Poor GNSS accuracy? [Bug 8]

[MarkHL]

Software version 4.80;

Here's a comparison of the GPSMAP 67 (Red) and a Garmin Epix (Blue) under a dense tree canopy on a mountain bike. The time interval for the 67's track is 1-second; similar to the 1-2 second interval on the Epix when it is laying a track in "Smart" mode.

The Auto setting for tracks on the 67 seems fairly accurate if I have a mostly open view to the sky (e.g., walking in my neighborhood) but loses a lot of detail under trees cover.

Garmin Explore Screen

Screenshot_20230412_120154_Explore.jpg

*** Image formatting corrected to comply with forum guidelines ***

[Gadwin]

To those who had a displayed gnss accuracy below 1.8m, is it still the case with firmware 4.80? Now I get almost always 1.8m, not right on the beginning like on the 66sr, but after a short time.

You wanted it, you got it. Your dreams have come true.

I am not quite sure what I got, because I had today some really strange tracks, the behavior was really utmost dubious. I have to do some more records, because it was unreal bad and I can't say for sure why. It is also, that the 1.8m appears almost everywhere, where it should not be the case.

[Nail]

So I got my GPSMAP 67 and noticed immediately, that the accuracy is worse than the GPSMAP 66sr. With the GPSMAP 66sr I have almost always 1.8m accuracy and it is really stable. With the GPSMAP 67 I have very often 3m and even this is not really stable. Sometimes it is 4m, sometimes 2.5m and is really wobbly. But almost never under 2m.

You expected, like the 66sr, your 67 to have a stable good accuracy. Your and others' wish has been granted.
Probably Garmin didn't put much work into it. Only this accuracy applies to the display, not reality as you say. If they really did, what's the difference between Garmin and VW?

[mimichris]

According to the Rinex graph I posted, what worries me is the way the GPS signal is managed, intermittently and thus favoring battery life at the expense of GPS accuracy. It's my opinion.

[JungleJim]

According to the Rinex graph I posted, what worries me is the way the GPS signal is managed, intermittently and thus favoring battery life at the expense of GPS accuracy. It's my opinion.

I wouldn’t base any conclusions on the RINEX data in its current form. The data is not formatted correctly (Galileo and BDS satellite numbers), contains duplicate observations and is incomplete (not all bands are included).

[Nail]

According to the Rinex graph I posted, what worries me is the way the GPS signal is managed, intermittently and thus favoring battery life at the expense of GPS accuracy. It's my opinion.

If you are right. It's not going to improve the quality of the gpx tracks. Garmin has announced long battery life and they can't back down from that. Your theory agrees with my observations.
Two weeks of testing practically did not show the superiority of 67 over 66sr, quite the opposite.

[Przekątny]

Reliable verification of position indications (not GNSS accuracy) is the reading of location data indicated by the device
in a place with known coordinates, such a place may be a geodetic control located in the open space.
Another way may be to read the coordinates in two characteristic points, physically measure the distance between
the points and compare the distance between the points saved in the device.

[MarkHL]

According to the Rinex graph I posted, what worries me is the way the GPS signal is managed, intermittently and thus favoring battery life at the expense of GPS accuracy. It's my opinion.

I came to the same conclusion after seeing a poorly defined track with the recording interval set to "Auto" and "Most Often". I boxed the unit up for return. However, I decided to to set the recording interval to "Time" and a 1-second" interval and I got a well define track under a tree canopy (mountain bike) where my 66s had previously struggled. I believe the accuracy has improved, I just have to force the the increased placement of track points.
Having the GPS lay a track point every second seems extreme, but it's practically what my Garmin Epix does on "Smart" mode. The other option is "1-second". The Epix also advertises battery life, but it lays a lot of points down. For example, yesterday, the Epix recorded @2500 points for an activity and the 67, @3000 points set at 1-second intervals. With the hope Garmin will make improve the accuracy of the tracks with future updates (given the satellite availbility) and having found a suitable work-around, I decided to keep the unit.

[Gadwin]

According to the Rinex graph I posted, what worries me is the way the GPS signal is managed, intermittently and thus favoring battery life at the expense of GPS accuracy. It's my opinion.

I came to the same conclusion after seeing a poorly defined track with the recording interval set to "Auto" and "Most Often". I boxed the unit up for return. However, I decided to to set the recording interval to "Time" and a 1-second" interval and I got a well define track under a tree canopy (mountain bike) where my 66s had previously struggled. I believe the accuracy has improved, I just have to force the the increased placement of track points.
Having the GPS lay a track point every second seems extreme, but it's practically what my Garmin Epix does on "Smart" mode. The other option is "1-second". The Epix also advertises battery life, but it lays a lot of points down. For example, yesterday, the Epix recorded @2500 points for an activity and the 67, @3000 points set at 1-second intervals. With the hope Garmin will make improve the accuracy of the tracks with future updates (given the satellite availbility) and having found a suitable work-around, I decided to keep the unit.

It would be disastrous if the 67 wouldn't be even more accurate than a 66s. But compared to the 66sr it struggles alot.

[MarkHL]

Even when the 67 was laying an ugly track (lack of detail) I saw the points that were laid were accurate. Just not enough of them, so I got straight lines where I should see corners. I wish they had a menu option that allowed the same algorithm to be used the Epix is using. However, the Epix uses a different chipset(?) and the watch is very sensitive to movement, so that might not be possible. The 67's release is not yet a month old, so I'm hoping for future improvements.