It all depends on your particular use, application or need in a GPS device. For Farm Management, slow moving accuracy is most important, for a walker going through heavily wooded terrain in the Brazilian jungles, sensitivity would be the most important metric. For us here in Oz, sensitivity is not that important, our sky is pretty open, except of course in the cities, particularly Sydney. Our bush is not that heavily wooded and pretty well any GPS works in our rainforests in most places.

So does having a super sensitive 50 channel GPS with every feature under the sun help you get a lock better than a less spec'd unit amongst the tall buildings of a city? Most likely not! Why you ask? Because if a receiver is more sensitive, it also means it will receive more echoes, reflections and noise from the surrounding environment. The density of tall buildings & electronic equipment in the city is of course more than in the suburbs, and each peice of equipment produces massive amounts of radio noise. This radio noise does affect your GPS receiver capabilities, and the reflections knock about the accuracy too.

All manufacturers and certainly retailers show the sensitivity, and they show accuracy with some weird codes and percentages after the figures. Common figures usually quoted are something like 7.5m w/o SA, means; 95% percent of the time, with Selective Availability disabled, the unit will be within 7.5m of the correct location.

There is a simple way to show this error. If we place a GPS in a spot for 5 min, and log the location each second for 300 secs, we get a scattergram showing the wander or drift of the given GPS. This test shows some interesting results, and it makes it clear why the choice of chipset becomes important for some specific needs.

Here we are attempting to show the difference in the performance of the various chipsets in common use in the market place. The test is not biased in any way to favour any particular chipset, we sell, have sold or support devices that use all these chipsets, so have no preference or bias. The idea is to show which chipset might be better suited to your particular need. For general car or boat navigation, it simply does not mater as the navigation software generally snaps you on the road, and what is 5m accuracy in a car anyway?. And with Marine, +/- 5-10m doesn't really matter either. A surveyor or Farmer though needs accuracy. Why does a Farmer need accuracy? A lot of Farmers these days use high tech farm management practices for such main reasons as cost savings, and better crop production with less pesticide use etc. The crop spraying tractors are fitted with GPSs so the area being sprayed is not overlaped as it was in manual spraying techniques. Thus saving spray and us!

The differences in drift are caused by many reasons, from simply noise or reflections in the signal reaching the GPS to the software algorithms used in the chipset computations for smoothing. Some chipsets do more averaging to show a more accurate position. Drift in a consumer grade GPS device is normal and to be expected.

• Sony
  
• SiRF Star II L/P
  
• SiRF Star III(With Static Navigation Function Disabled)
  
• Nemerix

If either or both of these conditions cease to become true, then position updates continue. Note that it is position updates, not the data stream that is placed on hold. NMEA data will still be sent when Static Navigation is enabled, but the position in the NMEA stream will not change. All other data such as time and date do still update. So from this we can see that a SiRF-III device with Static Navigation turned on will be useless for a bush walker, but great for a car PND unit as it will not start to wander when sitting at a set of traffic lights etc.