Private 5G Timing: Designing the Sync Fabric for Campus and Industrial Networks

Why private 5G is its own design problem

Private 5G networks — typically deployed by an enterprise to cover a campus, factory, port, mine or other defined geography — share their physical-layer timing requirements with public 5G. Massive MIMO, beamforming, carrier aggregation and inter-cell coordination all demand the same ±1.5 µs Class 6 (or ±1.1 µs Class 6A) time-error budget end to end. The radio doesn't care that it's part of a private network rather than a public one; it cares that the upstream timing source is precise enough to keep its waveform aligned with neighbouring radios.

What's different about private 5G is the operational context. Where a public mobile operator deploys tens of thousands of cell sites across complex transport networks they may or may not control, a private 5G operator typically deploys 5–50 radios across a single physical site, often connected by a transport network the operator owns end to end. This drastically simplifies the design space: the centralised vs distributed PRTC question is less interesting because the geographical footprint is small enough that one centrally located PRTC can serve every radio with a short, controlled PTP path.

Private 5G is also typically operated by a different team to public mobile networks. The operator may be a manufacturing organisation, a port authority, a hospital, a university — none of whom have decades of mobile network operations experience. This means the timing fabric has to be operationally simple in a way that public 5G timing fabrics often aren't, with strong defaults, clear observability, and minimal opportunities to misconfigure.

A reference architecture for private 5G timing

The reference architecture we recommend for most private 5G deployments has three components: a single hardware grandmaster at a central location (typically the same equipment room as the 5G core), a small number of PTP-aware boundary clocks in the transport network connecting that grandmaster to the radio sites, and PTP slave clocks at each radio. For redundancy, the grandmaster is doubled — a primary and backup, configured with G.8275.1 BMCA so failover is automatic.

GNSS feeds the central grandmaster from a roof-mounted antenna with clear sky view. The grandmaster runs G.8275.1 with the standard telecom defaults, distributing time over the operator-owned transport. Every device on the path is PTP-aware and configured for the same profile. The boundary clock chain is short — typically two or three hops between grandmaster and any radio, well within the Class 6 budget.

This is the boring answer, and it is correct for the vast majority of private 5G deployments. Operators who want to do something fancier — multiple GNSS sources, distributed PRTC, exotic hardening — should have a specific operational reason. "Future-proofing" or "because public 5G operators do it" are not specific reasons.

GNSS at the campus: the only really hard question

The interesting design question in most private 5G deployments is GNSS. The grandmaster needs a primary reference, and unless you have White Rabbit fibre back to a national time scale, that primary reference is a GNSS antenna on the roof of the equipment room. The challenge is that private 5G sites are often physically constrained in ways public mobile cell sites aren't.

Industrial sites (ports, mines, factories) frequently have steel structures, cranes, heavy machinery and electromagnetic interference that degrade GNSS reception. Campus environments (universities, hospitals, business parks) often have antenna placement restrictions driven by aesthetics, building structure or planning constraints. Both kinds of site can have intermittent multipath, partial sky obscurity and elevated jamming risk from passing equipment.

Three practical responses to this. First, take antenna placement seriously: a GNSS receiver is only as good as the antenna's view of the sky, and the difference between a poor placement and a good one is often the difference between a working timing fabric and a chronically degraded one. Second, use a multi-band, multi-constellation receiver; the redundancy across frequencies and constellations is what survives multipath and partial obstruction. Third, specify a holdover oscillator that can survive a realistic worst-case GNSS event — for industrial sites, this often means rubidium rather than OCXO, because the oscillator may need to bridge multi-hour environmental disruptions.

Coexistence with TSN and other deterministic networking

Many private 5G deployments coexist with Time-Sensitive Networking (TSN) for industrial control loops, robotics, AGVs and safety-critical systems. TSN uses its own variant of PTP (typically the IEEE 802.1AS profile, which differs from G.8275.1 in transport, message rates and BMCA configuration), and the operator typically wants both networks to share a common time reference so that 5G-delivered events and TSN-delivered events can be correlated.

Practically, this means the central grandmaster needs to deliver both G.8275.1 (to the 5G fronthaul) and 802.1AS (to the TSN network) simultaneously, on different ports. Not all grandmasters support this — many vendors ship products that handle one profile or the other, but not both at the same time. Confirm explicit per-port multi-profile support with the vendor before purchase.

When the same grandmaster does both profiles, the operator gets a single source of truth for time across the converged network. Latency-sensitive industrial control on the TSN side and beamformed 5G coordination on the radio side both reference the same nanosecond. This is the right architecture for any operator running converged 5G + industrial automation; it's also the architecture TimeBeat hardware is designed for.

Observability for private 5G timing

The same observability principles apply to private 5G as to public 5G fronthaul: every grandmaster, every boundary clock and every slave clock should be exporting health metrics to the operator's monitoring stack, with alerts on the metrics that correlate with downstream impact. The difference is operational scale — instead of monitoring tens of thousands of cell sites, the operator is monitoring tens of radios. This makes the per-device observability cheaper to deploy and easier to consume.

A private 5G operator with no prior mobile networks experience should expect to spend roughly as much effort on the observability side as on the timing fabric itself. The fabric is simple; the discipline of watching it is what keeps it healthy. TimeBeat Sync Insight exists specifically for this kind of deployment — it gives operators with no PTP background a working observability surface out of the box.