ublox NEO-6M accuracy errors in specific locations?

A recent post in another thread describes these interesting symptoms:

  • The tracked route suddenly “jumps” a few kilometers beside the motorway

  • At a given unlucky location, a steady constant false value is returned. Over hour’s the same (wrong) NMEA data.

  • This does not change, even when the GPS’s are directly connected to a computer thru a ftdi RS232 or ch340g (9600 baud).

  • At the same location, other modules provide correct data.

jeff-jordan says that “it’s not a signal reflection problem, or a problem with unprecise route data, the error is too big for such effects”.

Personally, I would suspect satellite reception or interference problems in those areas, and that other brands perform “better” in those conditions. It seems that we can eliminate the Arduino sketch, because a direct connection to a PC shows the same errors.

Questions for @jeff-jordan:

  1. Did you get low DOP values?

  2. Did the GSV sentences show a smaller number of satellites or lower SNRs?

  3. What is a “steady constant false value”?

Questions for everybody:

  1. Has anybody else seen these symptoms? There are always “problem” areas, and different modules perform differently, but is there any other factor to consider besides reception and interference?

  2. Has anybody seen a comparison of different modules under noisy conditions, something more than a comparison of spec sheets?


Just hooked up my neo-6m module using a sarantel helix antenna. The patch was taking took long to get a fix. Anyway hanging the GPS by window it appears that in terms of average coordinates I am within 25 meters. In terms of true (delta to my approximate average) I tend to be within 50 meters. I have never seen errors in the kilometer range. All stats are shown in the following figures.

My baud rate is 57600, 200ms update rate, nav5 mode automobile, sbas turned off.

You might also want to check out this post of Ground Planes and antennas ( it applies to M8N ) but may be appropriate.