How synchronous are two GNSS receivers?

Question

I'm not sure if this is the right place to post this, but I'm dealing with PPS (pulse per second) signals at the end of the day.

I am working with multiple GNSS receivers (from different manufacturers, if that matters). I am wondering if these multiple GNSS receivers are in sync with each other once they all have a fix. Considering that they should all be independently synchronized with the satellites' clocks, they should all be in sync with each other. Formalizing this a bit more, if we consider N distinct receiver+antenna systems, denoted \$(s_1, s_2, ..., s_N)\$, and denote \$t_{\text{PPS}i}^k\$ the UTC timestamp of the k-th PPS signal emitted by \$s_i\$, the following should be satisfied:

for any pair of systems, their PPS do not drift apart: for all pairs \$(i, j) \in [1, ..., N] \times [1, ..., N]\$ and all integers \$k_1, k_2\$, \$t_{\text{PPS}i}^{k_1} - t_{\text{PPS}j}^{k_1} = t_{\text{PPS}i}^{k_2} - t_{\text{PPS}j}^{k_2}\$
additionally, if all systems are configured so that their PPS are emitted at the same offset (e.g. all PPS are emitted on whole UTC seconds), then all the PPS signals are emitted at the same times; formally: for all pairs \$(i, j) \in [1, ..., N] \times [1, ..., N]\$ and all integers \$k\$, \$t_{\text{PPS}i}^{k} = t_{\text{PPS}j}^{k}\$

Is this correct? If not, by how much (roughly) would the synchronization be off between receivers?

Marcus Müller · Accepted Answer · 2025-06-26 17:08:30Z

for any pair of systems, their PPS do not drift apart: for all pairs \$(i, j) \in [1, ..., N] \times [1, ..., N]\$ and all integers \$k_1, k_2\$, \$t_{\text{PPS}i}^{k_1} - t_{\text{PPS}j}^{k_1} = t_{\text{PPS}i}^{k_2} - t_{\text{PPS}j}^{k_2}\$

yes, where drift means a growing difference; in correctly working GPS receivers, the mean drift is 0. Zero-mean jitter still exists.

However, that is a long-term statistic thing, and depending on the architecture of the GPS receiver, the clock from which the PPS is generated might be disciplined by a control loop that is rather slow, so that for seconds to hours, there might be a monotonously moving offset error between receivers.

And since different GPS receivers experience different errors, there can and will be drift on short-to-medium averages.

How much: It's an implementation aspect.

additionally, if all systems are configured so that their PPS are emitted at the same offset (e.g. all PPS are emitted on whole UTC seconds), then all the PPS signals are emitted at the same times; formally: for all pairs (i,j)∈[1,...,N]×[1,...,N] and all integers k, tkPPSi=tkPPSj

No, different models have different systematic offsets between GPS time and PPS, in addition to the jittering.

You would of course be right to assume the accuracy given in the GPS receiver's datasheet. These accuracies can range from nanoseconds to several milliseconds, even under good conditions. That's the economical justification why good GPSDOs are more expensive than cheap GPS receivers.

Thanks for your answer. I realize there may be some jitter associated to the PPS' period. The datasheets at my disposal mention a PPS resolution of roughly 10-100ns. From this I understand that the jitter of the delta between two receiver's PPS should be randomly (normally ?) distributed with a standard deviation of the order of 10-100ns as well. This could be an interesting experiment to test receivers relative to each other. — louisbzk
– louisbzk, Commented Jun 27 at 9:03

Stack Exchange Network

How synchronous are two GNSS receivers?

1 Answer 1

Hot Network Questions

How synchronous are two GNSS receivers?

1 Answer 1

Related

Hot Network Questions