How to Choose a 2D LiDAR Scanner
A 2D LiDAR scanner sweeps a laser beam across a plane and measures the time of flight of the reflected light, producing a distance profile of everything in its field of view. One compact unit can replace a whole row of photoelectric sensors: it sees a 270° or wider arc, distinguishes near from far, and can trigger different responses at different distances.
The catch is that "2D LiDAR" covers two quite different kinds of product. One kind outputs raw point-cloud data over Ethernet for software to process; the other converts detection into plain switching signals a PLC can read directly. Choosing between them — before arguing about range or resolution — is the single most important decision.
This guide takes you through the selection in six steps, using DAIDISIKE's DLD and SDLD series as worked examples: the DLD05A3 and DLD20A5 discrete-output obstacle-avoidance scanners, the DLD30T, DLD-50D, and DLD-100D Ethernet measuring scanners, the DLD50T8 that offers both interfaces, and the SDLD-05A AGV scanner.
What Is a 2D LiDAR Scanner and When Do You Need One?
A 2D (single-line) LiDAR measures distances across one scanning plane, typically using a 905 nm Class 1 eye-safe laser. Mounted horizontally on a vehicle it detects obstacles ahead and to the sides; mounted at a doorway or machine perimeter it monitors a whole area that would otherwise need many point sensors.
Reach for a 2D LiDAR when you need any of: wide-angle coverage from a single device, distance-dependent behavior (warn at 3 m, slow at 1.5 m, stop at 0.5 m), zones that change shape with vehicle state, or actual distance measurement for positioning and profiling. If you only need to know "is something at this one spot," a photoelectric sensor or laser distance sensor is simpler and cheaper.
One caution up front: a general-purpose 2D LiDAR is not automatically a certified safety laser scanner. If your risk assessment calls for a safety-rated device, that requirement drives the selection before any performance spec — see the final section.
Step 1: Discrete Switching Outputs or Point-Cloud Data?
Decide first how the scanner will talk to your system, because it splits the market in two.
Discrete-output scanners behave electrically like a photoelectric sensor. The DAIDISIKE DLD05A3 and DLD20A5 output three switching signals (NPN or PNP, rated DC30V/50 mA max) plus one status signal — no fieldbus, no protocol stack, no SDK. You configure detection zones once over USB; afterwards the scanner runs standalone and drives PLC inputs directly. This is the right choice for AGV retrofits and machine builders without a software team.
Point-cloud scanners output measurement data over Ethernet for navigation, mapping, and profiling software. The DLD30T (30 m), DLD-50D (50 m), and DLD-100D (100 m) all stream data over 100BASE-TX Ethernet. Choose these when a computer or robot controller will consume the data.
Two products bridge the gap. The DLD50T8N/P provides both Ethernet data output and a discrete NPN/PNP output — useful when a navigation computer and a hard-wired interlock must share one sensor (note its discrete output is a protective/pre-warning interlock signal, not a certified OSSD safety output). The SDLD-05A AGV scanner combines an Ethernet (RJ45, UDP) interface with NPN digital I/O — up to 4 inputs and 4 outputs — plus a USB Type-C configuration port.
Why the Same Scanner Has Two Very Different Range Figures
Nothing in a LiDAR datasheet causes more mis-specification than range, because the same unit is legitimately quoted with two numbers that differ by a factor of four or more — and they are not measuring the same thing.
Detection depends on how much of the emitted light comes back. Remission is that fraction: a white matte board returns roughly 90 percent, cardboard around 20, and dark clothing as little as 1.8 percent. A safety protective range must be guaranteed against the worst case a person can present, so it is specified at 1.8 percent remission. A measuring or obstacle-detection range is quoted at a far higher remission because it only has to work on cooperative targets.
The practical consequence: comparing a 20 m measuring range against a 5 m protective field is comparing two different measurements, and a scanner advertised at 30 m may guarantee only a few metres of protective field. When a range figure appears without a remission value attached, treat it as a marketing number until the datasheet says otherwise.
The same physics quietly reduces performance in service. A target that was white when the line was commissioned and is now grey with dust returns less light, and the usable range falls with it — one of the reasons a scanner that worked at handover can start missing objects months later without anything having been changed.
Step 2: How Much Range Do You Really Need?
LiDAR range depends on target reflectivity (remission), so always read the spec at 10% remission — roughly a dark work jacket — not just the headline figure at 90% (a white wall). The DLD-50D is rated 50 m at 90% remission but a guaranteed 20 m at 10%; the DLD30T detects 30 m at 90% but 10 m at 10%; the DLD-100D reaches 100 m at 90% and 40 m at 10%. Size your application against the low-remission number.
For obstacle avoidance, work backwards from speed: the scanner must see far enough that the vehicle can warn, slow, and stop within the detected distance. The DLD05A3 (0.05–5 m) is positioned for low-speed carts and close-quarters protection; the DLD20A5 (0.05–20 m) covers faster vehicles and longer aisles — the two share the same 50 × 50 × 76 mm housing, so an upgrade does not force a mechanical redesign. The SDLD-05A detects 0.1–14 m at 90% reflectivity and 0.1–8 m at 10%.
Do not ignore minimum range. A scanner blind below 0.3 m can miss an obstacle a docking robot is about to touch; the DLD05A3 and DLD20A5 start detecting at 0.05 m, keeping the near-field blind zone small.
Step 3: What Field of View and Blind Sector Can You Live With?
Field of view determines mounting position. A 270° scanner mounted on a vehicle corner covers two full sides at once; that is the geometry of the DLD05A3, DLD20A5, DLD30T, and SDLD-05A. The DLD-50D and DLD50T8 scan 280°, and the DLD-100D family includes a 360° variant (DLD-100DC) for mast or overhead mounting where all-round coverage matters.
Every non-360° scanner has a blind sector behind it — the DLD-50D, for instance, has an 80° blind sector spanning 320°–0°–40°. Point the blind sector at the vehicle body or the wall, and check that brackets, cables, and chassis edges do not intrude into the active arc.
Also verify the scan plane itself: a 2D scanner sees only one horizontal slice. Obstacles below the plane (forks, pallets on the floor) or above it (overhanging racks) are invisible, so mounting height is part of the safety concept, not an afterthought.
Step 4: What Angular Resolution and Scan Frequency Do You Need?
Angular resolution sets how far apart the measurement points sit on the target. At 0.1° resolution, points at 5 m range are spaced under 9 mm apart — fine enough to catch a chair leg. At 0.3°, spacing triples. The DLD05A3 and DLD20A5 offer selectable 0.1°/0.3° resolution; the DLD30T adjusts down to 0.08°; the DLD-50DH offers 0.08°/0.16°/0.32°.
Scan frequency sets how often the picture refreshes: 15/30 Hz on the DLD05A3 and DLD20A5, 10–30 Hz adjustable on the DLD30T, and up to 50 Hz or 100 Hz on the DLD-50D and DLD-100D variants (the 100 Hz DLD-50DP trades down to fixed 0.32° resolution).
The two trade off against each other on most platforms. Rule of thumb: prioritize resolution for detecting small or thin obstacles; prioritize frequency for fast-moving vehicles where reaction latency dominates. Check the resolution-frequency combination table for the exact model before committing.
An object narrower than this gap can sit between two measurements and go unreported at that distance. This is geometry, not a product limitation — it applies to every scanning LiDAR. It is also why a scanner that resolves a pallet at 2 m may miss the same pallet leg at 20 m.
Step 5: How Will You Configure Zones and Switch Between Them?
For obstacle avoidance, zone architecture matters more than raw specs. The DLD05A3 and DLD20A5 store 16 zone sets, each defining 3 nested detection zones mapped to the 3 switching outputs — the classic warning / slow-down / stop sequence, hard-wired. Zone sets are configured once over Micro-USB and switched in operation as vehicle conditions change: straight travel, turning, docking.
The SDLD-05A goes further for AGV work: its 16 zone groups each hold 3 independently editable sub-zones drawn as polygons (up to 15 vertices), fans, or rectangles, and the active group is selected live by a 4-bit digital input code (IN1–IN4) from the vehicle controller — no reprogramming. Its configuration software also provides hold-time filtering (default 200 ms) and object-size filtering to suppress nuisance trips, on top of the 50 ms hardware response time.
Sketch your zone geometry per driving state before choosing hardware, then confirm the zone shapes, vertex counts, and switching inputs the scanner supports. Running out of zone sets mid-project is a painful discovery.
Step 6: Power, Environment, and Mounting Constraints
On mobile platforms, size and power budget are hard limits. The DLD05A3 measures 50 × 50 × 76 mm at 150 g on DC9–28V; the DLD-50D draws just 2.5 W from DC9–36V; the SDLD-05A weighs 374 g including cable and draws ≤2.5 W. All fit small battery-powered chassis.
Match the IP rating to the environment: IP65 on the DLD05A3/DLD20A5 handles dust and light moisture; IP67 on the DLD30T, DLD-50D, DLD50T8, and DLD-100D suits harsher or outdoor-adjacent duty; the SDLD-05A is IP54, for indoor vehicles. Check ambient light immunity for sunlit sites — the DLD-50D withstands over 100,000 lux and the DLD-100D over 80,000 lux, and the DLD-100D adds multi-echo processing to help in dust, rain, and fog.
Operating temperature is typically −10 to +55 °C on the measuring models. All series discussed here use 905 nm Class 1 eye-safe lasers, so no laser-safety controls are needed around personnel.
What Degrades a LiDAR Reading in Service
A scanner measures light that comes back, so anything between the unit and the target competes with the echo it is trying to time.
Airborne particulate is the main one. Fog, steam, welding smoke and heavy dust scatter the outgoing pulse and return a partial echo from the air itself, which the scanner can read as a phantom object close in, or which can mask a real object further out. Rain and snow do the same intermittently, which is why an outdoor installation needs a scanner rated for it and a field configuration with tolerance for isolated returns.
Direct sunlight into the receiver raises the noise floor and shortens usable range, so scanners are usually specified with an ambient light immunity figure. Mounting the unit so that the sun cannot sit in its field at any point in the day is cheaper than compensating for it later.
Mutual interference matters as soon as there is more than one scanner. Two units whose fields overlap can read each other’s pulses unless the series implements interference rejection. Confirm that capability before designing a cell with several scanners rather than discovering it during commissioning.
Finally the window itself. A contaminated or scratched front window attenuates both the outgoing pulse and the echo. Units intended for safety duty monitor window contamination and report it as a fault — a feature worth having, because the alternative is a scanner that degrades silently.
Mounting Geometry Decides Whether the Scan Plane Is Useful
A 2D scanner sees one plane. Everything above and below that plane is invisible to it, and this single fact causes more field problems than any electrical parameter.
Mounted horizontally at ankle height on an AGV, the scan plane detects a person standing in the path but not a shelf overhanging at chest height, and not a fork protruding at knee height from a stationary pallet truck. Mounted higher it sees the obstruction and misses the feet. Neither is wrong; both need to be a deliberate decision recorded in the risk assessment rather than a consequence of where the bracket fitted.
On mobile platforms the plane also moves with the vehicle. Ramps, thresholds and floor joints tilt the plane into the ground or up into free space, producing either spurious stops or a gap in coverage exactly where the vehicle is least stable. Check the extremes of the route, not the flat section.
Two more geometric checks are worth running before drilling anything. Reflective floors and glossy machine panels inside the scan plane can produce mirror returns that appear as objects where none exist. And the blind sector behind the scan arc must face a wall, a structure or another guarded approach — an unguarded blind sector is simply an unguarded approach.
Is a 2D LiDAR a Certified Safety Laser Scanner?
Not automatically — and this is the most common specification mistake with LiDAR. A safety laser scanner in the regulatory sense is designed and certified as electro-sensitive protective equipment (e.g., per IEC 61496), with self-monitored safety outputs. A general-purpose 2D LiDAR, however capable, provides protective and pre-warning functions without that certification.
The DAIDISIKE DLD-series scanners described here are obstacle-avoidance and measuring devices; the DLD50T8's discrete output, for example, is explicitly a protective interlock signal rather than a certified OSSD output. If your risk assessment requires a certified safety device, tell us the application and required performance level, and we will confirm what is available — certification documents are available on request.
For many real installations the right architecture is layered: a certified safeguard (such as a safety light curtain or safety-rated scanner) covers the point of legal compliance, while general-purpose LiDAR adds early warning, slow-down zones, and navigation data that make the vehicle or machine both safer and more productive.