[ringmaster]

Does anyone know how accurate the speeds/distance data is on a Garmin when relying on a GPS signal for speed and distance rather than using Garmin's bike-mounted speed sensor? I would love to keep my frame and wheels clean and naked but don't want to miss out on too many Strava KOMs from poor accuracy

 

[deviousalex]

Does anyone know how accurate the speeds/distance data is on a Garmin when relying on a GPS signal for speed and distance rather than using Garmin's bike-mounted speed sensor? I would love to keep my frame and wheels clean and naked but don't want to miss out on too many Strava KOMs from poor accuracy

It's not as accurate, especially in places with tree coverage. 510 is better because it has Glonass but still not perfect.

On the other hand you won't be missing out on any Strava KOMs due to lack of accuracy because Strava KOMs are decided by GPS points not the speed data recorded via those sensors.

If you want to keep your bike "clean" get the newer ones that accelerometers instead of magnets.

 

[apn]

DCR looked into this a few years back (Edge 800), but I've not seen anything more recent;

2011 Sport Device GPS Accuracy In Depth: Part I | DC Rainmaker

2011 Sport Device GPS Accuracy In Depth: Part II | DC Rainmaker

GPS is good enough for me (I don't Strava) and I use the mag-less cadence sensor.

As deviousalex said, tree cover can be an issue (more so in the rain), as would the virtual canyons in the downtown core of major cities, as well as tunnels. Fortunately, most of us avoid cycling in the latter two.

 

[Migen21]

I've never really understood the need for a wheel speed sensor. If I understand how the data works, they would only be useful to tell you your *current* speed at any given point in time (which rarely ever interests me). Any averages (over time, distance, segments, KOM's etc...) would need GPS/location data.

I removed all of my speed sensors a couple of years ago and haven't missed them one bit.

Edit: My comments are framed in the context of normal outdoor riding. I could see a use for a speed sensor on your wheel for indoor riding (track, trainer, etc...), and possibly as a backup in areas where GPS may be unavailable or inaccurate (heavy tree cover, some city areas).

 

[MMsRepBike]

Current speed is not displayed on my screen. Average speed isn't either. No need for a sensor to tell me something I don't even want to know.

 

[apn]

Just noticed that the new Omata (Analog) GPS cycling computer quotes the following specs;

Speed Accuracy: 0.18Kph/0.11Mph
Distance Accuracy: 2.5m/8ft

Since the technology is established, most manufacturers will likely be using off-the-shelf receiver chipsets (rather than investing $$$ in IP) to handle the same GPS+GLONASS signals. Thus you can reasonably expect that other manufacturers products should be in the same ball park.

 

[deviousalex]

I've never really understood the need for a wheel speed sensor.

I have a Garmin Virb Elite that I create the occasional video with. Garmin is a bunch of lazy programmers and they don't have the ability to use an ANT+ speed sensor so I have to rely on GPS if I want that data overlaid on the video. It can be wildly inaccurate in areas with poor GPS accuracy and shows me going accelerating from 15 to 50mph in about 3 seconds while going down a descent where I'm really doing 25-35.

Just noticed that the new Omata (Analog) GPS cycling computer quotes the following specs;

Speed Accuracy: 0.18Kph/0.11Mph
Distance Accuracy: 2.5m/8ft

Since the technology is established, most manufacturers will likely be using off-the-shelf receiver chipsets (rather than investing $$$ in IP) to handle the same GPS+GLONASS signals. Thus you can reasonably expect that other manufacturers products should be in the same ball park.

There's more to it than just the chip, it's how the developers use the data from it. Some really smart devices can take noise out of the data by doing post processing in the software. Remember, GPS satellites aren't in geosync orbit and satellites are constantly moving in and out view of your GPS.

 

[bigbill]

I used the wheel sensor with my 305 when I lived in Sardinia because the roads weren't accurately mapped. Same thing when I was in Montenegro, Croatia, and Slovenia. My routes would be lines out in the Mediterranean on some rides. I replaced it with an 800 four years ago and it has been much better staying synched. I've never noticed any inaccuracies and if there are any, it's been in my favor. I've got a bunch of KOMs around here.

 

[scott967]

I have a Garmin Virb Elite that I create the occasional video with. Garmin is a bunch of lazy programmers and they don't have the ability to use an ANT+ speed sensor so I have to rely on GPS if I want that data overlaid on the video. It can be wildly inaccurate in areas with poor GPS accuracy and shows me going accelerating from 15 to 50mph in about 3 seconds while going down a descent where I'm really doing 25-35.



There's more to it than just the chip, it's how the developers use the data from it. Some really smart devices can take noise out of the data by doing post processing in the software. Remember, GPS satellites aren't in geosync orbit and satellites are constantly moving in and out view of your GPS.

https://en.wikipedia.org/wiki/Kalman_filter

scott s.
.

 

[apn]

There's more to it than just the chip, it's how the developers use the data from it. Some really smart devices can take noise out of the data by doing post processing in the software. Remember, GPS satellites aren't in geosync orbit and satellites are constantly moving in and out view of your GPS.

Agreed. It's not my area of specialty, but looking at some of the OTS product, Skytraq has a pretty comprehensive SOC solution, and likely includes (licenses) reference firmware for the onboard CPU.

http://www.skytraq.com.tw/datasheet/S1216V8_v0.9.pdf

The larger IC at the top of the Omata main board is likely one of these devices, or a functional equivalent.

From a HW perspective, all you need is a basic feature/control CPU, an ANT+/BLE chip, a battery charge circuit and something to drive the USB-C interface and Voila. Thus, I don't think there's any groundbreaking IP to justify the premium price.

I think we also agree on the legendarily poor Garmin software

 

[thisisthebeave]

The data recording setting has a big impact too. When I got a power meter I switched from "smart recording" to "1sec recording" and all my ride distances went down, especially ones with tree cover. There's a particular mountain bike loop I do that used to get me 24.3 miles, now it records at 23.1. There's a decent amount of tree cover for about half of it.

 

[wilde737]

I have a 510 that suddenly is horribly inaccurate, both speed and distance. I'll do one leg of a ride and my garmin will say i've gone 11.5 miles and starva says 17.7. I was wondering if maybe the battery on my sensor is getting low. I'll try that next.

 

[TmB123]

I find that if you don't use a wheel sensor then the current speed generally bounces around all over the place, even on my 510 with GLONASS turned on. Once you stick a sensor on it, rock solid readout. Regardless of what your Garmin says, Strava makes its own mind up as to how far you have gone based on GPS points. The 810 doesn't have GLONASS so will never be as accurate as a 510/520/1000 etc

 

[rsilvers]

It's not as accurate, especially in places with tree coverage. 510 is better because it has Glonass but still not perfect.

On the other hand you won't be missing out on any Strava KOMs due to lack of accuracy because Strava KOMs are decided by GPS points not the speed data recorded via those sensors.

That is true for Strava segments. But just for everyone's info, the total distance that Strava lists for your ride (which I care about a lot), as well as your average speed (which I also care about a lot), *is* based on your wheel sensor data, if you have a wheel sensor on a Garmin device:

https://support.strava.com/hc/en-us/articles/216919487-How-Distance-is-Calculated

Strava only over-rides this and uses GPS distance if their algorithm sees some out-layers in the GPS data.

 

[TmB123]

I'm not sure if it's still the same, but distance ridden within your privacy zone at the start or end of ride "used" to be deducted from your total mileage.