Electronics, Lighting, & Gadgets - How accurate are those Garmin GPS units?

I'm thinking about getting a 205 or 305 to replace my cyclocomputer. But I heard GPS is accurate to within 10-35 feet...it's not pinpoint accurate. So would this throw off the actual speed displayed and/or distance? I know my cylocomputer is very accurate and I don't want to lose out by switching to a GPS unit. I don't plan on running the wheel sensor

I'm thinking about getting a 205 or 305 to replace my cyclocomputer. But I heard GPS is accurate to within 10-35 feet...it's not pinpoint accurate. So would this throw off the actual speed displayed and/or distance? I know my cylocomputer is very accurate and I don't want to lose out by switching to a GPS unit. I don't plan on running the wheel sensor


pretty accurate....for example....i rode the 3 state 3 mountain Century back in May. my Garmin showed 99.28 miles for the entire 100 mile ride...so it missed 0.72 miles over 100 mies....as far as climbing acuracy goes...it is relatively accurate there too. 10-35 feet error isn't that bad.

My 305 speed reading is dead on with my flightdeck. Elevation on the unit is usually +/-20-30 feet, but that hasn't affected distance or speed readings.... yet.

The units are very accurate. Commercial GPS is now accurate to better than 10m 95% of the time. You can check the Edge by going into the Navigation menu and looking up the satellites that are being tracked.

The Garmin computers display a lot of different numbers; some are more accurate than others. Distance is the most accurate number. Gradient and calories numbers are very suspicious. I think the speed number tends to lag a little when you are accelerating, but it is OK once you are cruising.

I'm not sure I'd attempt an instrument landing by one but so far mine has been painfully accurate. . .except for the occasional HR-strap-spazfest.

Yes, that's correct. So at the start of your 50-mile ride your position might be off by 25 feet, and at the end of your 50-mile ride your position might also be off by 25 feet. Those errors could both be in the same direction, in which case the total distance error is zero, or they could be in opposite directions, in which case the total distance error is 50 feet. That's 0.019% off; I very much doubt you can measure the circumference of your tires that accurately!

Yes, I'm oversimplifying, but the intermediate errors don't jump around randomly (which would maximize their effect). If your GPS is telling you're 32 feet southwest of your true position, 30 seconds later you're probably also reading about 32 feet southwest of your true position, in which case there is no error in the distance.

The small positional errors at intermediate points tend to largely cancel out - that's why you may see slight speed fluctuations when you're pedaling on the flats with a rock-steady cadence. But unless your rides tend to be less than 200 feet, the GPS is probably more accurate than measurements with other devices that require an accurate measurement of your tire circumference. An 0.019% error in the size of a 700x23 tire is about one-third of a millimeter, and your tires probably vary that much based on their pressure that ride, their temperature, and how full your water bottle is.

Speed is accurate. Elevation is usually quite accurate. But cumulative elevation gain as reported on their website (MotionBased) is usually substantially higher than actual. My training rides on a flat course show up as 1000 ft up and 1000 ft down. I think they count every speed bump or minor road variation.

I've seen some threads lately where people report that when "converting" the data on the Garmin with add on programs (SportTracks and Motion Based), the numbers change. Doesn't the Garmin come with software so data can be uploaded to computer? If so, what is the need for the add on programs? I'm strongly considering getting a 305 with HR & cadence, but these threads have given me some pause. If the software is necessary, which do people recommend (& why)?

Dan

Yes, I'm oversimplifying, but the intermediate errors don't jump around randomly (which would maximize their effect). If your GPS is telling you're 32 feet southwest of your true position, 30 seconds later you're probably also reading about 32 feet southwest of your true position, in which case there is no error in the distance.


There's absolutely no possible assurance that this is the case. Positional accuracy is dependent on the angle at which you are receiving satellites to the horizon and number of satellites. If you move @ speed for 30 seconds or change directions, you're going to have a change in those factors.

I'm not saying the GPS is not accurate (I have two and use them often), but this statement is pretty off the mark.

Speed is accurate. Elevation is usually quite accurate. But cumulative elevation gain as reported on their website (MotionBased) is usually substantially higher than actual. My training rides on a flat course show up as 1000 ft up and 1000 ft down. I think they count every speed bump or minor road variation.

Is the cumulative elevation gain on MotionBased substantially different from the elevation gain number that is displayed on the Edge 305 itself? Is MotionBased using a different method for computing elevation gain?

I've seen some threads lately where people report that when "converting" the data on the Garmin with add on programs (SportTracks and Motion Based), the numbers change. Doesn't the Garmin come with software so data can be uploaded to computer? If so, what is the need for the add on programs? I'm strongly considering getting a 305 with HR & cadence, but these threads have given me some pause. If the software is necessary, which do people recommend (& why)?

Dan

The PC software is not required, but it is very useful. For one, it archives your data by date so you can go back and look at your past rides at any time. It also shows you all kinds of maps and graphs and profiles to help you analyze the ride. For example, you can display your heart rate and your elevation and your speed mile-by-mile so you discover where you were slacking. Or you could use the elevation profile to plan your attacks the next time you do the same ride. Also if you're just exploring or a friend is showing you a new ride, the map display will show you exactly where you were so you can easily repeat the ride later.

Positional accuracy is dependent on a LOT of other factors (atmospheric conditions, multipath reflections, etc.). Having spent quite a few years doing field GPS work with high-end (not consumer) units that record pseudorange data for each satellite (for postprocessing), I can report that moving for 30 seconds at bicycle speeds will - most of the time - cause very little change in these factors. Most of the time you will still be receiving the same satellites with the same atmospheric conditions, etc. You'll tend to jump to another set of satellites for a new solution for a bit, then jump to a new set, etc. The sateliites you're using are over 12,000 miles away, and moving a thousand feet or so makes no material change in satellite geometry. After all, the satellites themselves are moving about 350 times faster than you are!

This is especially true in consumer-grade units. Recreational users don't like to see their position jump around and don't understand what it means, so most recreational units employ various types of smoothing or velocity filtering to reduce small fluctuations in the positional solution from moment to moment.

My comments were made based on actual data from my personal experience over eight years of GPS use (from back in the old SA days) with over a dozen consumer, mapping, and survey grade GPS units.

I have a 301 which I use all the time. Distance and speed seem very accurate. Calories and elevation seem very inaccurate. When I connect it to the PC and look at the track on a map it follows the road correctly. If I start and stop at the same spot it shows that correctly. But when I do a loop it usually shows a net elevation gain or loss (which is impossible). If you are setting up your cyclocomputer by measuring one wheel revolution, then GPS is much more accurate.

Any GPS unit will find it harder to calculate your elevation than your horizontal position. This is mainly due to simple geometry.

Your GPS is essentially triangulating your position by using the calculated distance from four satellites at known locations. If all four satellites are in the same spot in the sky, you'll have a hard time triangulating. The ideal satellite geometry for calculating your horizontal position is to have four equally-spaced satellites all spread out in the sky right on the horizon. But the atmosphere gets in the way, so the best you can get is probably about 25 or 30 degrees up; close, but not ideal.

The ideal satellite geometry for calculating your vertical position is a similar setup tilted 90 degrees, with one satellite directly overhead, one satellite directly below you, and one each halfway between those two and opposite each other in the sky. The problem is that three of those four satellites are behind the Earth and you can't see them. The best you can get in practice is really quite a bit worse, since there's no way you can use a satellite anywhere near the position beneath your feet.

GPS accuracy is measured in a number of ways, including HDOP and VDOP (horizontal and vertical dilution of precision). In most scenarios VDOP (your vertical or elevation error) is 1.5 to 2 times larger than your HDOP.

However - that's why Garmin puts aneroid barometers in these things, to calculate elevation better! But I also find that I end up with elevation measurements that are pretty bad, and find my house has gone up or down by 100 feet after a short 15-mile ride.

I have a 301 which I use all the time. Distance and speed seem very accurate. Calories and elevation seem very inaccurate. When I connect it to the PC and look at the track on a map it follows the road correctly. If I start and stop at the same spot it shows that correctly. But when I do a loop it usually shows a net elevation gain or loss (which is impossible). If you are setting up your cyclocomputer by measuring one wheel revolution, then GPS is much more accurate.

Does the 301 use a barometric altimeter or the GPS altimeter? A barometric altimeter (like the Edge 305 uses) should be much more accurate for measuring cumulative elevation gain.

This is an interesting thread in that lately I've been testing one of those services that
allows you to use a gps enabled cell phone as a fitness gps.

The advantages of the Garmin units is that they update GPS position much more often
than the particular cell phone I have - mine updates its position every 5 seconds (Razr
v3m) and the newer ones do it ever 2.5 seconds but I believe the Garmin units do it
even more often but a 305 owner should be able to clear that up.

The disadvantages of reading speed every 5 seconds at 20 miles an hour will really
do a number on distance at times. Especially if your course changes rapidly at
high speed. The computer sometimes "cuts off" parts of where you rode if you
encounter that kind of thing.

See:
http://bimactive.com/ba/journal/post/4941/5383

This is a ride I did last weekend.

Look at the left side of the pic about halfway down where it says
"E Bush Lake Road". Just below that you'll see my GPS track looks
like I rode right thru the lake!

That is a downhill so I curved around that area but was at least
a ways away from the lake. GPS said 19.21 miles but my bike
computer reported 20.5 when I was done - so it's not completely
ideal but they corroborate altitude changes with known USGS data
for the area so it should be more correct than GPS data alone.

Would I get a 305 if I had the $? Yes. Will I keep using the service
I'm using with my phone? Also yes since I don't have a 305 and it's
nice to be able to save maps of where I've ridden for google earth
access so my wife can see where I've ridden.

Will I upgrade my phone to a faster reading phone? Yes, eventually.
The KRZR and some other newer phones are 2 to 2.5 seconds reading.

But if the $ came my way, you can bet I'd get a 305.

The 301 uses GPS satellites to triangulate and measure elevation, which is incredibly accurate. I use a SUUNTO vector watch when backpacking which measures elevation by ambient air pressure, and when compared to the my backpacking GPS and the GPS's of other people, it is consistently off by as much as 400ft in one instance, where all the GPS's have elevations that match within a couple feet.

Hullo - my first posting here.

I've used a Garmin FR 301 for riding and general outdoor activities since March 2005 (including racing) and found it to be accurate enough for my purposes. Unless you're in cityscapes, canyons or heavily forested areas, even the 301 gives accuracy down to 5-6 m. The Edge is said to be even more accurate because it uses a more advanced chip.
My preferred analysis software is SportTracks because it just offers so many more features and customisability than the Training Centre software supplied with the Garmin. SportTracks also allows editing and correction of the GPS track, elevation profile and heart rate graph. One other thing: it accepts uploads from a variety of makes and models, including Polar heart rate monitors; those files can be integrated with GPS tracks uploaded at a different time.
As to the disparity in indicated ascent/descent, this could be a result of your chosen smoothing or threshhold setting. For cycling, I have set my anything over a 2 % gradient as a climb (obviously anything lower than -2% qualifies as a descent). Everything between -2 and 2 % is regarded as flat. To an extent this eliminates inconsequential bumps and mild slopes, changes in road camber, etc. Also, because of the way GPS pinpoints your position it's possible that there will be occasional spikes, not to say inaccuracies, particularly when it comes to elevation. As I mentioned above, SportTracks allows these spikes to be shifted or deleted by clicking/dragging, without having to work directly on your history file's source coding.
Oh, and the 301 doesn't have a barometric altimeter.

The 301 uses GPS satellites to triangulate and measure elevation, which is incredibly accurate.

No, actually, that elevation is not incredibly accurate; no GPS will do that. But it will be very consistent among comparable recreational GPS units. Your friends will all have units that agree completely - on an inaccurate number!

Any barometric altimeter needs to be calibrated; if yours is really off by 400 feet, you either didn't calibrate it, it's broken, or you're in a thunderstorm! If you want it to work properly, you do need to calibrate it daily or at the start of each trip. It will give you very good results as long as the weather's not changing quickly. The benefit of the GPS units with barometric altimeters is that they can use the GPS elevation to calibrate the altimeter for you. The altimeter is excellent for relative elevation changes, and combined with the GPS should give a pretty good auto-calibrated result. It's not as good as a well-calibrated barometric altimeter, but it's a lot better than an uncalibrated one.

http://www.youtube.com/watch?v=taRPwYe1EYA

The ideal satellite geometry for calculating your vertical position is a similar setup tilted 90 degrees, with one satellite directly overhead, one satellite directly below you, and one each halfway between those two and opposite each other in the sky. The problem is that three of those four satellites are behind the Earth and you can't see them. The best you can get in practice is really quite a bit worse, since there's no way you can use a satellite anywhere near the position beneath your feet.


Taking a slightly different tack, a 'good' and useful satellite geometry
would be four satellites: 1) overhead, 2) due north (for example) near
the horizon, 3) near the horizon but 120 degrees clockwise from 2,
and 4) near the horizon but 120 degrees clockwise from 3. Put the
satellites closer together and they won't produce as accurate a result.
The trouble is that these 24 satellites are in 'low earth orbit' unlike the
geosynchronous satellites we are used to for e.g. TV and they move
all over the place so the accuracy is constantly changing. I don't know
how many satellites data current GPS units will actually use in their
computations but generally more visible is better.

Good reception from 3 satellites is the minimum required for
latitude/longitude and 4 are needed for altitude.

My Garmin Etrex Vista CX is very accurate at reporting distance. . . and average speed is fine. For a reading on my current speed, it is spotty.

When I first received it, I mounted it in lieu of my wireless computer. I have since remounted the wireless.

On downhills, the Etrex will often simply "freeze" at 14 mph for a while, then, jump to 35 MPH in one second. If, in the interim, you have exceeded 47 mph, then, the Etrex will have totally missed that top speed.

Actual distance travelled, however, is very accurate.

Caruso

I'm thinking about getting a 205 or 305 to replace my cyclocomputer. But I heard GPS is accurate to within 10-35 feet...it's not pinpoint accurate. So would this throw off the actual speed displayed and/or distance? I know my cylocomputer is very accurate and I don't want to lose out by switching to a GPS unit. I don't plan on running the wheel sensor

I've used a Etrex Yellow (the basic model) when I was out in Elmira, Ontario, Canada and got 2 meters accuracy to my grid point. All the other units I've tested (RINO 120, Etrex Legend, Etrex Legend C) inside the city of Toronto I've gotten as low as 5 meters accuracy to the grid point. My thinking is that because I was way out in the rural farmland when in Elmira that the accuracy was better without all the buildings and such to muck with the signal.

My Garmin Etrex Vista CX is very accurate at reporting distance. . . and average speed is fine. For a reading on my current speed, it is spotty.

When I first received it, I mounted it in lieu of my wireless computer. I have since remounted the wireless.

On downhills, the Etrex will often simply "freeze" at 14 mph for a while, then, jump to 35 MPH in one second. If, in the interim, you have exceeded 47 mph, then, the Etrex will have totally missed that top speed.

Actual distance travelled, however, is very accurate.

Caruso

Are you using battery save mode or normal mode? I know when I was testing the GPS's on battery save and normal mode battery save mode updated I think every 3 seconds where as normal updated I think every 1 sec.

http://www.youtube.com/watch?v=taRPwYe1EYA

Someone owes me a new keyboard. Who do I send the KB cleaning bill to? Seriously I could not stop laughing and sprayed the OJ all over when I saw that commercial. Best GPS commercial I've seen to date.

I'm thinking about getting a 205 or 305 to replace my cyclocomputer. But I heard GPS is accurate to within 10-35 feet...it's not pinpoint accurate. So would this throw off the actual speed displayed and/or distance? I know my cylocomputer is very accurate and I don't want to lose out by switching to a GPS unit. I don't plan on running the wheel sensor

I actually did the reverse, I replaced my Garmin 205 GPS with a Cateye Strada cyclocomputer. The reason is because 1) for my commute, I don't really care to map my routes, and 2) I got tired of waiting for the GPS to get a satellite lock before I can ride (my afternoon commute starts in a downtown area so it can take a minute or two to lock on). When I first got the Strada, I rode with both units for a couple of weeks. The distance and speed readings from the two were very close and very consistent. I don't think you'll lose any significant accuracy by switching.

Are you using battery save mode or normal mode? I know when I was testing the GPS's on battery save and normal mode battery save mode updated I think every 3 seconds where as normal updated I think every 1 sec.

I will have to check that . . . I'm not certain my unit even has a battery save mode. It seems to run forever on a set of batteries, though.

As I said, I'll double check . . . but, at times, it seems to respond as contemporaneously as the computer. At other times, not, so, my guess is that what I'm seeing is for reasons other than the battery use setting.

Caruso