Question 1

Why does Skybrush show TIMESYNC on drones with the MA-RTK-AIR-626 when RTK is fixed?

Accepted Answer

GPS time is the GPS week number plus the time of week. The MA-RTK-AIR-626 sends only the UBX NAV-PVT message, and ArduPilot's u-blox driver reads the week only from NAV-SOL or NAV-TIMEGPS. The week stays 0, ArduPilot treats that as no GPS time, and Skybrush Live raises TIMESYNC. Position and RTK are not affected.

Question 2

Why not take the whole time from the UTC date in NAV-PVT?

Accepted Answer

ArduPilot keeps GPS time as week plus milliseconds of week, and the receiver's time of week is the precise, epoch-aligned part. Converting UTC to GPS time also depends on the leap-second count. The Robofusion fix keeps the receiver's time of week and uses the date only to fill in the missing week, rounded to the nearest week so a leap-second difference cannot shift it.

Question 3

Does the fix change anything for genuine u-blox receivers?

Accepted Answer

No. The date-derived week is used only until the receiver supplies a week through NAV-SOL or NAV-TIMEGPS. A genuine u-blox receiver sends one of those, and from then on the driver behaves exactly as the official firmware does.

Question 4

How was the fix tested?

Accepted Answer

A 1,305-case test of the week formula covering 2020 to 2099, the seconds around GPS week rollover and receivers whose leap-second count is off by up to 2 s; ArduPilot SITL with a simulated receiver that sends only NAV-PVT, with the unpatched driver as a negative control; and a customer's Skybrush show fleet on MicoAir H743 V2 and H743 V2 AIO 45A flight controllers.

Question 5

Which firmware target does the MicoAir H743 V2 AIO 45A use?

Accepted Answer

MicoAir743v2, board ID 1179, the same target as the standalone H743 V2. The MicoAir743-AIO target, board ID 1176, is only for the older H743 AIO 35A, and the bootloader on a 45A board refuses it.

One post tagged with "skybrush"

Skybrush TIMESYNC and the MA-RTK-AIR-626: tracing a missing GPS week