This is a report on the issue of the Garmin GPSMAP 67 in which we observe an huge drift of the location it indicates when satellite reception is not optimal. Next picture shows the coordinate as recorded by the device overnight when placed stationary at the window. At the start the location is indicated correctly (green symbol).
Images have been taken from the Garmin Connect website. An investigation has been made into the problem. This document reports on the findings.
Device
| Device | Garmin GPSMAP 67 |
| Serial number | 7QH027772 |
| Firmware version | 8.90 |
| GPS Firmware version | 11.02 |
| GPS setting | Multi-glnss, Multi band |
| Recording | Interal, one record every second |
| Activity | Hiking |
| Expedition mode | Never |
Finding
Description
The coordinate that is indicated by the GPSMAP 67 drifts over time. It happens when the conditions for satellite reception are not optimal, especially if is placed stationary. Deviations of >100 m have been observed multiple times. The problem was reproducible: the device was placed stationary in house against the window, lying on its back. The problem is immediately resolved when switching the device off and on again: the indicated coordinate jumps to the ‘real’ position.
Unobstructed sky view
When the device has clear sky view and is upright (antenna vertical pointing up) we find good precision. Next scatter plot shows the result of 12 hours of recording (44284 measurements). Average distance to the mean value of the coordinate: 1.13 m, max distance 3.9 m. The circle indicates the CEP (Circular Probable Error). An excellent result, which we might expect given the error indication on the device of 1.8 m.
When the sky view is unobstructed but the device is lying horizontally (display facing upwared, the abnormal drift has been observed as well a few times. In next picture the device (FW 8.90) has horizontally under unobstructed sky conditions for 2h17′. Green symbol indicates the real position.
Firmware version
The intial test took place with latest firmware to date (25-01-2025): FW 8.90 (GPS FW: 11.02). Firmware was downgraded to FW 7.60 (GPS FW: 2.46). Also here we see the error, though it appears that the error is not as large (no statistical proof for this).
Effect of moving
When starting to move, the error seems to reduce. Two examples are shown in next image, FW version 7.60.
However there seems to remain a residual error after movement, which disappears on a power cycle. This is shown in next image for FW 7.60. The track is started after a period of stationary placement, so it indicates the error. Then the device was moved outside, reducing the error. Finally the device was power cycled and logging resumed. This resulted in the indicated coordinate to jump to the actual position.
GPS only vs Multi-GNSS
Effect has been observed in Multi-GNSS but also in GPS Only mode (FW 7.60)
Comparison with other GPSses
Next image image shows a comparison between 3 GPSses: Garmin Edge 1040 bike computer, Garmin Fenix watch and the Garmin GPSMAP 67 (FW 7.60). Devices were left stationary for 3h40′ and then taken outside for a small walk. Whereas the former two exhibit a pattern we might expect, the GPSMAP 67 clearly differs.
What we expect
The Garmin support site defines following for GPS drift:
‘GPS drift, in the most practical sense, is the difference between your actual location and the location recorded by a GPS receiver. Since consumer-grade GPS receivers (such as the ones Garmin makes) are not 100% accurate, this can cause a difference between your actual location and your recorded location.‘
‘The GPS location accuracy of Garmin outdoor watches is around 3 meters (10 feet), 95% of the time. This means, at any given time, your watch will save your location within 3 meters of your actual location.’
Note: given the better antenna we expect the GPSMAP 67 to perform better than the wearables.
As is shown in the pictures above the indicated location is far from ‘within 3 meters 95% of the time’. The statement ‘outside 3 meters 95% of the time‘ comes more close.
Interpretation
Maybe the GPSMAP 67 applies some kind of dead-reckoning based on digital compass course and accelerometers when satellite reception is not exactly optimal. In dead-reconing, based on a known position (fix), next position is calculated using speed and direction. When not comparing to real position, the position error increases. Due to sensor noise the course is rather random.
In my opinion, if this is the case, it must be feature that can be disabled. Furthermore, it should be applied only to short intervals and always the GPS location should be used (with some averaging) to correct the position.
However, this is guessing…
The fact that the problem immediately resolves when a powercycle is executed indicates that the device knows the proper GPS position. Therefore the problem can be solved.
Conclusion
The GPSMAP 67 shows an abnormal deviation of the indicated coordinate. In our findings this not to be expected given the description on the Garmin Support site.
Though movement improves the coordinate, residual error remains.
Deviations have been reported on the Garmin Forum (see: GPS about 30m off) which indicates it is not an incident or one device instance malfunctioning.
This problem renders the device unusable for practical use like Geocaching. You cannot trust the coordinate it indicates, especially when you pause for a moment under conditions where the sky is obstructed or the device is not exactly upright.
We hope Garmin solves the problem soon. The problem has been reported with Garmin Support Netherlands and is known under references Q#:1001450 and Ref:29213587K0.