The front of my bike is faster than the back!

[Cunobelin]

Inspired by the military GPS thread I decided to post this single sample experiment

Just out of interest I had borrowed my wife's Edge 705 yesterday and attached it to the handlebars of the Catrike, and My Edge 800 to the fairing on the front

I then started both computers and left for 5 minutes to acquire and settle.

Both are on auto record, so theoretically they should record identical trips. The display / analysis  then aligns the two tracks from first activity, again should be the same point

Not the case!

At some points they meet up, but when comparing on Ascent the Front of the bike (Edge 800) is consistently faster than the rear (Edge 705) and they even managed a distance of 125 metres from each other at one point - a very long wheelbase machine.

Interesting as it shows the accuracy of of the track these systems record is actually very much based on the unit itself

[Biggsy]

I've always found it perplexing or amusing when people have a lot of faith in GPS distance measurements, and also claim it's more reliable than a normal cycle computer.  Your experiment proves that a pinch of salt is required.

[Greenbank]

so theoretically they should record identical trips.

That's one assumption too far.

Cycling GPS are far from accurate when it comes to distance travelled or speed. It can do static position very well, the rest is just interpretation that just compounds errors with some (but nowhere near all) of the error being averaged out.

We've done this before lots of times before:-

http://yacf.co.uk/forum/index.php?topic=31482.msg583368#msg583368

[fuaran]

How many trackpoints did each device record? That could make quite a difference to the distance or speed recorded. ie if you have fewer trackpoints, it will be cutting corners more.

If they were set to smart recording, it's likely that the algorithm for that might be different between different models. If you want to compare them, best to set both to record 1 point per second.

[Cunobelin]

I fully realise the limitations, and have made no attempt to go into detail. I realise it may even be as simple or as complex as you require.

All I wanted to do is take the two units as they would be "out of the box" and in "normal use"

[Greenbank]

My point is that if you repeat the experiment a number of times you will probably find the opposite happening several times, i.e. the Edge 705 will record a longer distance than the Edge 800.

[Panoramix]

At some points they meet up, but when comparing on Ascent the Front of the bike (Edge 800) is consistently faster than the rear (Edge 705) and they even managed a distance of 125 metres from each other at one point - a very long wheelbase machine.

You worry me. I used to sail a bit and have been assuming that GPS position was safe within 100 meters since they've upgraded the signal. I thought this was safe by an order of magnitude and i know that some people use lower values. It is important to me because you need 100m accuracy to enter St Malo in poor visibility.

Were both antennas seeing the sky? ( I know that this has a big influence)

[frankly frankie]

You'd generally have exceptionally good sky view when sailing wouldn't you?  And possibly an external antenna.
Or 2 GPSs to be safe?

AFAIK the leisure GPS sytem is rated at about 15m accuracy.  Any one 'fix' could be a long, long way outside that, but you'd never be aware of it because set against that are the 95% of points that well inside that figure.
In practice on the bike an out-and-back along the same roads will easily record as two parallel(ish) tracks, suggesting a general accuracy of about 4m or so.

[mrcharly-YHT]

You'd generally have exceptionally good sky view when sailing wouldn't you?  And possibly an external antenna.
Or 2 GPSs to be safe?

AFAIK the leisure GPS sytem is rated at about 15m accuracy.  Any one 'fix' could be a long, long way outside that, but you'd never be aware of it because set against that are the 95% of points that well inside that figure.
In practice on the bike an out-and-back along the same roads will easily record as two parallel(ish) tracks, suggesting a general accuracy of about 4m or so.

A sky view interrupted by a socking great aluminium mast and lots of steel wiring. Possibly also interrupted by socking great ships, looming 20-30m higher than your decks.

I don't trust GPS accuracy in harbours. I've had to enter Den Helder with sub-20m visibility - fortunately we had a small radar rig.

[Panoramix]

You'd generally have exceptionally good sky view when sailing wouldn't you?  And possibly an external antenna.
Or 2 GPSs to be safe?

AFAIK the leisure GPS sytem is rated at about 15m accuracy.  Any one 'fix' could be a long, long way outside that, but you'd never be aware of it because set against that are the 95% of points that well inside that figure.
In practice on the bike an out-and-back along the same roads will easily record as two parallel(ish) tracks, suggesting a general accuracy of about 4m or so.

In term of "actual" accuracy, my experience is similar to yours. Freak fixes don't worry me as it is easy to spot one (Generally when you start breaking the boat speed record the GPS fix may be wrong)

You don't want to use 2 GPS as it can only worry you (you would just know that one is wrong). On a posher boat you would navigate with the radar and cross check with the GPS on a smaller one either GPS and cross check with echosounder (what i would tend to do) or dead reckoning cross checked with GPS or if none of the above is precise enough do an "hippodrome" (French term for making a big useless loop) and wait for the fog to dissipate. The trouble with St Malo is that once you proceed you've got to go to the end and it takes an hour to enter.

Sounds scary but when I was a kid we were doing dead reckoning cross checked with echo sounder and once got lost a quarter of a mile before St malo (my Dad blamed the nearby tide barrage for messing up the stream and his dead reckoning)

Edit: cross post with mrcharly

[Cunobelin]

Sorry if I have misled,  Ascent is the programme on the MAC not a hill

Both GPS have an identical access to signal as they are physically about three feet apart

[Panoramix]

Could you post the tracks somewhere?