ubx-cfg
Static ESP Testing - Phase 4
Overview
The objective of phase 4 was to determine the effects of limiting the number of satellites @ 5 Hz. The SYRAC GPS devices were all intended to use GPS + Galileo + BeiDou B1C (no GLONASS), and various satellite limits - 24, 28, and 32.
The clock rate of the ESP32 was also set to 80 MHz to see whether it would be capable of the 5 Hz rate, without dropping any frames. It was noticeable that almost all of the devices dropped frames, which inspired the 160 MHz tests for phase 5.
The tests were started on 2026-07-10 @ 2311 and the duration was over 10 hours. The results showed max sats = 28 to be optimal @ 5 Hz, and 5 Hz to be more accurate than 10 Hz. There was also a hint that D1 and D2 may have a slight performance advantage.
Configurations
The following device configurations were tested.
| ID | Constellations | Rate | Max Sats | ESP | Observed time distribution (ms) | Drops? |
|---|---|---|---|---|---|---|
| SSY1 | GPS + Galileo + BeiDou B1C | 5 Hz | 24 | 80 MHz | 199: 2, 200: 184003, 201: 2, 400: 6 | Y |
| S3 | GPS + Galileo | 5 Hz | 24 | 80 MHz | 199: 2, 200: 94680, 201: 2 | - |
| D1 | GPS + Galileo + BeiDou B1C | 5 Hz | 28 | 80 MHz | 199: 13, 200: 183228, 201: 13, 400: 6 | Y |
| D2 | GPS + Galileo + BeiDou B1C | 5 Hz | 28 | 80 MHz | 199: 5, 200: 183224, 201: 5, 400: 3 | Y |
| D3 | GPS + Galileo + BeiDou B1C | 5 Hz | 32 | 80 MHz | 199: 18, 200: 183737, 201: 18, 400: 5 | Y |
| D5 | GPS + Galileo + BeiDou B1C | 5 Hz | 32 | 80 MHz | 199: 13, 200: 183253, 201: 13, 400: 8 | Y |
| SY2 | GPS + Galileo + BeiDou B1C | 10 Hz | 32 | 160 MHz | 99: 18, 100: 363251, 101: 38, 199: 20, 200: 2098 | Y |
Notes:
- S3 was intended to be GPS + Galileo + BeiDou B1C, but it somehow switched to GPS + Galileo
- Devices using 3 systems @ 5 Hz + 80 MHz were all observed to drop frames, but SY2 dropped a LOT of frames
- SY2 @ 10 Hz + 160 MHz dropped a LOT of frames, and problems with speed and sAcc were evident in the data
Statistics
Click the GPY filenames for charts showing SOG, Sats, HDOP, and sAcc.
| File | Configuration | Mean | Median | Stddev |
|---|---|---|---|---|
| D2__2607112346.gpy | GPS + Galileo + BeiDou, 28 sats @ 5 Hz | 0.021 | 0.019 | 0.011 |
| D1_2607112346.gpy | GPS + Galileo + BeiDou, 28 sats @ 5 Hz | 0.021 | 0.019 | 0.011 |
| D5_2607112346.gpy | GPS + Galileo + BeiDou, 32 sats @ 5 Hz | 0.023 | 0.021 | 0.013 |
| D3__2607112345.gpy | GPS + Galileo + BeiDou, 32 sats @ 5 Hz | 0.024 | 0.021 | 0.014 |
| SY1__2607112345.gpy | GPS + Galileo + BeiDou, 24 sats @ 5 Hz | 0.025 | 0.023 | 0.014 |
| S3__2607120443.gpy | GPS + Galileo, 24 sats @ 5 Hz | 0.026 | 0.023 | 0.014 |
| SY2_2607112345.gpy | GPS + Galileo + BeiDou, 32 sats @ 10 Hz | 0.045 | 0.043 | 0.024 |
28 satellites @ 5 Hz (green) performed best and 32 satellites @ 10 Hz (orange) performed worst.
5 Hz consistently outperformed 10 Hz from the perspective of median values.
5 Hz consistently outperformed 10 Hz from the perspective of standard deviation.
Observations
- S3 was intended to be GPS + Galileo + BeiDou B1C
- Somehow it was switched to GPS + Galileo
- SY2 dropped lots of frames, and saw quite a few “mini spikes”
- Perhaps due to the M10 using “balanced” power mode
- 28 satellites performed best @ 5 Hz
- 32 sats was next best, and 24 sats slightly worse than 32
- D1 and D2 produced the most accurate results during phase 3
- D1 and D2 produced the most accurate results during phase 2
Conclusions
- Max of 28 sats provides the best accuracy, but drops frames @ 80 MHz
- Order of accuracy @ 5 Hz is 28 sats, 32 sats, 24 sats
- The cause of dropped frames at 5 Hz may be the ESP32 @ 80 MHz
- Earlier tests coped with higher rates without dropping frames
- Results suggest the accuracy of 10 Hz is significantly worse than 5 Hz
- The accuracy of SY2 @ 10 Hz was much worse than 5 Hz devices
- D1 and D2 may perform better than the S devices
- Phase 5 testing will ensure they use different configurations