Shoot a span with an OTDR plugged straight into the panel, and the trace opens with a huge spike — the launch reflection — followed by a stretch where nothing can be trusted. That stretch is the dead zone, and the span's first connector sits right inside it. A trace that starts "clean" has not proven the first connector at all: it has only proven that the OTDR cannot see it. The fix is unglamorous and cheap — a spool of fiber between the OTDR and the span — and it is the difference between a trace that is evidence and a trace that is decoration.
"Dead zone" gets used for two different blind spots, and confusing them is how a bad connector survives a test:
| Dead zone | What it hides | Typical size |
|---|---|---|
| Event dead zone | A second reflective event (another connector) close behind the first | 1–5 metres on handheld OTDRs |
| Attenuation dead zone | A small non-reflective loss (a splice, a bend) after a strong reflection, until the detector recovers enough to measure it | Tens of metres — much longer than the event dead zone |
The attenuation dead zone is the one that bites on real spans. A fusion splice 10 metres past a panel connector can be completely unmeasurable if the OTDR has not recovered from the connector's reflection by the time the splice's small loss arrives. The trace looks flat there — not because the splice is good, but because the detector was still blind. The same logic applies at the launch itself: the OTDR's own output connector reflection blinds the first stretch of whatever is plugged into it.
A launch cable is a spool of fiber — typically 100 to 1000 metres — connected between the OTDR and the span under test. It does not improve the OTDR. It moves the problem: the launch reflection and its dead zone now fall on the spool, where there is nothing to measure, and the span's first connector appears mid-trace, after the detector has settled back to the fiber's normal backscatter level. The first connector's loss becomes measurable, and a splice or bend shortly after it becomes visible too.
A receive cable (tail cord) at the far end does the same job for the last connector. Without it, the far-end connector's loss is unmeasurable for the same reason — the trace simply ends at the reflection, and any loss in that final connection is folded into the end spike where it cannot be separated out. A span tested with a launch cable but no receive cable has one proven end and one assumed end; the ticket should say which is which.
Long enough to clear the attenuation dead zone at the pulse width you are actually using — and pulse width is the lever most techs forget. A wider pulse reaches further down a long span, but it also widens the dead zone, because the detector takes longer to recover from a longer, stronger reflection. A spool that clears the dead zone at a short pulse width may not clear it at the wide pulse width you switched to for a 20 km span.
The check is on the trace itself, not on the spool's label: the launch reflection's recovery must settle fully back to the fiber's backscatter level before the first span connector appears. If the trace is still sloping down when the first connector's spike arrives, the spool is too short for that pulse width — use a longer spool or a shorter pulse width, and re-shoot. For most handheld work, a 100–300 metre spool covers short spans; longer spans or wider pulses call for 500–1000 metres. When in doubt, the longer spool costs a minute of setup and removes the argument entirely.
Sometimes you inherit a trace shot straight into the panel, or the launch box is on another truck. That trace is not useless — but its claims have limits, and the ticket must respect them:
The minimum useful note names the launch setup, the pulse width, and what was actually measured:
Span: Hall A ODF-2 ↔ Hall B ODF-5 (2.0 km SM). OTDR: launch cable 300 m + receive cable 150 m, pulse width 100 ns. First connector (Hall A): 0.4 dB (allowance 0.5 dB) — resolved clear of dead zone. Mid-span splice at 1.1 km: 0.08 dB (allowance 0.1 dB). Last connector (Hall B): 0.3 dB (allowance 0.5 dB) — resolved via receive cable. Total OTDR span loss 1.1 dB; LSPM end-to-end 1.2 dB @1550 nm — consistent.
And the honest fallback when there was no launch cable:
Span: Hall A ODF-2 ↔ Hall B ODF-5, OTDR shot WITHOUT launch cable. Mid-span clean (no events over allowance between 0.3–1.8 km). First + last connector loss NOT resolved by OTDR (launch dead zone). Ends covered by LSPM total: 1.2 dB @1550 nm, inside 2.7 dB budget — PASS on total loss.
The shape to copy: the launch setup (or its absence), the pulse width, each end connector's loss with its allowance, and whether the verdict rests on resolved events or on total loss. That note lets the next tech trust the trace exactly as far as it deserves — and no further.
See RackReady Free sample How to read an OTDR trace
Support: support@hasherventures.com
Free tool: turn your OTDR findings into a structured ticket note with the RackReady Ticket Builder.