I'm building a wall-mounted single-anchor UWB system (doorway exit-detection use case) that needs to determine whether a tag is inside or outside a doorway using AoA — including doors that are already open, so I can't gate on a door contact sensor.
Background: I was running Qorvo QM33120W/DW3000 (2-antenna PDoA) and hit a hard architectural wall; with the anchor mounted ~7ft up facing straight down, walking/circling under it produces false "outside" angle commits even on clean, strong signal.
My plan was to have the UWB antennas face directly down toward the ground so that I can capture both positive and negative angles (inside and outside) signals to determine when a tag moves from inside to outside (exit detection).
Currently with the Qorvo devices, if I hold the tag in the optimal straight forward/antenna on top, everything is fine but when moving/rotating the tag (DW3000), I start seeing different AoA values even when standing in the same place.
Sign convention issue: My expectation is a consistent convention — positive angle while inside, negative once the tag crosses to outside. In practice, while unambiguously still inside, I'm seeing both positive and negative angles, correlated with the tag being rotated and moved in random directions/speeds — not with actual crossing of the doorway. This is the specific symptom I'm trying to design around before committing to new hardware.
I'm now evaluating a move to NXP silicon and would appreciate community input:
- Qorvo 2D AoA boards use an RF switch to capture the UWB PDoA between the two antennas. Will switching to the NXP SR150 dual rx chain solve my AoA jumping issues?
- For a single overhead-mounted anchor with a tag moving through a doorway below it, is SR150's 2D/software-assisted-3-antenna-3D AoA mode sufficient to resolve front/back ambiguity, or does this really need SR250's 3 simultaneous RX chains for a true second baseline?
- Any recommended antenna geometry/mounting orientation for this specific top-down overhead use case?
- Any real-world experience with SR150's antenna-switched 3rd-antenna mode vs. SR250 in ambiguity-prone geometries like this?
Thanks.
Hello,
Hope you are doing well. For a single overhead-mounted anchor doing inside/outside doorway detection with a freely rotating tag, the Trimension SR250 is the right platform and is our recommended UWB product for IoT and Industrial anchor designs.
The core reason for this recommendation is hardware architecture. The SR250 integrates 3 simultaneous RX paths for one-shot 3D AoA with no external RF switches required, delivering both azimuth and elevation from a single ranging frame.
The SR250 also brings additional capabilities that are valuable for this use case:
- 360° AoA support with antenna diversity up to 9 antennas
- On-chip UWB radar (OCPD) for presence detection, a useful complementary layer alongside ranging-based crossing detection
- FiRa 4.0 and Aliro 1.0 ready, ensuring interoperability with the broader UWB ecosystem
Where to Start
- Dev kit: SR250 Development Board, Arduino-compatible, integrated PCB antennas, plug-and-play demos for ranging, AoA, and radar out of the box
- Software: SR250 UWBIOT for Zephyr OS
- Partner modules & kits: TrueSense (ETNA TS 250 DevKit), MobileKnowledge (MK UWB Kit Mobile edition 2.0), Amotech (SR250 Integrated 3D Antenna Module), all available via the NXP Partner Marketplace
Hope this helps.
Best Regards,
Ricardo
Thanks Ricardo, appreciate the detailed response.
Update: I've ordered both the Murata Type2BP (SR150) dev board and the NXP SR250UWBSHIELD dev board. Unfortunately, the SR250UWBSHIELD is showing a 14-week backorder on my end, so I'll be starting bench work on SR150 in the meantime and moving to SR250 once it arrives.
A few follow-ups while I wait:
1. SR150 vs. SR250 for my specific application: given that I only need azimuth (inside/outside), not elevation, what benefit would I actually see moving from SR150 to SR250 beyond the architecture difference (true 3 simultaneous RX chains vs. SR150's 2 native chains + switched 3rd antenna)? Is the accuracy/stability improvement meaningful enough for a single-axis use case to justify the wait, or would SR150 likely get me there on its own once I have my mounting/multipath mitigations in place?
2. RF switching and tag rotation: one specific symptom I'm trying to root-cause, on the Qorvo hardware, if I stand in exactly the same spot and just rotate the tag on the X plane (no position change at all), the AoA value jumps around noticeably even appearing to flip polarity. Qorvo's 2D AoA uses an RF switch to time-multiplex the two antennas rather than sampling them simultaneously. Do you think that switching architecture is a likely contributor to this rotation-correlated instability, or is this more consistent with something else (tag radiation pattern/polarization sensitivity as it rotates, multipath, etc.)? Trying to understand whether SR150/SR250's simultaneous RX sampling would be expected to resolve this specific symptom or if it's a separate issue I'd need to address regardless of chip.
3. 2D vs. 3D AoA for my use case: since I only need to know inside vs. outside (azimuth), and elevation isn't meaningful for my application, is there any accuracy or reliability benefit to running full 3D AoA anyway, or would I be better off configuring azimuth-only and ignoring elevation? Specifically wondering if the elevation measurement helps discriminate reflected/NLoS signals from valid ones even if I don't use elevation for the actual inside/outside decision.
4. Power draw, 2D vs. 3D: this will be a battery-powered install. Is there a meaningful current draw difference between running 2D-only vs. full 3D AoA on SR250, given it's 3 simultaneous RX chains either way? Or is the power cost basically fixed once all 3 chains are active regardless of which axes I use in software?
5. Multipath mitigation for my specific install: the beacon will be wall-mounted above a doorway with a nearby glass door and a ceiling roughly 5ft above the unit. I've been seeing what looks like reflection-driven angle instability on my current Qorvo hardware (correlates with tag rotation, worse in higher-ceiling rooms, worse with the glass door closed). Any recommended mounting practices, antenna beamwidth/pattern choices, or firmware-side filtering (FOM/NLoS thresholds) specifically for suppressing near-field reflections off glass and adjacent walls? Or are there any features on either the NXP SR150 or SR250 that would suppress this issue?
6. Radar-based motion gating for battery savings: I'd like to use SR250's on-chip radar (OCPD) purely as a low-power wake trigger: stay in radar mode, only spin up full ranging/AoA once motion is detected. With the anchor mounted above a ~2m-high doorway, antennas facing straight down, roughly what motion detection range/coverage should I expect directly below the unit? Trying to size the radar's effective "wake zone" against the doorway footprint. I would want to detect movement as the person is walking toward and away from the doorway.
Thanks again for the help.