) will be offset from reality by as much as 1/4 mile, eventually over time it does settle in and fix itself but I often have garmin activities with questionable data from the first mile (even if I sit and watch the GPS accuracy screen to make sure I have a good fix before starting the activity)
I’ve had a Garmin 220 since released and see mixed results. One day street running in Bangkok I started getting pace warnings as it had me going in a straight line through town at 1 minute km speed.
Other times in other Asian locations I get a solid GPS then it loses the plot for 5 or 10 minutes showing me crossing rivers and going through apartment buildings.
Additionally, when pushing the start button to begin a run, I always get the auto pause and then slow followed by fast pace warnings for the first km.
It’s a great piece of kit but definitely needs some fine tuning. Btw, the update recently had little effect on performance.
Cheers.
Just got my 620 and have only had two runs with it (moved up from the 410). First run was a route that I have run MANY times so I know every half mile spot by heart. Was surprised when I got to 1.5miles and the watch said 1.7. Got home and uploaded results and saw a block detour down a side street around 1.2m that I did not take. Track basically looked like I ran down the street, turned around and ran back. I did download the update after the run and did not notice a problem on my second run. But I do have concerns.
One thing that occurred to me the other day after originally coming across Fellrnr’s detailed analysis of Garmin GPS unit accuracy is that DCRainmaker doesn’t typically do TRUE accuracy tests.
I am a huge fan of DCRainmaker’s reviews, but for just about every review of his that I can remember, his method is to compare the measured distance of the GPS watch he is reviewing to other models that he has on hand. This is not a true test of accuracy, just a test of repeatability (or precision as Fellrnr defines it in his review). I’m not sure why this never occurred to me before, but comparing the measurement of one device to another is not a valid means of accuracy unless the device being compared to is a perfectly calibrated known standard. In this case, the true test of accuracy would be to use a measure course, like Fellrnr did.
By the way, nice writeup, Pete. This was a great read!
Good point, Brad. DC Rainmaker seems like he’s detail-obsessive enough that he might just start doing that – you should suggest it!
That said, he certainly spends a bit of time looking over the maps of his routes, and I would think he’d mention substantial deviations from roads such as what Pete’s experienced. Perhaps he’s just been lucky with his 620, as I haven’t seen him mention any accuracy/mapping problems for it.
I am sorry, but that statement is not entirely correct; in fact, I remember at least one recent occasion when DC Rainmaker did refer to the “wobbles” of the GPS unit under review, the Adidas Smart Run GPS. Please check the GPS accuracy section of that review over at link to dcrainmaker.com.
Hi Pedro, I think we’re actually in agreement here – I trust that DC Rainmaker would mention any substantial accuracy problems with his GPS units. As best as I remember, he didn’t have any problems with that in his 620 review (indeed, he uses the 620 as a comparison in the Adidas review you linked to, and states that the routes for the 620 are “astoundingly perfect”). But if I misinterpreted what you’re saying, let me know.
The only thing I’m not sure about is why DC RM is getting such great accuracy from his 620, whereas Pete & FellRnr (and many others, apparently) aren’t getting such pinpoint readings. Maybe there’s enough variation in manufacturing to affect this? I’m just confused as to why there’s such apparent variation between different units of the same model.
Hard to know. Maybe different batches of units, different measuring conditions, different firmware installed, etc. I’d say min is doing far better than Fellrnr’s unit for example under the conditions I use it (mostly roads). I’ve only recorded about 20 runs with mine so far, so two with big anomalies, one with a small error is only about 10-15% of runs. Two runs since latest firmware have been fine, but will be looking at every run for awhile going forward prior to a review.
The problem I have with Fellrnr’s results is they are not related to what I have seen. In my numerous tests, the TomTom is the most accurate, with the 310xt consistently measuring long. Footpods, even calibrated, were inaccurate outside the calibration pace.
Peter, have you messed around with the data recording setting on your 620? I think the default is the “Smart” data sample setting. I’ve recently switched over to the “Every Second” data sample setting. It’s still too early to tell, but I *think* it’s tracking more accurately (likely at the expense of battery life and file size).
That’s been my experience with my trusty 305 – the “Every Second” recording is at least a little more accurate (haven’t had the motivation to really test it), plus I just like the greater (perceived?) accuracy for interval work.
One other thing I haven’t seen anyone mention here is cloud cover – perhaps Pete’s wonky tracks were taken under thick cloud cover or really foggy conditions? I know that thick clouds/fog definitely slow down my GPS lock time, and I assume accuracy as well.
Not yet, need to play with this as well.
Ever since I bought my 620 in December last year that I have been using the 1-sencond mode and I have had flawless readings.
It is also true that, in general, the weather has been nice, with clear to sparsely clouded skies on the majority of my runs – you see, I live in the south hemisphere, where is nice and warm…
Jealous of your weather :) Clouds, cold, snow and wicked wind for most of the winter for me.
Pretty soon it will be the other way around… :-)
This tiny precious blue bubble of ours is already in countdown tilting to directing the north pole and the north hemisphere towards the Sun. Good news for you!
It’s still early, but 2.60.0.0 has shown a nice improvement in accuracy over 2.50.0.0.
I agree. I’ve now done several runs with the 2.60 firmware update. I have yet to see any of those strange meanderings I observed in earlier versions.
THis guy noticed the same thing, check it out:
link to fellrnr.com
Whoops, should have read the entire article first.
After a heart attack scare at such a young age, mid 20s, i decided to take my health into my own hands. I went for and love the top end garmin 620 gps watch definitely worth the $450, is aesthetically pleasing than others on the market, and improves/measures my running and I have lost 40lbs. I enjoy waking up every morning for a run and I notice I sleep much better at night when I do. Incorporating diet as well with weight training I am a whole new person and I feel great. Couldn’t do it w/o the garmin though. Thanks!
I am thinking of buying the 620 for my wife , is there a consensus that the software update has helped , I run with the 910 , but my wife likes the 620 due to its appearace
I am thinking of buying the 620 for my wife , is there a consensus that the software update has helped
Since writing this I’ve found it to be fine on roads, but on trails I have less confidence. If I run another trail race I may opt to use my old 205. Besides the occasionally accuracy issue and Wifi connectivity problems, I like most everything about the 620.
Updated GPS to 3.10 and Firmware to 2.70; however, issues with the GPS still bothers me. :(
I find that mine is great about 95% or more of the time. But there are rare occasions when I have issues. Seems to be dense tree cover that causes a problem.
Thanks for the post. I just noticed a major issue with accuracy for my Forerunner 620. I ran a marathon recently and the Forerunner was short by over 500m – that means it recorded 42.76km over what was an IAAF certified 42.195km course. SO when the Foreunner told me I had finished (in under 4 hours) I still had over 500m to run! The extra half kilometre blew my time out by 2 and a half mins.
I realised this along the course as the Forerunner was beeping well before the km signs, and this gap got larger as the race went on. I’m a bit disappointed!
One thing to keep in mind is that a GPS watch will almost always measure a marathon long unless you run the tangents perfectly. For example, I wore both the FR10 and FR610 in a marathon last spring and both were more than 500mm long. Very typical for a marathon race so I wouldn’t necessarily blame the watch.
My own testing puts Forerunner 620 on average with most GPS units on the market today. You’re welcome to check out my data in this article. It’s in Danish, but the schematics should be quite easy to digest.
link to wp.me
Thanks, google translate did a good enough job that I could read it :) Since writing this post I have not had many issues with the 620, so my fears have mostly been alleviated. Need to get a full review together soon!
I’m working on that as well, Garmin bringing out the latest update with bike support, bike sensor support, second datafield on lap banner and auto footpod calibration added some more work to the list. ;-)
link to pushperformance.dk
I am the disappointed owner of three Garmin Forerunner 620s. I can’t wait for Garmin to reimburse me for this flawed gadget.
link to valerietoth.wordpress.com
The latest issue with Garmin’s firmware 2.90 is crashes and reboots during an activity:
link to forums.garmin.com
Had this happen to me after the update. Only once though and after reboot no issues since.
Pete – a little update/info from some recent experience I’ve had with the FR620 and Edge 810 during cycling training. I’ve been getting ready for a century ride, and have had my FR620 in Cycling mode (love this new option) while using my Edge 810 for navigation on my longer training rides. I primarily use the Edge when cycling, but it’s nice to have additional stats on the FR620 (total distance, total time, time of day) so I don’t over-clutter my Edge screen and can focus on the map (I keep that down to speed and cadence + nav).
Anyway…I have noticed that, by the end of a 50-60 mile ride, the FR is 1-2 miles short of the Edge’s distance. On my last ride, I checked a few times during the ride, and it’s pretty noticeable already after 10-15 miles, and just builds. My sense is that the FR is just not powerful enough to keep up with the speed of cycling where I am pretty consistently doing ~3′ pace, and clocking >20 mph much of the time (avg speed ends up in the high 18s as I am doing lots of climbing training since my century is in the mountains outside San Francisco).
So, I wonder if this is all related. I can’t imagine the GPS, processor and battery are as powerful in the FR as in the Edge given size constraints. That can lead to a weaker reception and processing speed to deal with things like high speed and obstructions like trees. I see this when I run to work, which finishes in downtown Boston. My pace suddenly drops into the 5′-5’30” range when I get amongst the tall buildings, even though I’m really doing 7’30”-8′. The odd thing is that the GPS track in GC when I sync and overall distance and time are both accurate.
Just thought I’d share this little bit of testing. I do hope some firmware updates can help, but I wonder if there’s just some physical limitations to the design. That would really suck given the cost of the thing (and that it’s generally a great watch otherwise).
Thanks for the info Bryan. I really feel like the 620 is very sensitive to tree obstruction. I still have times when the route drops out for a bit and has me running through the woods, houses, or cutting corners I did not cut. 95% of the time it seems to be fine, but there are times when it seems to have these weird little hiccups. Still not sure why.
Hi, I have my GFR6200 on 3.0 firmware version (both GPS and sw) but…. yesterday I got the same error. The run is in open roads with very little house/tree obstruction. bye, jose.
Correction: GPS version is 3.10
Same issue here. With the current firmware version and in September 2014 ;(
It may be trees in my case. But the Strava app on my Android phones works fine.
Returned my 620 to Amazon.
It seems that Garmin acknowledges an issue with FR620 GPS accuracy in the release of FR920XT. Interested readers may want to follow this thread: link to forums.garmin.com
My 620 is complete piece of junk. This watch is completely unusable for serious runners using it for pacing. I never regret something that much as me replacing my 610 with the 620. Everybody reading this: NEVER EVER buy this device.
This issue I could confirm on a second 620 as well. The first had the same issues (it was replaced because it unrepairably killed itself during the fw update…horraaaay also :D).
I was a former polar user and switched to garmin for open standards (ant+)…but I’m gonna be a polar user again.
Hello, I was wondering if there is a way to edit or modify the tracks in the event the GPS has not been accurate. For not being accurate the whole workout is sometimes messed up (e.g. pieces where I run at a speed of 30 km/h or s0).
THanks
Same GPS problem here different 500 m form running road. I was disappointed with Garmin FR 620.
I entered a 12km race in Feb this year and was surprised to cross the finish line when my 620 registered 10.96 km. My next race, a half marathon in Apr, was registered as 19.54km.
Finally my 3x1600m track repeats on Monday were registered as distances of 1.32km, 1.28km and 1.32km!
GPS watches are not suitable for tracks. Use a traditional stopwatch. This is not a garmin issue. Every GPS watch I have owned has been ill suited for track work.
I have used a Garmin 610 in Seoul, Korea and other places around the world for a number of years with generally repeatable and stable results. Recently, I obtained a 620 and started running some of the same routes in Seoul and have found the 620 basically unusable with very significant deviations from the actual route (on the map) and consistently wrong splits and totals for distance. There is no doubt in my mind that it is significantly worse than the 610. I have the latest software so not sure what else might be coming. I am trying to contact Garmin service but so far had to get back to work before I made it through the waiting queue.
I have a Garmin 410 with heart rate monitor. It used to be accurate (I believe) and consistent – if I went for the same 10k run the measured distance would vary by a few yards/metres, and my heart rate would rise quickly in a smooth curve then level out and slowly rise. But then Garmin installed a software update back in around April 2013 and after that the 10k would vary by up to 5% (0.5k) and my heart rate would shoot up to 250 bpm, plummet to 50-60 and then throughout the run every now and then shoot back up to 250 or so before dropping back. Now when I go for a run it usually measures short compared to everyone else’s watches (Garmin, TomTom, Fitbit and the rest), the worst being round Grizedale Forest when my watch said 15 miles, my friend’s 16.5 miles. Garmin’s response when I contacted them was to say it’s out of guarantee, buy a new one. I’d had it for 14 months and it worked perfectly until their software upgrade. I keep checking for new updates but nothing, still mesures short and apparently I can walk on water, as the map often shows.
It would be great if they released a correction to their last update so my watch would be accurate again, but I now just use it as a guide to how I’m doing. It’s annoying but I’m not serious (or good) enough to get too upset.