This site uses cookies to offer you a better browsing experience. Find out more about how to manage cookies, or dismiss this message and continue to use cookies.

"Quality" of GPS hardware - smartphone vs wrist watch

I've been a fairly long time user of a GPSr (old yellow eTrex was my first one bought about 1 month before SA was disabled) and have used them for Geocaching, navigation and more recently running.

I keep finding myself getting into debates with other runners about using smartphones vs wrist watches to track running. I've used both, using a Nokia symbian phone in 2010 with Sportstracker through to today using a Garmin vivoactive watch.

The tone of the debates often starts with someone complaining that their "Strava" app on their smartphone has created a dodgy trace, or measured a race they've done as longer / shorter than it should be. Many posters then come back with "should have used a Garmin watch" (or similar) on the basis that "its well known that Garmin watches have better GPS than smartphones".

My assertion is that, whilst my experience is that both smartphones and watches give acceptable results for runners (with watches often seeming to output less "noisy" routes) that what they are seeing isn't a hardware issue (in reference to the GPS chips etc), but most likely a usage issue in that they aren't able to give their smartphone a good view of the sky (well, the antenna at least) as they'll wear it in an armband, or on a belt etc which will inherently make it harder for the phone to maintain an accurate lock. And secondly what they may be seeing is actually the software on the watch being dedicated to what they are doing - i.e. running - and then Garmin making cute programming options to try to smooth out the trace, knowing that most runners don't actually zig-zag left to right every second or so.

So my question is - is there any hardware reason why a watch should be better than a smartphone? Are there "rubbish" GPS chipsets out there in some phones? Do Garmin running watches have "better" GPS hardware within them? Or is it all down to software / firmware which means that smartphone GPS usage is inherently compromised a little to manage power drain etc whereas running watches are optimised to just do what they are supposed to do?




  • Tim 1481 Points
    A few points come to mind. First is battery. First, let's separate devices into categories of devices that are primarily designed for recreation/fitness (like a Garmin fitness watch, Garmin Edge GPS for bikers, eTrex, etc) and those devices which are designed for full time carry like smartphones and watches like the Apple Watch.

    The recreation/fitness devices have different battery goals for the end user as the user will often use them just for the fitness activity, as opposed to a smartphone where the user demands not only it to last during the activity, but "all day" battery life.

    Although there are tons of variables (size of battery, driving large color screens, etc) I think it is fair to boil it all down and say that for fitness devices, the device will send full power to the GPS chip, 100% of the time. On the other hand, a smartphone will attempt to conserve battery life whenever possible and will reduce power to the GPS chip taking sub second cat naps whenever it thinks it can to preserve battery life.

    Second, a smartphone will augment location determination through other sources such as proximity/triangulation of cell phone towers, detecting nearby wifi networks that have been geolocated, etc. Those additional data sources can also introduce additional errors.

    Third, I agree that the position (sky view) is also important. I'll often carry my Garmin Edge while hiking. I might put it in a top jacket pocket, top pocket of my backpack, etc. In those positions it never gets the same track quality as it does when mounted on the bars of my bike. The same thing would happen with a phone. Your body blocks the GPS signal quite well so if it is against your body then you're blocking a good portion of its view, as opposed to being on your wrist, handlebars, etc.

    Finally, the GPS chipset phone makers are putting into their phones are typically a little different than those going into dedicated GPS devices for reasons citied above. (A smartphone will opt for a GPS with lower power requirements, and will augment the location with data from other sources.)

    To my knowledge, based on talks with Garmin engineers a few years ago, they are not doing any "smoothing" of their own to the tracks. They are recording the data they get from the GPS chipset... which itself could be doing some smoothing, but perhaps more accurately described as "error correction".
  • Boyd 1997 Points
    edited April 2017
    Another approach might be to use your smartphone app with position data from an external bluetooth GPS receiver. They are small and can be clipped to an optimal position on your body. I have been impressed with the Garmin GLO, but there are other options, some of which are even more accurate (and expensive). The GLO records 10 position fixes per second, Garmin's dedicated GPS units only record one position per second (not sure about their watches). Bluetooth receivers can be used on both iOS and Android with any app.

    See this:
  • privet01 227 Points
    edited April 2017
    You sort of ask this in regards to fitness so my response is more geared that way. And is more just for me to get this "off my chest" as I've had plenty of the same discussions you seem to be getting into.

    For sport and fitness devices. The only thing that matters is time. So if you did a particular activity as well or better than the previous then you are good. Doesn't matter how bad the log of your track looks, or what your HR, cadence, Calorie burn, power, elevation gain/loss or anything else looks. Out of all those things, track is indeed the least useful piece of data for your activity metrics.

    Similarly, what's going to benefit a 200 lb. (90 kg) cyclist more for speed? Carbon fiber aerodynamic rims or loosing 20 lb. (9 kg) of bodyweight ?

    Almost every time, they'll spend the money on the rims..................
  • larkim 6 Points
    @privet01 - I agree to a point. Track does matter *if* you're looking to compare your "training" pace to a potential "race" pace. If I train for 7mins per mile as a runner using a device which (unknown to me) is measuring 1 mile as 1.1 miles due to "wobble", and then I turn up to race a half marathon and try to get past each mile marker in 7 minutes I'll find I'm way behind pace as the mile marker will come up way before when my phone "pings" at me, and I'll have an unrealistic sense of how fast I am going. In one sense of course that doesn't matter, run to perceived effort etc etc. But that sort of experience could ruin at least one race day for which lots of effort has been put in.

    Ref spending money on losing 50g on a bike component vs losing weight around your stomach - agreed!

    @tim - helpful, thanks. One point I'm never entirely sure of on a technical perspective is how much influence those non-GPS data points have when a smartphone is updating location. I'd always assumed that cell tower triangulation and wifi geolocation were just used as helpful initial indicators to make initial fix easier. Is that a wrong assumption on my part, do these data points actually get utilised throughout the activity? If so, one straightforward piece of advice might then be to switch off wifi and data connection before heading out for a run, these points would then deliberately get ignored. Also, helpful insights into the power management impacts. Though it seems to me if an app like Strava is initialised I'd again assume that this puts the GPS chip into "full power" mode on a par with the power drain on a watch, in the same way that my watch doesn't use GPS all the time, but only when I ask it to start a run / ride etc.

    Finally @boyd - interestingly this is precisely what I used to do once upon a time with a bluetooth GPS dongle that I used with a device that didn't have GPS built into it, and (whilst it wasn't up to the Garmin GLO standards) it was a very fine device in terms of my perceptions of its accuracy.
  • privet01 227 Points
    Understand that in the case of devices like Garmins, the values you see on the device for distance traveled, elevation gain/loss, average pace, average speed and many others pieces of data might show entirely different when you look at the track log. I don't use Strava very much anymore, but when I did, they didn't use the accumulated values of the device that was passed to them in the activity log. They took the position points of the track and built their own "model" of what you did. But that is based on track points that are usually 1 second or better apart. So at 20 mph (32 kph) you are covering 30 feet per second or 9 metres per second. A lot of the curves and corners get left out of your distance as well as the bottoms and tops of your hills for the elevation gain/loss.

    And then we also have to argue whether gps truly only does horizontal distance, or does it know your path is longer going up an incline and accumulate that slant distance? I don't know. But I trust the distance derived from the speed sensor of my bike more than the gps distance.

    Still, if I do the same route day after day, time is still the only thing that matters. You don't win or lose the competition or body weight or heath by how accurate the track is. You are trying to glean info from it that will always be dodgy at best.... IMO of course.
  • Tim 1481 Points
    I'd agree that augmented data from cell towers and wifi geolocation will predominantly help with initial fix, as you suggested. I'd also say that they typically do more harm than good. But they can be an additional source of errors as well. It probably does more harm than good, so I'd leave them on for fitness activities, but I've seen indications they can also be an additional source of errors.

    I'd further agree that GPS based apps on smartphones will kick the GPS chipset into full gear. However there are APIs available to developers that can change just how much polling they want the GPS to do. Developers will try to balance "battery shaming" and potential bad reviews regarding battery life while using their app versus the need for accurate data. I'd agree something like Strava would probably be polling the GPS about as much as the APIs will allow them to.
Sign In or Register to comment.
↑ Top