gps-details

Garmin - Static Testing

Author: Michael George

Created: 10 Feb 2025

Introduction

Static testing is very common in GNSS research and literature, but until now has not been feasible on Garmin watches. The primary reason has been that Garmin set very low speeds to zero, making it impossible to do static testing.

The application APPro has access to the GNSS data and has added a developer field called “sog” (speed over ground) to the FIT. This is a quick investigation to assess the feasibility of using the “sog” field for static testing.

Approach

The test used two 1 Hz Mini Motions in addition to a Garmin Forerunner 255 and Fenix 7 Pro Sapphire Solar.

The devices were fixed to a mast extension, and placed in an open garden for 6 hours.

Findings

Timing Issues

Although timings were not the primary focus of this investigation, it is worth noting how the event timings in the Forerunner 255 were not consistent. The timings of FR 255 location events drifted considerably during the 6 hours of testing, whilst the fenix 7 Pro remained relatively constant.

It is also noticeable that the timings of the fenix 7 pro event had occasional blips. The interval between F7 anomalies was consistently 30 minutes.

Measured Speed

The speeds measured by the Garmin watches were consistently low throughout the 6 hours of testing, but with an occasional spike.

The speeds from the Mini Motions were slightly lower, and without any noticeable spikes.

Distributions

The Garmin distributions can be plotted, but limited to the resolution of the speeds being recorded, which is around 0.018 kts.

The Motion distributions don’t have the same limited resolution as the Garmin watches, so can be plotted in more detail.

Results

The p50 and p95 results of the F7 Pro are slightly better than the FR 255, but quantization errors may have contributed to this slightly. It was mentioned previously that the resolution of the speeds being recorded is around 0.018 kts.

The Mini Motions performed better than both of the Garmin watches during the 6 hours, roughly twice the precision of the fenix 7 Pro. Understanding the relative performances may prove to be useful when testing Garmin watches in the future.

	FR 255	F7 Pro	Mini 1040	Mini 1042
p50 (kts)	0.036	0.018	0.010	0.012
p95 (kts)	0.073	0.054	0.027	0.031
p99 (kts)	0.109	0.109	0.039	0.054
mean (kts)	0.033	0.028	0.012	0.014
max (kts)	2.086	1.360	0.121	0.216
distance (km)	0.364	0.310	0.137	0.152

Note: The distance was calculated from the sum of the speeds (m/s) during the 6 hours of testing.

Summary

This was a quick and easy test to establish the feasibility of static testing for specific Garmin watches.

The results are encouraging and suggest this approach may prove to be useful, potentially complementing any real-world testing.

The ability to capture near-zero speeds is currently limited to a private beta of APPro, but will ship in the next public release.