How Is In-Building Two-Way Radio Coverage Measured and Passed?

In-building radio coverage is measured with a grid survey: each floor is divided into roughly 20 equal cells, and the signal is checked at a point near the centre of each cell against two criteria — a received signal level, measured in dBm on a calibrated analyser, and Delivered Audio Quality (DAQ), a 1-to-5 listening score. A floor passes when a set percentage of cells meet both thresholds, with adjacent failures treated more harshly than scattered ones. The working numbers most integrators design and test to are −95 dBm and DAQ 3.0–3.4, with 90–95% of general-area cells passing and 99% in critical areas such as stairwells and lift lobbies.

I run this test constantly, and almost everyone asking the question has read a US fire-code answer without realising those exact figures are American. Here is what the survey actually involves and which numbers carry over to a Singapore building.

What is a grid survey and how many points get tested?

The floor plate is overlaid with a grid — the common convention is about 20 cells per floor, sized so each covers a comparable area. A test point is fixed near the centre of each cell, and once chosen you are not allowed to hunt around inside the cell for a stronger spot; that rule is what stops a survey being gamed. At each point the surveyor keys up a portable on the system frequency and records two things at once: the received power on a spectrum analyser or calibrated test set, and the audio intelligibility a listener at the far end scores by ear.

For a re-test on an existing system, a shortcut is common: rather than re-walking all 20 cells, the survey re-checks the two or three cells that scored worst last time, on the logic that if the weakest points still pass, the floor still passes.

What dBm level and DAQ score count as a pass?

Two independent thresholds have to be met at each point:

• Received signal level: −95 dBm is the accepted floor for reliable UHF (400–470 MHz) DMR or analogue voice to a portable. Below that you are eating into fade margin and the audio starts breaking up as the user moves or turns.

• Delivered Audio Quality (DAQ): a 1–5 subjective scale. DAQ 3.0 means “speech understandable with slight effort, occasional repetition needed”; DAQ 3.4 is the stricter public-safety bar. Most codes set the pass at 3.0, mission-critical work at 3.4.

• Area pass rate: 90% of general-area cells is the common general-coverage grade; serious integrators design to 95%. Critical areas — exit stairs, fire lift lobbies, pump rooms, the fire command centre — are held to 99%.

• Adjacency rule: two non-adjacent cells may fail on a floor and it still passes; two adjacent failing cells fail the whole floor, because adjacent dead cells mean a real dead zone a user can walk into, not measurement scatter.

The −95 dBm and DAQ numbers are worth separating in your head. dBm tells you the signal is physically present; DAQ tells you it is actually usable. A point can show −92 dBm but still fail DAQ if noise, desense or multipath is corrupting the audio — which is why any honest acceptance test scores both and does not accept a dBm reading alone.

Do these thresholds apply in Singapore, or are they US numbers?

The pass percentages (90% general, 99% critical), the DAQ 3.0/3.4 grades and the −95 dBm figure originate in the US International Fire Code Section 510 and NFPA 1225. Singapore has not adopted an equivalent numeric public-safety in-building radio coverage code — the local framework is IMDA's COPIF for provisioning and SCDF's Fire Code for the hard-wired Emergency Voice Communication System, neither of which fixes a two-way radio coverage percentage. I cover that split in detail in our SCDF/IMDA in-building radio coverage explainer.

That does not make the numbers irrelevant here. They are simply good engineering, and in the absence of a Singapore statutory grade they are the benchmark competent integrators design and commission to anyway. When we hand over a system in Singapore, we test it to 95% area coverage at −95 dBm and DAQ 3.4, because that is the level at which a security or facilities team's radios genuinely work everywhere they need to — not because a local code forces the number.

How is the survey actually carried out on site?

The uplink matters as much as the downlink, and a proper survey tests both. Downlink is the repeater or DAS talking to the portable; uplink is the 4 W portable talking back. Because the portable is far weaker than the 40 W repeater, the uplink is frequently the limiting path — a point can receive fine but fail to be heard. A survey that only measures the signal reaching the handset, and never checks the talk-back, is only half a test. The surveyor walks the grid with a portable and a reference receiver at the repeater, logs both directions per cell, and marks any cell that fails either.

Related questions

What is DAQ 3.4 versus DAQ 3.0? Both are points on the 1–5 Delivered Audio Quality scale. DAQ 3.0 is “understandable with occasional repetition”; DAQ 3.4 is a stricter grade requiring speech to be understood with rarely any repetition, and is the level used for critical public-safety coverage.

How many grid cells can fail and still pass? On a ~20-cell floor, typically two non-adjacent cells may fail and the floor still passes (around 90–95% pass rate). Two adjacent failing cells fail the whole floor, because that indicates a contiguous dead zone.

Is a dBm reading enough, or do I need a DAQ score too? Both. dBm confirms the signal is present at the required strength; DAQ confirms the audio is actually intelligible. Noise, desense or multipath can wreck DAQ even where dBm looks acceptable, so a credible acceptance test records both at every point.

How Suneast handles this

Suneast has designed and commissioned in-building UHF coverage in Singapore for over 40 years, and we hand over on measured results, not assurances. Every system is grid-surveyed on completion — downlink and uplink at each cell, dBm and DAQ logged together, adjacency rule applied — and we design to 95% area coverage at −95 dBm and DAQ 3.4, with the remaining margin usually the lift cars and shielded rooms we call out explicitly rather than quietly exclude. If your radios drop out in the basement, the stairwell or the back-of-house, our network enhancement service starts with the same survey we use to prove the fix — so the pass or fail is a number on a sheet, not an opinion.