67 vs 66sr - Accuracy

[JungleJim]

While they both use very similar GNSS chipsets (AG3335M in the GPSMAP 67, AG3335MN in the eTrex SE), the eTrex SE version does not support Multi-Band.

Thanks for pointing that out, didn't know the eTrex SE does not have multi-band!

[mimichris]

When searching for the type AG3335MN from Airoha indicates "Search by Key Words AG3335MN: 0 Search Results"
can't find this type of GNSS.

[JungleJim]

When searching for the type AG3335MN from Airoha indicates "Search by Key Words AG3335MN: 0 Search Results"
can't find this type of GNSS.

The Airoha site only mentions the AG3335M series (emphasis mine). But if you search a bit on the internet there are multiple (reliable) results for AG3335MN, see for example this one or this one that show the AG3335MN exists And the difference between M and MN is indeed that MN only supports multiband for NavIC, not for the other GNSSes.

[GPSrChive]

FYI - The GNSS chipset information is already available from the 'Specifications' page for these models at GPSrChive.

[gpsjorgen]

I don't think it has much value to make a statement on accuracy or precision based on one coordinate measurement. In next images in which I put to the test 4 GPSses this is shown (Garmin GPSMAP 67, Garmin Edge 1040, Garmin Fenix 7, TomTom Adventurer). I proceeded as follows:

1. Switch on and wait for a good fix
2. Log a waypoint
3. Walk away 20-100 m in random direction
4. Walk back
5. Repeat from step 2, 15-20 times

The walking was done to make the measurements independent, since GPSses tend to have some dead-band filtering.

I executed this on three days on a Dutch grid reference point. I converted the coordinates to the Dutch RDNew (easting and northing in meter) and plotted the location with respect to the reference (so [0,0] is the reference point). Next charts show the result. The dots are measurements, the red circle indicates the mean value of the measurements, its radius is the average precision error.

accuracy measurments

test22.png

test24.png

test23.png

Note: the waypoint coordinate of the Edge 1040 and GPSMAP 67 were rounded off, so the track coordinate was taken at the timestamp of the waypoint.

Apparently the GPSMAP 67 is the most accurate. Most measurements show an systemtic error in NE direction. I cannot explain. I believe the referencepoint is with cm accuracy, but I cannot have this confirmed...

I did more testing on which I will publish a blog in a few days.

[gpsjorgen]

I put the GPSMAP 66sr and GPSMAP 67 to the test at a known reference point in the Netherlands. I logged 30 waypoints with a little bit (~20 meter away and back) of walking between the individual loggings to get independent measurements.

In next graph I plotted the measurements, where I used the track coordinates i.s.o. the waypoint coordinates, since the waypoint coordinates in the GPX file have limited accuracy (6 digits)

The cross in the center represents the reference point, the individual dots the logged waypoints at this reference point. The circle reprents the average distance to the mean of the waypoints (precision).

Test results plot

test34.png

GPSMAP 67: accuracy 1.1 m precision 0.9 m
GPSMAP 66sr; accuracy 1.1 m, precision 0.8 m
Fenix 7: accuracy 1.8 m, precision 0.8 m

The Accuracy is the average error with respect to the reference
The Precision is the average distance to the mean of the 30 waypoints.

So both devices are equally accurate in this measurement.

[GPSrChive]

Are you going to do any test using the built in Waypoint Averaging feature?

[gpsjorgen]

Are you going to do any test using the built in Waypoint Averaging feature?

As far as I understand the feature you can only log 3 waypoints to be averaged. I did not test this.

I did however an extensive comparison of the GPSMAP 67 and GPSMAP 66sr and reported the results in next blogpost.

Spoiler

blogpost

Bottomline:

• with respect to accuracy and precision under clear sky conditions devices seem to be comparable with respect to accuracy and precision.
• in heavy forrest hiking the GPSMAPS stand out to for example the Fenix 7 wearable. The GPSMAP 67 seems to be a bit more precise and accurate
• the GPSMAPs especially perform well during hiking; for cycling use the Edge 1040, for swimming Fenix 7

However, I found a serious issue on the GPSMAP 67: the coordinate logged starts to drift when you are standing still; this drift can be huge and sometimes result in persistent error. In my opinion this makes the device unreliable.

Spoiler

blogpost

[GPSrChive]

Are you going to do any test using the built in Waypoint Averaging feature?

As far as I understand the feature you can only log 3 waypoints to be averaged. I did not test this.

What makes you think this?

[Nail]

However, I found a serious issue on the GPSMAP 67: the coordinate logged starts to drift when you are standing still; this drift can be huge and sometimes result in persistent error. In my opinion this makes the device unreliable.

I had the same thing, with two GPSMAP67 units. You write on your blog that turning it off/on solves the problem, that's correct. And did you try to start the route when your unit drifted into the next street before you pressed OFF/ON?

Here are my attempts, red is 66sr, green is 67, which drifted. Both units start the route from the same place.

Spoiler

67vs66sr3.jpg

It took some time for 67 to correct its position and travel with 66sr.

One day this problem disappeared and 67 stopped drifting, but now I see it is back again.