GPS about 30m off

Seems I cannot edit my own message above.

So in reply to myself I can add that switching from Multi-GNSS to GPS only gives the right location again.

Regards,

  Edwin.

Ah, found the "edit link" now  to edit my own messages..

Edwin, this falls in the category of advice that's too late.  But it would have been useful to know what the device said its accuracy was when it was that far off.  

Since it only occurs with multi-GNSS and resolves with a power cycle, it might be a bug. You should probably open a support ticket. 

The fact that you can see the difference on the map causes me to wonder. Which map are you using. And what datum?

Regardless, be aware that the accuracy estimate (that 1.8m on the satellite page) is based on probability. “There is x% probability that your actual location is within y meters of the reported location.” The choice of x is up to Garmin. This is a balancing act. Marketing vs real world. All you see is the distance. You have no idea what the fixed probability is. Typically on the order of 85%. Point is that (with that example) 15% of the time, the discrepancy will be larger than the reported accuracy.

Edwin said:

On the map I can see the gps is at least 30m off.

the question would be what map? Where it came from and who did it draw with what basic chart datum etc. There is no correct or wrong position, just matter of input data

There is a simpler solution than worrying about map datums in the 67i.  Type the coordinates from the 67i into Google Earth and zoom all the way in.  If the location shown is the same as the one on the map in the 67i, then it's not a map issue.  If the location shown is different by about 30m in the opposite direction of the apparent 67i map error, then a map issue is the likely reason for the discrepancy.  I presume Edwin was using the same map when he got correct positions.  So I doubt it's a map issue. 

Okay, thanks for all your replies!

I have been trying to think of a test script that could give some insight.

I thought that the gps, when in gnss mode, did not recover well from a bad signal reception. It seems not be the case. After a lot of trying, the test script is really simple.

Set the gps in gnss mode, put on a (plastic) table in the backyard and wait.

It starts really well. The gps finds it location fast and according to the map it seem quite accurate. (Standard topo Europa map that came with the device, sometimes using the satellite map layer I downloaded with the gps).

Don´t touch the gps, just sit and wait and after about 5 minutes it is about 20 meters off.

Then I pick up the gps, walk those 20 meters and back to the table. The gps is about on the right location as on the start of this little test a bit of though. I can now watch the coordinates change slowly until it is about 20m off again.

I will now do this test with only gps (no gnss) and report back.

Edit: the given variation on the gps is always 1.8m.

Edit2: I can confirm that in gps-only mode this odd behavior does not occur.

Regards,

   Edwin.

Edwin, your test might be easier and avoid manual record keeping if you turn on track recording and set to a fixed time interval.  

Thanks, I didn´t think of that. Will try and test some more. (Probably tomorrow).

I have the same problem, If in satellite System you configure in demo mode and after Restore default position and elevation are ok again.

It seems it is like a cleaning.