OTDR Pulse Width Explained: Choosing the Right Setting

What OTDR Pulse Width Actually Is

An OTDR works by launching short optical pulses into the fiber and measuring the light that returns. Pulse width is the duration of each pulse, expressed in nanoseconds. Common values range from 5 ns (a very short pulse) to 20,000 ns (a very long pulse).

In the time domain, pulse width is just a number of nanoseconds. In the spatial domain, the pulse occupies a length of fiber equal to the pulse duration multiplied by the speed of light in fiber. Light travels through fiber at roughly 200,000 km/sec (about two-thirds of vacuum c, due to the index of refraction). A 5 ns pulse is therefore about 1 meter long in the fiber. A 100 ns pulse is 20 meters long. A 1000 ns pulse is 200 meters long.

This length-in-fiber is what determines the OTDR's resolution. If the pulse is 200 meters long, you cannot distinguish between two events that are 100 meters apart -- they get smeared into a single broad event on the trace. For background on how the OTDR uses these pulses to build a trace, see OTDR testing basics.

The Core Trade-Off: Resolution vs Dynamic Range

Pulse width sets a trade-off between two competing test requirements: resolution (the ability to discriminate closely spaced events) and dynamic range (the distance the OTDR can measure before signal disappears into noise).

Short Pulses Give Better Resolution

A short pulse can resolve closely spaced events because two adjacent reflections do not overlap in time. With a 5 ns pulse, you can distinguish two connectors 2 meters apart. With a 100 ns pulse, those two connectors merge into one event because the reflection from each lasts longer than the gap between them.

Resolution matters most in FTTH and patch panel environments where events are close together. A splitter cabinet may have 8 to 32 output connectors within a few meters. To verify each output independently, you need pulse widths short enough to resolve them as separate events.

Long Pulses Give Better Dynamic Range

A long pulse delivers more total energy into the fiber and produces a stronger return signal. This lets the OTDR measure further before the backscatter level falls below the noise floor. Dynamic range is the OTDR specification that quantifies this -- typically expressed in dB, with 5 dB more dynamic range translating to roughly 30 km more reach on standard single-mode fiber.

Long pulses are essential for backbone and long-haul testing where the link may be 50, 100, or even 200 km long. A short pulse simply will not reach the far end -- the trace fades into noise long before the end-of-fiber spike.

The Net Effect

You cannot have both. A pulse width that gives 1-meter resolution will not reach a 100 km link. A pulse width that reaches 100 km cannot resolve events 50 meters apart. Real-world testing uses different pulse widths for different parts of the trace: a short pulse for the near end where most events live, and a longer pulse to reach and characterize the far end.

Pulse Width Reference Table

Approximate values for typical OTDRs. Specific dead zones and dynamic ranges vary by instrument; check your OTDR's spec sheet.

Choosing Pulse Width by Link Type

FTTH Drops (Under 2 km)

Use 5-10 ns pulse width. The drop is short and packed with events: connector at the OLT, splitter, drop fiber, ONT connector. You need fine resolution to distinguish the splitter outputs from each other. Longer pulses blur the splitter into a single event and you cannot verify which port has the problem. Pair with a 100-300 meter launch fiber so the OTDR's dead zone does not eat the first connector.

FTTH Feeder and Distribution (2-10 km)

Use 10-30 ns pulse width. Long enough to reach beyond the first splitter and characterize the distribution fiber, short enough to resolve cabinet events. If the link includes a 1:32 splitter, factor in the ~17 dB splitter loss when choosing pulse width -- you may need a longer pulse to maintain signal beyond the splitter.

Access Network and Short Metro (10-30 km)

Use 100-300 ns pulse width. Resolution drops to 20-60 meters but events on access networks are typically further apart anyway. The extra dynamic range lets you reach the central office or hub from any drop or remote site. For very short metro segments (5-10 km), 30 ns may be sufficient and gives better event resolution.

Metro and Regional (30-80 km)

Use 1000 ns pulse width. Resolution is roughly 200 meters, which is fine for backbone fiber where splice points are typically kilometers apart. Bidirectional testing is essential at these distances because gainer artifacts at splices become large.

Long-Haul (80+ km)

Use 3000-20000 ns pulse width depending on link length. At 200 km, 10000 ns is standard. Plan for long acquisition times -- a 3000 ns trace at full averaging takes 3-5 minutes; 20000 ns can take 10+ minutes per wavelength. Document with longer averaging because long-pulse traces need more averaging to maintain noise floor.

The Multi-Pulse Strategy for Documentation

For acceptance testing or detailed link characterization, no single pulse width gives complete coverage. The standard approach is to acquire multiple traces at different pulse widths and combine the results.

• Short pulse trace for events near the OTDR end. Resolution is excellent for the launch fiber, first connector, splitter outputs, and any near-end splices. The trace fades into noise before the far end but that is fine -- the long-pulse trace covers that section.
• Long pulse trace for events near the far end. Resolution is poor near the start (everything blurs) but the trace reaches all the way to the end-of-fiber event. Far-end connectors and final splices are characterized accurately.
• Combine the results in your test report. Use the short-pulse measurements for near events and long-pulse measurements for far events.

This is the workflow for documentation-grade OTDR testing on any link more than a few kilometers long.

Pulse Width and Dead Zones

Every pulse width creates two corresponding dead zones: the event dead zone (minimum spacing between two reflective events to detect them as separate) and the attenuation dead zone (minimum distance after a reflective event to make accurate loss measurements). Both scale roughly linearly with pulse width.

The Event Dead Zone

For a 5 ns pulse, event dead zone is approximately 1 meter. For 100 ns, it is roughly 20 meters. This is the minimum spacing for the OTDR to count two reflections as separate events. Closer than that and they merge into a single broader event on the trace.

The Attenuation Dead Zone

The attenuation dead zone is typically 4-5x larger than the event dead zone. For a 5 ns pulse, attenuation dead zone is around 5 meters. For 100 ns, it is around 80-100 meters. Inside the attenuation dead zone, the trace is recovering from the reflection and any loss measurements are unreliable.

Practical Impact

If you have two connectors 5 meters apart and you use a 100 ns pulse, both connectors fall inside each other's dead zones. You see one merged event with combined loss instead of two separate events. The fix is a shorter pulse: 5-10 ns gives event dead zones small enough to resolve them. We cover this in detail in OTDR dead zones explained.

Pulse Width and Averaging Time

Pulse width interacts with averaging time to determine trace quality. Short pulses produce noisy traces because each pulse delivers less energy; longer averaging compensates by combining more pulses to reduce noise. The two settings are linked in practical use.

Short Pulse, Short Averaging

For quick FTTH spot checks, a 5-10 ns pulse with 10-15 seconds of averaging gives a usable trace in under 30 seconds total. The trace is noisier than documentation-quality but adequate for verifying connectivity and finding obvious faults.

Short Pulse, Long Averaging

For acceptance documentation on FTTH and access fiber, use 5-10 ns pulse with 60-180 seconds of averaging. The longer acquisition reduces noise floor enough to detect small splice losses (0.02-0.05 dB) and low-reflectance events.

Long Pulse, Short Averaging

For long-haul reach checks where you just need to see the end of the fiber, use the longest available pulse with minimal averaging. The trace is rough but reaches the far end.

Long Pulse, Long Averaging

For long-haul documentation, long pulses with long averaging (3-10 minutes) deliver clean traces with accurate event analysis at long distances. This is the slowest configuration but produces the highest-quality long-haul traces.

Common Pulse Width Mistakes

• Long pulse on FTTH. A 300 ns pulse on a 1 km drop blurs the splitter and patch panel into a single event. You see "the link works" but cannot identify which connector is the high-loss one. Use 5-30 ns for FTTH.
• Short pulse on long-haul. A 30 ns pulse will not reach 50 km. The trace looks normal up to about 5 km then dives into noise. You think the link ends at 5 km when actually you just ran out of dynamic range. Match pulse width to link length.
• Trusting auto pulse width for documentation. Auto picks a single value optimized for "reach the end" not "characterize every event accurately." For documentation work, choose pulse widths manually.
• Not adjusting pulse width when wavelength changes. Loss is higher at 1310nm than 1550nm by about 0.13 dB/km. The 1310nm trace runs into noise sooner. You may need a longer pulse at 1310nm than at 1550nm to characterize the same link.
• Forgetting splitter loss. A 1:32 PON splitter adds ~17 dB to the link budget. The pulse width needed to reach 5 km after a 1:32 splitter is much longer than for unsplit 5 km fiber.

Real-World Pulse Width Examples

Example: 500m FTTH Drop with 1:32 Splitter

The link is short but the splitter adds 17 dB of loss, dropping the post-splitter signal substantially. Use 5-10 ns pulse width to characterize the OLT-to-splitter feeder, then potentially a 30-100 ns pulse for the post-splitter section if signal is weak. Two acquisitions, two pulse widths, complete characterization.

Example: 25 km Metro Link with No Splitter

Standard 30-100 ns pulse width handles this comfortably. Resolution is in the 6-20 meter range, fine for the few connectors and splices typical on metro fiber. Single-pulse acquisition is sufficient.

Example: 80 km Backbone Fiber

Use 1000-3000 ns pulse width. The longer pulse is necessary to maintain signal at 80 km. Resolution is 200-600 meters, which is fine because backbone splice points are typically kilometers apart anyway. Plan 5-minute acquisitions per wavelength.

Example: Patch Panel Verification

Patch panel connectors may be inches apart. Use the shortest pulse width available, typically 5 ns. Even then, very tight panel spacing may merge connectors into single events. The OTDR is not the right tool for very dense patch panels; use a power meter and light source.

OTDRs and Test Equipment

The Fiber Ranger OTDR covers pulse widths from 5 ns to 20 microseconds, which handles every link type from FTTH drops to long-haul backbone. For most FTTH techs, the 5-30 ns range is what gets used 90% of the time.

For complete OTDR test workflow, see how to launch an OTDR test step by step and OTDR trace interpretation. Pair your OTDR with a fiber inspection scope and optical power meter for complete fiber link certification.

How Pulse Width Interacts With Wavelength

The same pulse width does not produce identical traces at 1310nm and 1550nm. Fiber attenuation differs, dynamic range differs, and noise floor differs. The same OTDR with the same pulse width will show slightly different reach at the two wavelengths.

1310nm: Higher Attenuation, Less Reach

At 1310nm fiber loses about 0.35 dB/km. A pulse width that gives 60 km reach at 1550nm gives roughly 50 km reach at 1310nm. For the longest links, you may need a longer pulse width at 1310nm than at 1550nm to achieve the same reach.

1550nm: Lower Attenuation, More Reach

At 1550nm fiber loses about 0.22 dB/km. The same pulse width travels farther. For backbone links it is common to use a shorter pulse width at 1550nm than 1310nm because the lower attenuation gives the same reach with less spatial smearing.

1625nm: Live-Fiber Testing

1625nm is used for live-fiber maintenance OTDR testing because it sits outside the operational windows of GPON and most other transmission systems. Pulse width choice at 1625nm follows the same rules as 1550nm because attenuation is similar.

Pulse Width Quick Reference

For technicians who just want a starting point, here are pulse width recommendations by link type. Use these as initial settings, then adjust based on what the trace shows.

• FTTH drop, 50-300m: Start at 5 ns. Move to 10 ns only if noise is unacceptable.
• FTTH distribution, 300m-2km: 10-30 ns is the sweet spot for most distribution-side traces.
• Building backbone, 2-10km: 30-100 ns. Longer pulse helps push through splice losses on multi-segment runs.
• Metro link, 10-40km: 100-500 ns. Most modern OTDRs in this range have automatic pulse-width selection that gets you close.
• Long-haul, 40-150km: 1-10 microseconds. Reach is the constraint and resolution becomes secondary.
• Ultra long-haul, 150-300km: 10-20 microseconds. Specialized OTDRs only -- most field instruments cap at 100km reach.