This morning while recording GPS data from NMEA sentence RMC at 10 readings per second from two independent Garmin GPS-19X, the recorded time stamp jumped backwards in time from 122339.0 to 122336.1. This happened on both GPS-19X units that were running. Both GPS receivers had a complete unobstructive view of the sky. They are co-located about 4 feet apart. Both had been running for over 24 hours. How can this 3 second backward time stamp change happen?
A backward jump of about 3 seconds in NMEA GPRMC timestamps, occurring simultaneously on two independent receivers, is a well-documented phenomenon tied to how GPS receivers handle leap seconds or, more commonly, a GPS week rollover or time correction event. Here is what most likely happened:
GPS time correction / receiver re-synchronisation
GPS receivers occasionally perform an internal clock correction when they refine their estimate of GPS time against the satellite signal. If the receiver's internal oscillator has drifted slightly ahead, a snap correction backward is possible. Seeing it on both units simultaneously is actually the strongest clue here: it points to a real event in the GPS signal itself (or a common firmware behaviour), not a hardware fault in either unit.
Leap second or week boundary
Check whether the event occurred close to a GPS week boundary (Saturday night / Sunday morning UTC) or a scheduled leap second. Both can cause receivers to briefly report an incorrect time before settling. At 10 Hz you are sampling fast enough to catch the transition, which most users at 1 Hz would miss entirely.
Satellite constellation event
A sudden change in which satellites are being tracked (for example, a satellite going unhealthy mid-solution, or a new one being acquired) can briefly disturb the receiver's time solution. With both units co-located and both jumping at the same moment, a satellite-side event is the most plausible single cause.
To narrow it down further, it would help to know the exact UTC date and time to check against leap second and GPS week schedules, and whether your NMEA log shows any change in satellite count or fix quality (GSA/GSV sentences) immediately before the jump.
This is an automated reply from the GPS Forums AI assistant. AI can make mistakes, so please double-check anything important.
Prefer no AI replies to your own posts? Manage your AI reply settings
