UWB vs BLE Channel Sounding: Airtime Limits at Warehouse Scale

What changed now that Bluetooth 6 channel sounding hardware ships?

The Bluetooth SIG published the 6.0 core spec in September 2024 with Channel Sounding, a ranging method that combines phase measurements across narrow 2.4 GHz channels with round-trip timing. Through 2025 the silicon arrived: Nordic's nRF54 series and NXP's KW47 and MCX W72 ship with Channel Sounding support, and tag vendors are building on them. So the comparison is no longer slideware against slideware.

Vendors quote sub-meter accuracy in clean line of sight. Early reports from racked warehouses land closer to 1 to 3 meters, and the physics explains why: an 80 MHz band resolves multipath to about 12.5 nanoseconds, roughly 3.7 meters of path difference, while a 500 MHz UWB channel resolves about 2 nanoseconds, around 60 centimeters. That accuracy gap has been argued to death. The constraint that gets less attention is airtime: how much radio time one position fix consumes, and what that does to tag count, update rate, and battery life at warehouse scale. Accuracy is not the only budget.

Why does one position fix cost more airtime on BLE channel sounding?

Channel Sounding is a point-to-point procedure. An initiator and a reflector step across up to 72 narrow channels exchanging tones and packets, and a full procedure occupies tens of milliseconds. The result is one distance to one anchor. A 2D position needs ranges to at least three anchors, 3D needs four, so one fix means three or four full procedures run back to back from the tag's point of view.

UWB supports two-way ranging too, and a DS-TWR exchange on a Qorvo DW3000 class radio completes in a millisecond or two. But the mode that matters for warehouse RTLS is uplink TDoA: the tag transmits a single blink shorter than a millisecond, every synchronized anchor in earshot timestamps it, and the location engine solves position from the time differences. The tag never waits for a reply, and adding anchors improves the fix without adding airtime. The Bluetooth 6 spec defines no equivalent broadcast mode. One blink beats four conversations.

Run the capacity math for a 5,000-tag warehouse

Take a mid-size facility: 5,000 tagged assets, one position fix every 10 seconds. That is 500 fixes per second.

• UWB TDoA: 500 blinks per second at 0.2 to 1 ms each consumes 10 to 50 percent of a single channel's airtime. Slotted access handles that comfortably, and a second channel near 8 GHz is sitting free if you need headroom.
• BLE Channel Sounding: 500 fixes times three procedures times roughly 30 ms demands about 45 seconds of radio time per second of wall clock. You can shard zones and run pairs concurrently, but you start 45x over budget, before Wi-Fi interference forces retries.

The escape valves are all painful: shrink the fleet per zone, stretch updates from seconds to minutes, or buy and mount more initiator hardware. Battery follows the same arithmetic. A tag that transmits under a millisecond per fix runs multi-year on a coin cell; a tag that holds four tens-of-milliseconds conversations per fix does not. Airtime sets the ceiling, not the datasheet.

Does 2.4 GHz coexistence break channel sounding at peak hours?

Channel Sounding lives in the same 83.5 MHz of 2.4 GHz spectrum as everything else in the building. Warehouse Wi-Fi on channels 1, 6, and 11 already blankets most of that band, and the floor adds handheld scanners, voice-picking headsets, label printers, and legacy BLE beacons. A Channel Sounding procedure hops across the band by design, so it crosses active Wi-Fi channels on nearly every run. Corrupted steps mean retries, longer procedures, and noisier distance estimates exactly when traffic peaks, which in a warehouse is also when you most need positions.

UWB operates in channels 5 and 9, near 6.5 and 8 GHz, under low-power regulatory limits. In industrial buildings that spectrum is close to empty, and 802.15.4z preambles are designed for detection at very low received power. Interference planning for a UWB deployment is usually a non-event; for a 2.4 GHz ranging system it is a standing work item. Your busiest hour is channel sounding's worst hour.

When is BLE channel sounding still the right call?

Channel Sounding earns its place wherever the airtime math is not binding:

• Bay-level tracking at low rates. If 1 to 3 meter accuracy and a fix every few minutes answers the question, which aisle is the pallet in, the airtime is affordable and the tag carries a single 2.4 GHz radio.
• Access and proximity. Channel Sounding was pushed forward by secure car access and digital key work, and door-level logic like is this badge within 2 meters sits squarely in its lane.
• Phone interaction. UWB has shipped in iPhones since the iPhone 11 and in flagship Androids; Channel Sounding phone support in 2026 is still thin. Verify per device, not per spec sheet.

One catch: Channel Sounding needs new silicon such as nRF54 or KW47, so an installed BLE beacon fleet does not upgrade by firmware. Many real deployments will run both radios, Channel Sounding for bulk slow-moving inventory, UWB for forklifts, AMRs, and high-value assets at 1 Hz and up. Pick per asset class, not per religion.

The radio is the cheap part. The ranging stack is not.

Whichever radio wins your bakeoff, the chipset is the smallest line item. The hard engineering sits above it: synchronizing anchors across a million square feet, scheduling thousands of tags without collisions, rejecting multipath off steel racking, and fusing RF ranges with inertial or vision data when line of sight dies. Teams routinely spend 18 to 24 months rebuilding that layer, then discover the methods they converged on are already patented by someone else.

Position Imaging licenses hundreds of granted US patents in radio-frequency ranging, real-time positioning, computer vision, and machine learning, IP that Apple, Bosch, and other major firms cite in their own filings. Licensing that stack instead of rebuilding it puts your team on the warehouse workflow rather than on re-deriving ranging math, and it works on both sides of the UWB versus Channel Sounding bet because the scheduling, sync, and fusion problems are radio-agnostic. The portfolio covers the parts this article just told you are hard. Build the product, license the physics.