What Fade Margin Should an In-Building Two-Way Radio Coverage System Be Designed To?

Design an in-building two-way radio system to at least 10 dB of fade margin for a standard commercial building and 15 dB for mission-critical or public-safety use. That means the worst point in the covered 95 percent of the floor should receive about -85 dBm (or -80 dBm for 15 dB), not the -95 dBm design threshold itself. A bare -95 dBm is only where the radio starts to fail, so a system designed exactly to it will pass a static test and then drop calls whenever someone turns a corner, holds the radio differently or a lift door opens.

Why isn't -95 dBm enough on its own?

Because -95 dBm is a threshold, not a target. It is the level at which a DMR portable still decodes voice reliably in the average case, roughly 24 dB above a typical -119 dBm (12 dB SINAD) datasheet sensitivity once body loss, antenna inefficiency and receiver implementation allowance are taken off. Every real reading scatters around the predicted value. If you design so the predicted level equals -95 dBm, roughly half of all positions land below it by definition.

Digital radio makes this worse. As I explain in the digital radio cliff effect article, a DMR receiver does not degrade gracefully. Audio is clean at the threshold and gone a few dB below it, so the margin is what stands between a working call and a dropped one.

What is fade margin, and how do you size it?

Fade margin is the extra signal level you deliver above the threshold so that normal variation does not push a real position below it. You size it from the statistics of indoor signal variation, not by feel.

Indoor UHF shadowing is commonly modelled as log-normal with a standard deviation of roughly 6 to 8 dB. To have a given location meet the threshold 95 percent of the time you need about 1.645 standard deviations of margin:

• 6 dB standard deviation: about 10 dB margin

• 8 dB standard deviation: about 13 dB margin

• 99 percent location probability (2.33 standard deviations): about 14 to 19 dB margin

That is where the 10 dB and 15 dB working figures come from. They are engineering practice, not numbers written into an IMDA or SCDF rule, so treat them as design policy to be agreed with the client and consultant and verify the acceptance criterion in the tender.

How does fade margin change the antenna count?

Directly. Adding 10 dB of margin means each antenna must deliver -85 dBm at its cell edge instead of -95 dBm. Indoors, where path loss climbs steeply with distance and every partition adds several dB at 450 MHz, giving up 10 dB shrinks the usable radius of each antenna substantially, so you place more antennas closer together. That is why a design to -95 dBm is cheaper on paper and why the difference shows up at survey. My per-floor antenna spacing article covers how that radius translates into positions on a plan.

Which path limits the margin, downlink or uplink?

Almost always the uplink. A 4 W portable (36 dBm) transmits 10 dB less than a 40 W repeater (46 dBm), and its signal must survive the same cable, coupler and split losses on the way back to the repeater receiver. A worked example at the worst antenna point, using assumed values: downlink EIRP of +20 dBm per antenna after cable, coupler and split losses, less 65 dB path loss and 3 dB body loss, delivers about -48 dBm at the portable. On the uplink the portable's 36 dBm, plus about 2 dBi at the DAS antenna, less the same 3 dB body loss, 65 dB path loss and roughly 26 dB of chain loss back to the head-end, arrives at the repeater receiver near -56 dBm, about 8 dB below the downlink figure. Margin must therefore be stated for both directions and the smaller one reported. Quoting a downlink-only margin is the most common error I find when reviewing other designs.

Can a design pass a measured test and still lack margin?

Yes. A grid survey with a handful of dBm readings can pass at the average level while individual cells sit only a few dB above threshold. Pair the level test with an audio quality check, since a raised noise floor can cost you the margin even when the dBm reading looks fine. The DAQ 3.0 versus -95 dBm article sets out why both tests are needed.

Related questions

Is 10 dB fade margin enough for a hospital or data centre? Usually not. Where a missed call has safety or operational consequences, design to 15 dB and confirm the client's required coverage grade in writing.

Is a 20 to 30 dB fade margin realistic indoors? Those figures come from point-to-point microwave links where fading is deep and rare. For an in-building UHF DAS, 10 to 15 dB captures the shadowing statistics, and going beyond it mostly adds antennas and cost.

Does fade margin replace the -95 dBm threshold? No. The threshold is the floor, the margin is added on top, and the design target is the sum: about -85 dBm at 10 dB margin.

How Suneast approaches fade margin

Suneast designs in-building UHF radio coverage for Singapore buildings, stating both downlink and uplink budgets and the margin over -95 dBm in every proposal, so that the acceptance test and the design assumption are the same number. Typical projects are designed to 10 dB for commercial buildings and 15 dB where the client requires mission-critical coverage, with the target agreed at design stage rather than discovered at survey. Speak to Suneast for a site assessment and a design that states its margin.