Laser Distance vs Ultrasonic vs Photoelectric Sensors

Distance Sensors: Laser, Ultrasonic or Photoelectric

Short version. Answer one question first: do you need a number or a yes/no. If you need a number (how many millimetres to the object), that is a laser or ultrasonic sensor with an analogue output. If you only need the fact that an object is there, take an ordinary photoelectric sensor; it costs several times less and is slightly faster. From there, two parameters cut the choice: what surface your target has, and how fast you have to measure. On the Dx35, Dx50 and DL50 the datasheet prints one accuracy figure for the whole range, ±7 to ±10 mm, but the condition beside that figure differs from datasheet to datasheet: 6 to 90% remission on the DT35 and the DT50-2B215252, 90% only on the DT50-P1113, DT50-P1123 and DS50-P1112, and no footnote at all beside the ±7 mm of the DL50-P1123, which measures to a reflector; on the DL100 it is ±2 mm with a footnote about ±4 mm over 150 to 180 mm. The range of a laser, though, drops by a factor of 2.5 to 4 on black. The ultrasonic datasheets have no remission-factor column at all, and their response delay is 150 to 325 ms against 0.83 ms for the fastest laser — two numbers you can only compare by order of magnitude, because the datasheets define them differently, as shown below.

Every figure below comes from manufacturer datasheets we opened in this pass: eight of the nine Sick laser sensors (we could not open the DT50-2B215552 datasheet), two Pepperl+Fuchs ultrasonic sensors and two Sick photoelectric sensors for comparison. Where a number is ours (a division, a percentage), it says so. Our distance sensor section currently holds 11 items, and what follows shows how they actually differ rather than restating a price list.

First Decision: a Number or a Fact

This is the most expensive mistake in selection, and it happens in both directions. One customer pays for a laser sensor with a 16-bit analogue output to catch a box passing on a conveyor. Another tries to build thickness control on a diffuse photoelectric sensor with a potentiometer.

The speed comparison does not favour the measuring devices, which surprises people. The photoelectric GRL18S and GRTB18S have a response time below 500 µs at a 1,000 Hz switching frequency. Our fastest laser sensor, the DT50-2B215252, manages 0.83 ms in its fastest mode. So a plain photoelectric sensor reacts at least 1.7 times faster (that is our division, and "under 500 µs" is an upper bound — the datasheet gives no exact figure); it simply does not tell you how far away the object is.

We do not restate the photoelectric circuit types (diffuse, retro-reflective, through-beam) here; they are covered in our guide to photoelectric sensors. Task-based selection in general sits in choosing sensors for automation, and if your job is tank level, go to level sensors: that has its own rules and this article does not cover them.

The Number That Is Not in the Price List: Remission Factor

When a laser sensor is described as "12 metres", that is 12 metres to a target with 90% remission, meaning something white. Sick datasheets print three ranges rather than one, and the gap between them decides whether you buy the right sensor or replace it a month later.

The ratios in the last column are ours, from dividing the datasheet figures. The practical conclusion is simple: if you measure to black rubber, dark coal or matte black paint, read the 6% column and forget the number in the product name. For the DT50-P1123 that means 2.5 metres instead of ten. And one more thing: on the DT35-B15251, DT35-B15551 and DT50-2B215252 the maximum ranges in this table are declared for the Slow speed setting (footnote 1), so not at the same time as the shortest response time.

Now the interesting part. Neither Pepperl+Fuchs ultrasonic datasheet has a remission-factor column at all. In its place sits "standard target plate 100 × 100 mm", so the range is specified against a plate of a defined size. That is an observation about how two manufacturers build their datasheets, not our verdict on the physics: the ultrasonic datasheet simply does not describe range through surface reflectivity, it describes it through target size. If your object is dark but large and solid, ultrasound looks at that job differently from a laser.

Accuracy: ±10 mm Absolute against ±1% of Full Scale

Here the two manufacturers measure with different rulers, and the figures cannot be compared head to head.

Sick prints an absolute error: typically ±10 mm on the DT35-B15251 and DT35-B15551, ±7 mm on the DT50-2B215252, ±10 mm on the DT50-P1113, DT50-P1123 and DS50-P1112, ±7 mm on the DL50-P1123 and ±2 mm on the DL100-21AA2112. "Absolute" here means one thing only: the datasheet gives one figure for the whole range rather than one per distance. The condition under which that figure holds, however, differs per datasheet, and that is exactly what gets skipped when models are compared with each other.

The difference between "6 to 90%" and "90% only" is practical, not formal. The DT35 datasheet gives its ±10 mm on black too, within the corresponding range, while the DT50-P1113 datasheet promises ±10 mm on white only: what happens at 6% remission it does not specify, and we are not going to specify it on its behalf.

Pepperl+Fuchs prints a relative one: deviation of the characteristic curve ±1% of full-scale value, repeat accuracy ±0.1% of full-scale value, and temperature influence another ±1% of full-scale value. Converting that to millimetres (our arithmetic): on the UB2000-F42-I-V15 with its default upper limit of 2,000 mm that is ±20 mm, and on the UB4000-F42-I-V15 at 4,000 mm it is ±40 mm.

Do not confuse resolution with accuracy. The 0.7 mm resolution of an ultrasonic sensor is the smallest step it can display, not how close that number is to the truth. The truth is in the "deviation of the characteristic curve" row: on the UB2000 it is 28.6 times coarser than the resolution (±20 mm against 0.7 mm), and on the UB4000 57.1 times (±40 mm against 0.7 mm). Both ratios are our division.

A word on repeatability, because it is easy to misread. The datasheets mean by it a statistical spread under unchanged conditions: ≥0.5 mm on the DT35, where the row is annotated "equivalent to 1 σ" and applies to a 6 to 90% remission factor; ≥2.5 mm on the DT50-P1113 and DT50-P1123; ≥2 mm on the DL50-P1123; and 0.5 mm on the DL100, noted as "statistical error 1 σ, environmental conditions constant, min. warm-up time 10 min". On the DT35-B15251 that row carries three footnotes at once: 1 σ, a 6 to 90% remission factor, and a pointer to the repeatability characteristic lines at the end of the datasheet. There are five of those graphs, one per speed setting from Super Slow to Super Fast, and each carries three lines — 6%, 18% and 90% remission — that diverge with distance. So ≥0.5 mm is the best corner of those graphs, not a constant. So that row does not answer the question "will the device read the same as yesterday". Other rows answer that one: on the ultrasonic sensors it is temperature influence of ±1% of full scale (another ±20 mm or so at a two-metre limit), and on the laser sensors it is the warm-up time (up to 15 minutes on the Dx50, typically 7 minutes on the DL100 below −10 °C). If you need stability between shifts, measure after warm-up and at a stable temperature.

Speed: 0.83 ms against 325 ms

This is the widest gap between laser and ultrasound in the whole article, but it has to be read carefully: the two manufacturers define the "response time" row differently.

The gap between the UB4000 and the DT50-2B215252 works out at roughly 390 times (our division of 325 by 0.83; the conditions on those two rows differ — see below). Note the pattern inside the ultrasonic pair itself: the two-metre UB2000 runs a transducer at about 175 kHz with a 150 ms response, while the four-metre UB4000 runs at 85 kHz with a 325 ms response. So within one family a longer range comes together with a lower transducer frequency and a longer response. That is a correlation taken from two datasheets, not our physics: the documents do not print the cause, and we are not inventing one.

Now the conditions, without which that table reads wrong — and they differ between datasheets, there is no single rule here.

• DT35-B15251, DT35-B15551, DT50-2B215252. Both rows, response time and switching frequency, carry the same two footnotes: 5) the value depends on the selected speed setting (Super Fast … Super Slow) and 6) it is specified for lateral entry of the object into the measuring range. So the 0.83 ms on the DT50-2B215252 and the 2.5 ms on the DT35-B15251 are not "the response time" in general, they are the fastest setting on lateral entry.
• DT50-P1113, DT50-P1123. These datasheets have no switching frequency row at all. Their 20 / 30 ms response time carries two different footnotes: 3) it depends on the averaging setting, fast/slow, and 5) lateral entry of the object into the measuring range. Super Fast … Super Slow modes are not printed here.
• DS50-P1112. Here both rows are present, 10 / 50 ms and 50 / 10 Hz, and both carry one footnote — dependence on the averaging setting, fast/slow.
• DL50-P1123. No switching frequency row, and the 15 / 30 ms response time carries two footnotes: 3) the averaging setting, fast/slow, and 4) lateral entry of the object into the measuring range.
• DL100. Its 2 ms response time carries no footnotes, but the datasheet separately specifies a maximum movement speed of 15 m/s.

Besides the response time, Sick prints a separate output time row: 0.33 / 1.33 / 3.33 / 10 / 30 ms on the DT50-2B215252 and 1 / 2 / 4 / 8 / 32 ms on the DT35-B15251. The datasheet prints it as its own row and does not state how it relates to the response time, so we neither add it to the response time nor discard it: if your machine cycle is anywhere near these figures, that is a question for the manufacturer's technical support, not for our arithmetic.

The Pepperl+Fuchs datasheet has none of that: it prints "response delay approx. 325 ms" with no conditions at all. So our 390-times ratio divides two differently defined rows. Its order of magnitude is honest — between laser and ultrasound it really is hundreds of times. Its second significant digit is not, and a conveyor cycle time should not be calculated from it.

On the laser sensors the response time is not fixed, you pick it with a speed setting, and you pay for speed with accuracy. The DT35 has five modes from Super Fast to Super Slow, and its rated 12,000 mm is declared for the Slow setting specifically (footnote 1 in the datasheet). So "12 metres and 333 Hz" at the same time is not on offer. The same holds for the DT35-B15551 and the DT50-2B215252: in all three datasheets the maximum range is taken at the slow setting and the minimum response time at the fastest one, and those two numbers do not exist together.

Laser Class 1 against Class 2: What It Costs

If your site requires a class 1 laser (often an occupational safety requirement for areas where people walk), be ready to pay for it in either speed or range. We stock two pairs of sensors that differ in exactly this parameter, which makes the cost visible.

What matters here is separating what the datasheets print from what they do not. On the DT35-B15551 the max. output is the same, 250 mW, the duty cycle is 1/500 against 1/250, and the switching frequency is 166 against 333 Hz. Those are three separate rows, and the datasheet claims no causal link between them — neither do we. On the DT50-P1123 the max. output is cut from 180 to 120 mW while the response time does not change at all (20 / 30 ms, as on the P1113): the difference in that pair shows up in range on dark surfaces. The normative reference for the laser class — IEC 60825-1:2014 and EN 60825-1:2014 — appears in all eight datasheets we opened, and in both DT35 sheets in particular.

One thing worth knowing before you pay for class 1: the footnote "do not intentionally look into the laser beam, never point the laser beam at people's eyes" appears in the class 1 and the class 2 datasheets, in identical wording. What exactly separates class 1 from class 2 is defined by IEC 60825-1; our datasheets name only the class number and a reference to that standard, so we are not retelling the classification here.

Reflector against Natural Surface: How 50 and 100 Metres Happen

Two of our sensors measure over tens of metres, and both do it against reflective tape rather than the object. In the target row of their datasheets it does not say "Natural objects", it says Reflector.

• DL100-21AA2112: 0.15 to 100 m on diamond grade tape, accuracy ±2 mm, repeatability 0.5 mm, response time 2 ms, resolution selectable (0.1 / 0.125 / 1 / 10 / 100 mm). Supply 18 to 30 V, consumption under 250 mA at 24 V. Two rows in this datasheet decide whether the device fits at all: Scope — Indoor, meaning the manufacturer specifies it for indoor use, and a maximum movement speed of 15 m/s (acceleration up to 15 m/s²). If the sensor has to sit outdoors, that is agreed with the manufacturer rather than decided for them. This is a positioning device, not a detection one: it typically measures to tape on a moving carriage.
• DL50-P1123: 200 to 50,000 mm on Diamond Grade film, accuracy ±7 mm, repeatability ≥2 mm, response time 15 or 30 ms, laser class 1, IP65. Initialization time ≤250 ms, warm-up up to 15 minutes.

The film is a separate item, not part of the sensor package. The same "Diamond Grade" both datasheets are specified against is in our catalogue as REF-DG. Ordering the sensor without it makes little sense: against a natural surface these two models will not deliver their tens of metres.

The DL100 spot grows with distance per its datasheet formula, 5 mm + (2 mm × distance in metres). At 50 metres that is 105 mm (our substitution into the datasheet formula). So the tape at that distance has to be wider than 10 centimetres, or part of the beam misses it. People usually remember this during installation.

The Dead Band: Why Ultrasound Cannot Measure Close Up

An ultrasonic sensor has a zone near the membrane where it does not work, and it is a large one. On the UB2000 it is 0 to 60 mm; on the UB4000 it is already 0 to 200 mm. The datasheets give this as a separate "dead band" row and offer no explanation, so we are not guessing at one. All that is visible is that the band grows with the range. The practical conclusion is single and hard: nothing is measured closer than 200 mm to the face of the UB4000.

The P+F datasheets hold one more row that is easy to miss: adjustment range. The UB2000 has a sensing range of 60 to 2,000 mm but its switching limit can only be set between 90 and 2,000 mm. The lower end of the adjustment range sits above the lower end of the sensing range. Same on the UB4000: sensing from 200 mm, adjustment from 240 mm.

Note that the "dead band" row belongs to the ultrasonic datasheets. The laser datasheets have no such row; there the lower end is simply written into the measuring range: 50 mm on the DT35, 200 mm on the Dx50 units whose datasheets we opened, 150 mm on the DL100 — so the minimum is non-zero there too, but it is a different datasheet row, not a dead band. If you need to measure from 20 mm, none of the ten devices whose datasheets we opened fits, and that is a case where it is better said up front.

Outputs and Integration: Where the Problems Hide

An analogue output has a load limit, and it needs checking against your controller input before the order, not after.

• 4 to 20 mA, permissible load resistance: ≤450 Ω on the DT35-B15251 and DT50-2B215252, ≤300 Ω on the DT50-P1113 and DT50-P1123, 0 to 300 Ω on both ultrasonic sensors. With a long cable plus a shunt at the PLC input, 300 Ω runs out faster than you would think.
• A voltage output is not on every model. 0 to 10 V into ≥50,000 Ω is printed by the DT35-B15251 and DT50-2B215252 datasheets; the DT50-P1113 and DT50-P1123 only have the 4 to 20 mA current output into ≤300 Ω. And a separate warning from the DT35-B15251 datasheet: the supply must be above 13 V when the analogue voltage output is used and above 18 V with IO-Link, even though the supply range itself starts at 12 V.
• Analogue output resolution: 12 bit on the DT35-B15251 and 16 bit on the DT50-P1113, DT50-P1123 and DT50-2B215252. Here the cheaper DT35-B15251 loses to the more expensive DT50 family, which is not a catalogue error: both figures are printed in the respective datasheets. The documents do not explain why, so we are not inventing a reason.
• IO-Link V1.1 on the DT35-B15251 at 38.4 kbit/s with process data, parameterization and diagnosis. The same channel sets the hysteresis anywhere from 0 to 11,950 mm. If you plan to reconfigure the sensor from the PLC rather than with buttons, that is the row to read.
• Digital outputs — the biggest trap in this section. Push-pull PNP/NPN is printed by four datasheets only: DT35-B15251, DT35-B15551, DT50-2B215252 and DL100. On the DT50-P1113, DT50-P1123, DS50-P1112 and DL50-P1123 the type row simply says "PNP". Count it the other way round, because it is clearer: push-pull with NPN is offered by only four of the eleven items — DT35-B15251, DT35-B15551, DT50-2B215252 and DL100. The remaining Dx50 and DL50 units are PNP only, and neither ultrasonic sensor has a digital output at all: one analogue 4 to 20 mA output and nothing else. So if your input accepts NPN only, the choice in this section narrows to four items rather than widening. We did not open the DT50-2B215552 datasheet, so it is not part of this count. The DT35-B15251 has one or two outputs, up to 100 mA, under 3 V of voltage drop, and its output Q2 switches between current, voltage and digital modes.

We covered PNP and NPN logic and wiring into controller inputs separately: connecting a sensor to a PLC, NPN against PNP — and if your input is NPN-only, start there, because the choice narrows to those same four items. On housings and conditions: the DT35, the Dx50 units and the DL100 are all IP65 with protection class III, while the ultrasonic UB2000 and UB4000 are IP54, worth noting if the device sits anywhere that gets hosed down. The ultrasonic operating range is −25 to 70 °C with an ABS housing.

Connectors are uniform, but worth checking before the order: every measuring sensor in this line uses an M12 5-pin male connector (the DL100 an M12 that is SPEEDCON compatible, plus separate PROFINET ports), the ultrasonic UB2000 and UB4000 also M12 5-pin, and the photoelectric GRL18S and GRTB18S M12 3-pin. The matching cable is not part of the sensor's scope of delivery: we stock, for example, the YF2A15-100VB5XLEAX, an M12 cable 10 m long.

And a word on commissioning, the second most common question after the output. Here we state only what the datasheets print, because we did not open the operating instructions for these models in this pass. The datasheets print two things: the row "teach-in, scaling and inversion of the analog and digital output" (with an "Easy teach option" added on the DT50-2B215252) and the row "multifunctional input: laser off / external teach / deactivated", meaning teach can be triggered by an external signal rather than by hand only. What exactly is pressed, and in what sequence, the datasheet does not say: the DT35-B15251 indication row reads simply "LEDs" (no display; the dimensional drawing labels "control elements"), while the DT50-2B215252 reads "3 × LED, LC display". The step-by-step procedure sits in the operating instructions, and we are not paraphrasing it. The second route is IO-Link: V1.1 at 38.4 kbit/s on the DT35-B15251, carrying parameters, diagnosis and the same 0 to 11,950 mm hysteresis. The practical difference is that teach has to be done on each device separately, while parameters over IO-Link can be pushed from the PLC and restored after a sensor is replaced.

What We Have: All Eleven Items

As queried from the catalogue on 2026-11-10. This is not a statement about stock: check the product pages for what is current.

That last row is left blank deliberately. We did not open the DT50-2B215552 datasheet in this pass, so we quote no figures for it: an empty cell beats a guessed number. Write to us and we will open the datasheet against your specific question.

And what usually goes with the sensor. These are separate items, not part of the package:

And one remark about money, because it is easy to draw the wrong conclusion in either direction. At the short end of the line ultrasound really is cheaper: the UB2000-F42-I-V15 is the cheapest of all eleven items, and the cheapest laser sensor, the DT35-B15251, costs about a third more. But that is the only place where "ultrasound because it is cheaper" holds. The second ultrasonic unit, the UB4000-F42-I-V15, already costs more than either DT35, while offering a percentage of full scale instead of an absolute error and a response time in the hundreds of milliseconds. And the most expensive item in the line, the DL100-21AA2112, differs from the cheapest by nearly an order of magnitude — and what you pay for there is not range but ±2 mm, 2 ms and PROFINET. So the comparison to make is not between principles but between two specific items for your task. Current figures are on the product pages.

How We Choose: Four Questions in Order

1. A number or a fact? If a fact, stop reading, take a photoelectric sensor and save the money.
2. What surface? Dark and matte means the 6% column in the datasheet, not the number in the product name. If reflective tape can be stuck to the object, the DL50 and DL100 open up with their tens of metres — we also sell the Sick "Diamond Grade" foil itself, as REF-DG.
3. What is the minimum distance? The 200 mm dead band on the UB4000 (a row that only the ultrasonic datasheets carry), or the 200 mm lower end of the measuring range across the Dx50 family, rules out jobs that measure up close.
4. How many times a second? Less than once a second, and the difference between 0.83 ms and 325 ms means nothing to you, which makes ultrasound a reasonable option. If the object moves, read the frequency column and remember that the maximum range is declared for the slow setting.

What We Ask Before an Order

Surface material and colour, minimum and maximum distance, the accuracy you need in millimetres (not in percent), the measurement rate, and the input type on your controller. Five items. The third is where people usually stumble: "it has to be accurate" is not a requirement, while ±5 mm at two metres is one, and against it you can see what fits and what does not.

All eleven items are in the distance sensor section; all Sick sensors are collected in the brand filter, and the full catalogue in the sensors section. If what you need is a switching sensor rather than a measuring one, go to photoelectric sensors. Not sure which is yours: send the five answers above through contacts and we will check them against the datasheets.