10 GPSr - 10 different co-ordinates

[+kbootb]

 

I hope this picture is readable.

 

I happened to have 10 identical Garmins, fresh out of their boxes. I configured each one and put it on the grass bank while waiting for the children I was training to arrive. After 20 mins they were still showing slightly different readings, so I took a picture to show you.

 

In case you can't read the numbers they range from

 

N 52 deg 15.174 to .178

E01deg 37.577 70 .588

 

No two units ever had the same co-ordinates.

 

This is a very open site, heathland at the top of a cliff, great view of the sky and lots of satellites visible.

 

Just goes to show how much variation there can be.

[+kennamatic]

That would explain all my DNFs!

 

Of course the ones I find I are down to my skill!!!

[+John Stead]

Yes but they're all in slightly different positions!

Did not realize they could be so accurate!!

Could have been interesting if you had set the same waypoint into each of them to see what distances they came up with

[+brdad]

What I find interesting - if you load the satellite page, you will find that some of the GPSrs are using different satellites than others.

[+kbootb]

Yes but they're all in slightly different positions!

Did not realize they could be so accurate!!

Could have been interesting if you had set the same waypoint into each of them to see what distances they came up with

 

Actually, that's what we did. I had set up a course around the site on one unit, saved the waypoints on the computer and uploaded them to all units so they were identical.

 

The students had one GPS for each pair and to start them off we did the first one together. They were all crowded round me as they were 'trained' in going through the pages, so the units were not more than a metre apart. The distance to go varied by a max of 3 metres and when we got to the first waypoint they ended up about 2 metres apart when they declared they were at 'zero' (some of course never got to zero, but all got to 1 metre to go).

 

We were running the activity over 3 days and had similar variation each day. They all clustered with the same accuracy, but interesting to note that the point that they all got to varied by about 5 metres from one day to the next. Certainly accurate enough for them to work out which 'feature' of the site we were trying to get them to.

 

They were pretty impressed by the accuracy.

[+kbootb]

What I find interesting - if you load the satellite page, you will find that some of the GPSrs are using different satellites than others.

 

Yes we were surprised by that. These were fresh out of the box and we switched them all on at the same time so they were doing a cold start. Interesting that they found the satellites in different orders, but out of all the satellites on offer, only 2 were the first ones found. Once this was found they put 3 or 4 satellites in the correct place and then found another, which put a few more in the correct place.

 

But as you said, they didn't all have the same list of satellites and didn't have them all in the same place in the sky.

[+Team Sieni]

They all clustered with the same accuracy, ...

Did that correlate well with the estimated accuracy that the GPSr was reporting?

 

Really fascinating stuff.

[+John Stead]

Yes but they're all in slightly different positions!

Did not realize they could be so accurate!!

Could have been interesting if you had set the same waypoint into each of them to see what distances they came up with

 

Actually, that's what we did. I had set up a course around the site on one unit, saved the waypoints on the computer and uploaded them to all units so they were identical.

 

The students had one GPS for each pair and to start them off we did the first one together. They were all crowded round me as they were 'trained' in going through the pages, so the units were not more than a metre apart. The distance to go varied by a max of 3 metres and when we got to the first waypoint they ended up about 2 metres apart when they declared they were at 'zero' (some of course never got to zero, but all got to 1 metre to go).

 

We were running the activity over 3 days and had similar variation each day. They all clustered with the same accuracy, but interesting to note that the point that they all got to varied by about 5 metres from one day to the next. Certainly accurate enough for them to work out which 'feature' of the site we were trying to get them to.

 

They were pretty impressed by the accuracy.

 

Many thanks Keith - I found that quite fascinating. I wonder if anyone will repeat the experiment given the opportunity of so many identical GPSrs being available?

[+FollowMeChaps]

Most of us get by with just the one unit. You must have some serious Agoraphobic tendancies with a serious fear of getting lost!

[+Gralorn]

Lottery Grant ???????????? We notice this sort of thing with just two GPS's, hers will often differ to mine in time to destination and yet we are walking 'hand in hand' (cue Bluebirds) until we are almost at the cache. So that experiment with 10 does not really surprise me..... 10 !!

[+The Forester]

Lottery Grant ???????????? We notice this sort of thing with just two GPS's

 

The odds against getting 10 GPSrs to show the same co-ords simulateously are pretty much the same as the odds against winning the Lottery.

 

Even if you believe that it's possible to start 10 GPSrs exactly simultaneously; and even if you stacked them vertically atop eachother so that the antennae did not obscure eachothers view of the low horizon angle satellite which are the ones which most influence Lat/Long output co-ords, you'd have to have 10 matched GPSrs which were previously shut down at the same time/place as eachother and had very exactly the same previous life history of self-calibration of their internal temparature sensors and internal oscillators.

 

To start them all simultaneously, you have to recognise that all the GPS satellites are transmitting their navigation messages at a datarate of 50 bits per second, so you'd have to get them all looking at exactly the same satellites, all starting within less than a 1/50th of a second of a second of eachother.

 

Can't be done, I think. Not even on a "fastest finger first" basis.

 

Nice try, though!

[+kbootb]

They all clustered with the same accuracy, ...

Did that correlate well with the estimated accuracy that the GPSr was reporting?

 

Really fascinating stuff.

 

Yes, all results well within expected 'error'.

 

We set the waypoints up about 4 months ago when we did a pre-visit to plan the trip. This was the first unit I have had with an averaging function. We left the unit averaging for about 5 mins at each point.

 

That was interesting, watching the 'accuracy' work down from about 15m to around 3m.

 

So, makes me wonder why they give 3 decimal poitn accuracy when it clearly isn't and, what sort of distance does each decimal place equate to in terms of metres across the ground.

[+The Forester]

watching the 'accuracy' work down from about 15m to around 3m.

 

The "Accuracy" figure is a misleading misnomer.

 

Accuracy is the difference between the indicated position and the true position. To know the accuracy of the position, your machine would have to know the true position. If your machine knew the true position, then surely it would tell you the true position, in which case the accuracy would always be zero.

 

Your machine does not know the true position, which is why it cannot tell you the true position, which is why it cannot tells you its accuracy.

[+kbootb]

watching the 'accuracy' work down from about 15m to around 3m.

 

The "Accuracy" figure is a misleading misnomer.

 

Accuracy is the difference between the indicated position and the true position. To know the accuracy of the position, your machine would have to know the true position. If your machine knew the true position, then surely it would tell you the true position, in which case the accuracy would always be zero.

 

Your machine does not know the true position, which is why it cannot tell you the true position, which is why it cannot tells you its accuracy.

 

Exactly why I put 'accuracy' in inverted commas. 'Error' would be more accurate, but people without a science background confuse 'error' with 'mistake' rather than 'limitations'. If it read 'certainty' or 'limit to accuracy' I'd be happier.

[+The Forester]

Exactly why I put 'accuracy' in inverted commas.

Exactly why I grumble about Garmin claiming their Accuracy figure to be an indication of "accuracy"!

[+kbootb]

Thanks to Master Mariner's post in another thread that gave me the distance for minutes of Lon and Lat I have done some calculations.

 

Taking the most northerly and southerly of the readings gives a range of 11.118 meters

 

Most Easterly and Westerly gives a range of 12.86 metres

 

The greatest north-south distance from the averaged centre is 6.86 metres

 

The greatest east-west distance from the averaged centre is 6.99 metres

 

The greatest direct distance of any one unit from the averaged centre is 7.67 metres, which happens to be the one at top left, which is probably about 20 cms from the centre of the array.

 

There you go, the geek in me is now fully satisfied.

 

I would say that this all fits well within the expected 'accuracy' of these units.

[+Pharisee]

Silver-Fox and I were on the Island of Harris recently and needed to check the location of a cache that was 'under review'. We put our GPS60s side by side at the cache site and left them 'averaging'. After 20 minutes or so, they were both reading exactly the same. Mine was indicating an 'accuracy' of just 4.7 feet which is the best I have ever seen. It was showing an almost full strength signal on 10 satellites with a 'D' in all the bars.

 

OK... it was on a hill, on the coast and the nearest tree was probably 30 miles away but I still reckon it's pretty impressive

[+GeoCornetto]

Last week there was a large group of us out caching. We put all our GPSrs together just to find they all upset each other. My yellow eTrex messed up djsnipers gpsmap 60csx and so on.

Does anyone know why?

[+Simply Paul]

When setting a new cache in Scotland (the lighthouse keeper is maintaining it, before there are any moans) with Silver-Fox the other week we left of GPSi on the same rock, averaging. Although mine's an ancient yellow running ancient firmware, and his is a bells and whistles thing with colour mapping (60s?) they both showed within 0.001 minutes N/S and E/W of each other after only 3 mins averaging. There's no science behind GPS accuracy - it's pure luck!

[NickPick]

Last week there was a large group of us out caching. We put all our GPSrs together just to find they all upset each other. My yellow eTrex messed up djsnipers gpsmap 60csx and so on.

Does anyone know why?

 

I can only suggest that yours upset the others by re-radiating some of the signal. All antennas re-radiate the signal they are receiving (this is how a TV detector van can detect if you're watching your tv). The others must have detected this signal and accepted it as a valid signal, which has confused their calculations. I suppose this would be similar to the effect of multi-path where a signal gets bounced off a wall or cliff (used to show up on TV as ghosting, but shows up on the GPS as receiving the sattelite signal slightly later than it should for its location).

 

That's my thoughts on the possible reason. I'm sure that the Forester will come along and correct me if I'm wrong. (and I'd be happy if he did)

[+careygang]

I got new 4 yellows out of their packages (brought for the Scouts) and tried them next to each other; got pretty much identical readings and accuracy levels from each, certainly what I'd expect.

[+PopUpPirate]

I always said Etrex were c***

[+Team Sieni]

Just had a pedantic think and isn't the correct term "precision" here, rather than "accuracy" or "error" (I know I used accuracy above!)

 

I recall learning that you could have equipment that was precise but inaccurate (eg a scale that is precise to milligrams but is wrongly calibrated and is out by 2 grams)

[Iwuzere]

All antennas re-radiate the signal they are receiving (this is how a TV detector van can detect if you're watching your tv).

I believe the detector vans actually detect a signal that the TV generates in order to translate the received signals to a fixed frequency where the demodulation takes place. See http://en.wikipedia.org/wiki/Local_oscillator

 

[+The Forester]

The reradiating aspects of a GPSr antenna are certainly true, but don't play a part in this phenomenon. If you connect two detatchable GPS antennae together with a length of co-axial cable, you can pop one of them outside on a building's roof and then rig the other one indoors. The outdoor one will pick up the satellites' signal and the signal will be reradiated indoors by the other antenna. We use this method to check and test any number of GPS receivers in a lab or workshop environment.

 

I don't doubt that if you were to take ten identical eTrexes out of their boxes indoors; cold-start them; then connect the indoor antenna to its bit of electric string as I've just described: then you might be able to get them all to behave like the clones they are and show identical responses to the reradiated signal. Doing that out in the wild will produce the variety of output shown in the root post picture of this thread.

 

The reradiation of signals by tellies is a different thing. That reradiated signal comes from internal oscillators within the receiver which are used to process the incoming signal. Those signals cannot be confused by a telly, just as a GPSr would not be confused by the radiated local oscillator signals of a nearby GPSr (I think).

 

There is a phenomenon called "parasitic element" in which two identically tuned antennae can mess eachother up somewhat by 'sucking' the signal at their tuned frequency and therefore depriving a neighbouring antenna of the full strength of the broadcast signal. I suspect that this was going on in the picture. The ten GPSrs are arrayed in a SouthWesterly orientation, which means that a low-angle signal from Northerly satellites would be eagerly sucked up by the units in the bottom-right corner of the picture and the signal would be much less well received by the units in the top left-hand corner of the picture. For this reason I would not expect the units to be showing the same signal strength from the satellites as eachother, nor even necessarily using the same set of satellites at all. Remember that it's the low angle satellites, especially those below about 30°, which contribute most to the determination of Lat/Long. The ones much above 60° above the horizon do contribute to the third and fourth dimensions of a GPS fix, but do not help us with the two dimensions which most interest geocachers.

[+GeoCornetto]

The reradiating aspects of a GPSr antenna are certainly true, but don't play a part in this phenomenon. If you connect two detatchable GPS antennae together with a length of co-axial cable, you can pop one of them outside on a building's roof and then rig the other one indoors. The outdoor one will pick up the satellites' signal and the signal will be reradiated indoors by the other antenna. We use this method to check and test any number of GPS receivers in a lab or workshop environment.

 

I don't doubt that if you were to take ten identical eTrexes out of their boxes indoors; cold-start them; then connect the indoor antenna to its bit of electric string as I've just described: then you might be able to get them all to behave like the clones they are and show identical responses to the reradiated signal. Doing that out in the wild will produce the variety of output shown in the root post picture of this thread.

 

The reradiation of signals by tellies is a different thing. That reradiated signal comes from internal oscillators within the receiver which are used to process the incoming signal. Those signals cannot be confused by a telly, just as a GPSr would not be confused by the radiated local oscillator signals of a nearby GPSr (I think).

 

There is a phenomenon called "parasitic element" in which two identically tuned antennae can mess eachother up somewhat by 'sucking' the signal at their tuned frequency and therefore depriving a neighbouring antenna of the full strength of the broadcast signal. I suspect that this was going on in the picture. The ten GPSrs are arrayed in a SouthWesterly orientation, which means that a low-angle signal from Northerly satellites would be eagerly sucked up by the units in the bottom-right corner of the picture and the signal would be much less well received by the units in the top left-hand corner of the picture. For this reason I would not expect the units to be showing the same signal strength from the satellites as eachother, nor even necessarily using the same set of satellites at all. Remember that it's the low angle satellites, especially those below about 30°, which contribute most to the determination of Lat/Long. The ones much above 60° above the horizon do contribute to the third and fourth dimensions of a GPS fix, but do not help us with the two dimensions which most interest geocachers.

 

Wow. How do cats work?