I’m just wondering what type of GPS receivers you guys use or which ones you would recommend.
I have been using the handheld Garmin E-Trax (single freq) and don’t seem to be getting very accurate results with it (prob around average 4-5m out)
I’m at university and have access to a large range of equipment, so fire away!
Please note that 4-5 accuracy is pretty good for a commercial GPS receiver, it is about the same as my GPSmap 60CSx in good conditions. I usually do not trust the claims the unit makes normally, it tends to be a bit optimistic overall.
There are lots of configurations possible here. Outdoor GPS like Garmin, PDA with builtin GPS or with a mouse, Nokia N800, car GPS like TomTom. It all depends on how you intend to mainly use it for (e.g. outdoor sport, car, etc.). Do you plan on cycling a lot for OSM then an outdoor is recommended (shock and water resistant), if you plan only do mapping by car the a TomTom or PDA is a good choice (PDA the better of the two I think).
So lets assume you are in the market for an outdoor GPS. The Dutch OSM group is also searching for an outdoor GPS with map showing capabilities and I recommended them the Garmin Etrex Venture Cx because it has: SD-card for large maps, color display, ‘unlimited’ tracklogging on SD-card, 12 channel GPS receiver, USB connector and routing capabilities. Overall it seems like a good tradeoff feature/money wise.
Ofcourse, I own a GPSmap 60CSx which is a beast of a GPS and I really love it, but it’s pretty expensive as well.
I’ve got a garmin 76cs, which is almost identical to the 60cs, except it floats, and has tide predictions. (+visual/size differences) The ‘x’ is vital for osm though. I would agree that the 60csx is a great gps. I really wish I had the 76csx or 60csx, as the inbuilt memory really isn’t enough for OSM, and that is the only problem I have had with my gps.
The figures for how accurate gps’s are don’t seem to mean to much I find. I tried out my gps on road to the next village, and then took another track riding back from that village on the other side of the road.  The 2 tracks seem to line up pretty well, and stay parallel the entire way (red lines are the top/bottom of the hill). So at riding speed in the summer in a none built up area with nothing blocking the satellite, the accuracy seems to be within a couple of feet. Unfortunately it can sway well off by 30 foot and still confidently tell you it knows where you are if your not it the most suited place for a gps such as riding around a woods in the summer. (Although standing still for 10seconds tends to give it time to get it’s act together. Recording at a low amount of points per second seems to produce less accurate data though for some reason.
Only 2 things that I would say are poor about the 60/76csx are the cost, and the metal bits in the battery compartment…they seem to snap rather easy.
I’m surprised you say the E-trex is giving you bad data though. I have noticed a difference in the accuracy and sensitively between my gps and my sisters etrex though. Hers will loose signal way quicker. This may be to do with the size of the units, and the bigger unit may be able to detect weaker signals. I’m not sure though.
I agree with “Recording at a low amount of points per second seems to produce less accurate data though for some reason” as my results got worse when I set it to 1 point per second. What do you feel is best for just walking?
Thanks for the info guys.
I actually meant the reverse of how I think you read that. A low amount of points per second such as 1 point per 5 seconds seems to produce poor track logs, while a high amount such as 1 every second seems better.
I varie the amount of points per second depending to what level of detail I am mapping to, (and also how much room I have left). For walking if you map super fine detail* then I would suggest a point every second still or 1 every 2 seconds, but if there is little detail then 1 point every 10 seconds is fine. For example when I was in Scotland the other week I had it at 1 every 10 seconds, because the landscape was bare, and little more than grass and rocks. Trodden paths covered huge distances without much change in direction. The rarity of mappable features meant it was easier to way point them when I crossed them, than make them visible in the gpx track (from clusters of dots, or shapes for example). The down point with this is when the signal was weak and there was a high chance of not adding a way point then 1 missed point would make the gap huge. Sometimes there was minuets between points and they were highly inaccurate. If it had been at 1 point per ‘a low amount of seconds’, then I could have clearly seen where the signal cut off and worked out where the data was moor or less reliable, and could have enshored that there wasn’t a long period between points.
- (gates/hedges/gaps/stiles/buildings/trees etc.)
Usually one should adapt the sample interval depending on my way of travel. Walking usually only requires a 5 sec. interval, while cycling every 2 sec and in a car about every second will produce a good quality track. But to be honest, as my tracklog is unlimited, I usually forget to set it (or I’m too lazy) and therefore most of my tracklogs are with a 1 sec. interval.