SparkFun Forums 

Where electronics enthusiasts find answers.

Tips and questions relating to the GPS modules from SFE
By BitSplitter
Hi all!

I've built a SMS GPS Tracker for my car. I'm using Telit GM862 Quad GPS module. It seemed very practical.. having both GSM and GPS functions in one module. The device is actually working fine except for one problem. GPS position readings vary a lot. I'm using the external active GPS antenna. If there are no buildings around my car then everything is OK. As soon as I park near some building the readings vary +- 20 meters at least.

Now, I'm thinking of using some other GPS receiver which can be more accurate if sorounded by tall buildings.
I'm considering using UBLOX LEA 6S module. Nevertheless, surfing the internet I've found about the SIRF STAR III chipset.
Can somebody tell me the diference between UBLOX LEA6S and some GPS receiver based on SIRF STAR III chipset?
Also, can someone provide a part number of some GPS receiver based on SIRF STAR III chipset (I mean the module which can be placed on the PCB)?

Thanks in advance!
By trillium
What kind of antenna are you using, and is it pointing any other direction than up. I guess I mean, does it have a good sky view?

What you are describing is called multipath. Its an area where GPS receivers vary a great deal in their ability to handle the selective loss of information from some satellites and the confusing reception of other satellite signals twice or more times, with different delays, caused by buildings.

Some of it is unavoidable, but perhaps there is some setting that has gotten messed up. I would try resetting your module's GPS related settings to defaults. Are you talking about deep "urban canyons" or merely places where one building is next to you. IMO a GPS should be able to handle the simplest case of one building or two adequately enough to only vary a bit in the fix displayed. But in places like New York City or Tokyo, where the sky is a blue strip of at the most, maybe 30 degrees width, far, far above, all bets are off!

Similar problems exist in forested areas, but I would expect the RF behavior to be somewhat different because trees probably do not reflect quite as much RF as metal framed buildings, rather, they absorb much of it.

I saw a program a few days ago that looked as if it might offer a lot of utility for testing antennas - its on David Taylor's web site..its called GPSHorizon - its only windows - so I have not tried it yet.. (I'm mostly a Mac and Linux person) But it lets you compute a GPS polar plot, which you can use to evaluate an antenna setup.

Caveat.. I am just a beginner in this, so I may be missing some important, obvious thing.