Pulse Width and Range Resolution in a Laser Rangefinder
*Two targets close together can arrive at the detector as one return. Pulse width decides how close is too
close — and it is one of the few specifications that cannot be improved after the fact. This article explains
the relationship, works it through the published figures, and covers what a shorter pulse costs.*
In short: range resolution in metres is approximately the pulse width in nanoseconds multiplied by 0.15. A
4 ns pulse separates surfaces about 0.6 m apart; a 12 ns pulse needs about 1.8 m. In this range the short pulses
belong to the low-energy models and the long pulses to the high-energy ones, so the two requirements have to be
traded rather than both satisfied.
What pulse width means
Pulse width, quoted in nanoseconds, is the duration of the optical output. It is measured at full width,
half maximum — the interval between the points where the pulse has fallen to half its peak power. Every model in
this range is specified that way unless stated otherwise.
Two different quantities share similar names and are worth separating now, because confusing them is a common
source of error:
Optical pulse width is the duration of the light. This is what the specification table means by "Pulse
width", and it is what determines range resolution.
Drive pulse width is the duration of the electrical pulse the driver supplies. It is much longer — hundreds
of microseconds rather than nanoseconds — because it is the energy storage time, not the emission time. For the
ER800 the drive pulse is up to 3 ms and the optical pulse is 7 ns; the two differ by a factor of about 400,000.
The datasheet gives both, and they are not interchangeable.
Why pulse width sets resolution
A laser pulse is not a point in time. It has a beginning and an end, and the light is spread over that interval.
When the pulse strikes a surface, the return has roughly the same duration as the outgoing pulse.
If a second surface lies behind the first, its return arrives later by the round-trip time to cover the extra
distance. If that delay is shorter than the pulse, the two returns overlap and the detector sees one longer
pulse rather than two. The system cannot tell whether it saw one thick object or two thin ones.
The boundary is conventionally taken as half the pulse length in distance, because that is where the two returns
are separated by the pulse duration and become distinguishable:
range resolution = c × pulse width / 2
where c is the speed of light, 299,792,458 m/s. The factor of two accounts for the light travelling out and
back.
For a convenient rule of thumb, the arithmetic collapses to:
resolution in metres ≈ pulse width in nanoseconds × 0.15
So a 5 ns pulse resolves about 0.75 m, and a 10 ns pulse about 1.5 m.
The published figures
Every pulse width in the ERDI range, with the resolution it gives:
| Pulse width | Range resolution | Models |
|---|---|---|
| 3.5 ns | 0.52 m | MINI-100 |
| 4 ns | 0.60 m | ER40, ER70 |
| 5 ns | 0.75 m | ER20, ER200, ER300 |
| 6 ns | 0.90 m | ER100, ER400, ER500 |
| 7 ns | 1.05 m | ER800, ER900 |
| 12 ns | 1.80 m | ER2000, ER4000, ER10000 |
| 13 ns | 1.95 m | ER8000 |
Read down the right-hand column and the pattern is clear. The finest resolution belongs to the MINI-100 at
3.5 ns. The coarsest belongs to the ER8000 at 13 ns.
The tension in the range
Look at which models appear in which row, and a design tension becomes visible.
The short pulses are in the low-energy models. The 3.5 to 7 ns pulses belong to the micro, compact and
standard tiers, running from 100 µJ to 900 µJ.
The long pulses are in the high-energy models. Everything from 2 mJ upward has a 12 or 13 ns pulse.
This is not an accident of design. Generating more energy in the same cavity means storing more energy and
releasing it over a longer interval; the pulse lengthens as the energy rises. It is a physical consequence, not
a specification that could simply be improved.
The practical consequence: **an application that needs both long range and fine separation cannot have both from
a single standard model.** It has to choose, or the system has to solve it another way.
Working out which matters more
The question to ask is what the instrument must distinguish.
If the target is isolated — a building face, a distant vehicle on an empty road, a reflective marker — the
return is one pulse and range resolution is irrelevant. What matters is whether the return is detectable at all,
which is a question of energy. The high-energy models are the right choice and the 12 ns pulse costs nothing.
If the target has structure — a cable against a hillside, a vehicle against the road behind it, a person in
front of a wall, vegetation over ground — the two returns must be separated. A 12 ns pulse will merge them into
one up to 1.8 m apart. A 4 ns pulse separates them from 0.6 m.
If the target is small and close to something else, the resolution requirement may rule out the energy tier
entirely, and the design has to find range another way: a larger receiver aperture, a more sensitive detector,
or a different measurement principle.
What a shorter pulse costs elsewhere
Pulse width does not only affect resolution. Three other parts of the design are affected, and they are worth
knowing before specifying.
Receiver bandwidth
The detector and the analogue chain must have the bandwidth to preserve the pulse. As a rule the bandwidth
should be at least the reciprocal of the pulse width, and in practice comfortably more.
| Pulse width | Bandwidth needed |
|---|---|
| 4 ns | ~250 MHz, in practice 500 MHz or more |
| 7 ns | ~140 MHz, in practice 300 MHz |
| 12 ns | ~85 MHz, in practice 200 MHz |
A 4 ns pulse needs a faster receiver than a 12 ns pulse. That is more expensive, draws more power, and in a
noisy environment collects more noise bandwidth — which is to say that a shorter pulse and a faster receiver
together make detection harder, and some of the energy advantage is spent there.
Peak power
Two pulses carrying the same energy do not have the same peak power. Halve the duration and the peak doubles.
| Model | Energy | Pulse width | Approximate peak power |
|---|---|---|---|
| ER70 | ≥70 µJ | 4 ns | ~17.5 kW |
| ER800 | ≥800 µJ | 7 ns | ~114 kW |
| ER2000 | ≥2 mJ | 12 ns | ~167 kW |
Peak power is what the optical coatings, the output window and the fibre, if any, must survive. It is also what
the eye-safety calculation deals with. A short pulse is not automatically the gentle choice.
Detector and timing resolution
The time-to-digital converter must resolve intervals finer than the pulse, or the resolution the laser offers is
lost in the electronics. For a 4 ns pulse that means a TDC with resolution well under a nanosecond; for a 12 ns
pulse, considerably less demanding.
Specifying a 4 ns source and pairing it with a timing chain that resolves to 10 ns wastes the laser.
Two returns that should not be merged
There is a case where a longer pulse is deliberately chosen: when the aim is to measure the last return rather
than the first.
In vegetation or camouflage, several surfaces return light — leaves, branches, then the ground. A receiver
configured to detect the last return above a threshold can penetrate canopy, and a longer pulse with more energy
per pulse is often what makes that possible at range. The resolution is coarse, but the measurement is of the
ground rather than the leaves, which is the point.
This is the same trade seen from the other side: resolution is not always the objective.
What to write down before choosing
Pulse width is one of the specifications that cannot be corrected in software or in a later design revision. It
is decided when the model is chosen. Before that:
- What is the closest separation the instrument must resolve? Convert it to a maximum pulse width:
pulse width ≤ separation / 0.15 in nanoseconds against metres.
- Is the target isolated? If yes, step 1 does not apply and the energy tier can be chosen on range alone.
- What bandwidth can the receiver afford? A short pulse needs a fast, noisier chain.
- What peak power must the optics survive? Short pulses concentrate the same energy into less time.
- Does the timing chain resolve fine enough? A TDC coarser than the pulse discards the advantage.
- Is a last-return measurement wanted instead? If so, the reasoning inverts.
The models, with both pulse quantities
The published figures, including the drive pulse width, which is the one that must be matched by the driver:
| Model | Optical pulse width | Drive pulse width | Pulse energy | Max repetition rate |
|---|---|---|---|---|
| MINI-100 | 3.5 ns | ≤1 ms | ≥100 µJ | 10 Hz |
| ER100 | 6 ns | ≤2.4 ms | ≥100 µJ | 10 Hz |
| ER200 | 5 ns | ≤2.4 ms | ≥200 µJ | 10 Hz |
| ER300 | 5 ns | ≤2.4 ms | ≥300 µJ | 10 Hz |
| ER400 | 6 ns | ≤2.4 ms | ≥400 µJ | 10 Hz |
| ER500 | 6 ns | ≤2.4 ms | ≥500 µJ | 10 Hz |
| ER800 | 7 ns | ≤3 ms | ≥800 µJ | 10 Hz |
| ER900 | 7 ns | ≤3 ms | ≥900 µJ | 5 Hz |
| ER2000 | 12 ns | ≤4 ms | ≥2 mJ | 5 Hz |
| ER4000 | 12 ns | ≤4 ms | ≥4 mJ | 5 Hz |
| ER8000 | 13 ns | ≤4 ms | ≥8 mJ | 5 Hz |
| ER10000 | 12 ns | ≤4 ms | ≥10 mJ | 5 Hz |
| ER10 | 5 ns | ≤0.3 ms | ≥10 µJ | 5 kHz |
| ER20 | 5 ns | ≤0.3 ms | ≥20 µJ | 2.5 kHz |
| ER40 | 4 ns | ≤0.4 ms | ≥40 µJ | 1 kHz |
| ER70 | 4 ns | ≤0.4 ms | ≥70 µJ | 500 Hz |
The drive pulse width is worth noting for a second reason: it changes by more than a factor of ten across the
range, from 0.3 ms for the ER10 to 4 ms for the ER10000. A driver designed for one tier will not necessarily
serve another.
Related reading
— where pulse width sits in the selection order
— why the wavelength was chosen
- LiDAR point density — how repetition rate and pulse width
combine in a scanning system
Next steps
If the separation requirement is known, the Contact / RFQ form is the quickest route
to a recommendation — include the closest distance the instrument must resolve and the required range, and the
two together usually identify a model. Datasheets and 3D models for every model are on the
Laser Sources page.