Timing Without the Roof

How neutral-host and private 5G networks achieve TDD-grade sync when GNSS isn’t practical

Neutral host and private 5G deployments don’t fail because the radio is hard. They fail because the building is hard.

Anyone who has delivered 5G inside a stadium, hospital, high-rise, shopping centre, or transit hub knows the pattern: radios go in quickly, fibre is available, switch ports are plentiful… and then timing becomes the blocker. Roof access is delayed for weeks. Landlord approvals bounce between stakeholders. The only feasible antenna route is too long, too lossy, or too risky to implement.

Yet the network still has to meet the realities of 5G timing. In TDD deployments, phase/time alignment is what keeps neighbouring cells from stepping on each other. Industry guidance commonly targets sub-microsecond to low-microsecond end-to-end time error, and a widely referenced figure is ±1.5 µs to a common time reference for TDD phase/time alignment.

So what do you do when GNSS is unavailable at the exact places neutral host and private 5G are most valuable?

The core shift: “Timing distribution” rather than “GNSS everywhere”

Traditional thinking puts a GNSS antenna at every site. In modern in-building neutral host, that approach often creates the most cost, the most schedule risk, and the most points of failure.

A more scalable approach is to treat precise time like any other critical utility:

1. Acquire a high-quality time reference where it’s easiest (headend, MMR, comms room, rooftop of one accessible building, or a nearby aggregation site).

2. Distribute time to baseband and radios using deterministic methods.

3. Engineer for holdover and graceful degradation.

This is where Timebeat deployments typically win: not by pretending the building will magically allow rooftop GNSS, but by designing architectures that deliver “TDD-grade timing” even on real-world networks.

Three deployment patterns that work in practice

Pattern A: GNSS-fed “Open Time Appliance” at the headend (minimal roof work)

If you can get any GNSS feed—even to a single location—concentrate it.

• A single GNSS-fed timing appliance becomes the authoritative reference (PTP Grandmaster function).

• PTP is then distributed over the in-building transport to BBUs/O-DUs and, where needed, onward to RUs.

This drastically reduces rooftop complexity: one antenna, one route, one set of approvals.

Key design notes:

• Place the timing appliance close to the transport core (short, stable path).

• Use redundant power and network.

• Treat it as critical infrastructure: monitoring, alarms, and documented change control.

Pattern B: PTP over the network (when you really can’t do GNSS)

Open RAN and telecom profiles recognise multiple PTP transport approaches, including profiles where the network provides partial timing support (e.g., G.8275.2), versus full timing support with boundary/transparent clocks (G.8275.1 concepts).

In neutral host/private 5G, the transport is often a “general network” rather than a timing-purpose-built one. That’s the hard part: packet delay variation (PDV), jitter, and asymmetry.

Timebeat’s approach in these environments is not “hope the network behaves.” It’s to apply techniques that make packet timing usable:

• PTP packet filtering and intelligent selection (so outliers don’t corrupt the clock servo)

• Robust servo behaviour under PDV

• Asymmetry mitigation strategies (measurement, topology choices, and operational guardrails)

• Traffic engineering (QoS marking, VLAN isolation, prioritised queuing, controlled contention points)

The practical outcome: baseband units can sync to a PTP source without requiring a local GNSS antenna at every radio location—exactly what in-building neutral host needs.

Pattern C: Hybrid timing with holdover (designed to degrade gracefully)

Even when you can use GNSS, a hybrid architecture is what makes it operationally strong:

• Primary: GNSS (or other absolute reference) at a reachable point

• Secondary: PTP distribution from a stable master

• Tertiary: disciplined holdover in the timing appliance, so transient issues don’t become outages

The message to an enterprise buyer is simple: timing should not be the fragile part of your private 5G.

What to tell a stakeholder who says “PTP over a general network can’t work”

They’re not wrong—if you’re using generic assumptions.

It’s true that best-effort switching and mixed traffic introduce jitter and asymmetry, and those can damage a naive PTP implementation.

But it’s also true that neutral host and private 5G aren’t measured in textbook diagrams—they’re measured in outcomes:

• Can you hit the required timing budget for TDD features?

• Can you keep it stable during peak traffic?

• Can you deploy without weeks of rooftop approvals?

A “timing-grade” network is not only achieved by perfect switches. It’s achieved by engineering: filtering, prioritisation, topology discipline, and resilient timing appliances.

The takeaway

If your sales conversation starts with “We need roof access for GNSS,” you’re starting at the highest-friction point.

A better conversation is: 

“We can deliver TDD-grade timing using a set of deployment patterns—from centralised GNSS-fed timing appliances to engineered PTP distribution—designed for in-building realities.”

That’s the difference between a timing strategy that looks good on paper and one that ships in the real world.