Knowledge base · Measurement
Index of refraction: the one number the whole length rests on
An OTDR gets distance from time: it measures time and converts it into metres through the fibre's index of refraction. An error in the fourth decimal of that number puts you metres out over a kilometre.
Check the IOR in your own file
The interface calls it IOR; cable datasheets call it the index of refraction, or the group index.
By Ilya (engineer, 6 years in fibre optics) ·
Where it comes from
The value can come from three places, and they do not always agree.
The instrument's settings
The file holds whatever value the technician had set at the moment of the measurement. Often it is the instrument's default, left over from the previous site.
The cable's datasheet
The real value for this fibre type at this wavelength. That is the correct one, and it differs between manufacturers, and between wavelengths of the very same fibre.
A check against the design
If the optical length along the trace disagrees with the design and the closures sit "almost" where they should, the cause is usually here rather than in the cable. It takes a minute to check: put in the datasheet value and see whether things line up.
The instrument's field is a wide one. Yokogawa lets the operator enter anything from 1.30000 to 1.79999, to five decimal places, which is wider than the corridor outside which our screen calls the figure implausible. The instrument accepts any value inside that field, and the file carries whatever was typed.
The table the instrument carries inside
An Anritsu µOTDR reflectometer keeps a list of 30 fibre types inside it: the operator picks the fibre, and the instrument fills in the index of refraction on its own. That list is transcribed below in full, along with what is not in it. It is a transcript of somebody else's table, not our list of recommendations: the right number for your line is in your cable datasheet.
Single-mode
| Fibre | 1310 nm | 1550 nm | 1625 nm | 1650 nm |
|---|---|---|---|---|
| Alcatel ESF | 1.46400 | 1.4645 | — | — |
| Alcatel SF | 1.46400 | 1.4645 | — | — |
| Alcatel Teralight | — | 1.4645 | — | — |
| Alcatel TL Metro | 1.4690 | 1.4692 | — | — |
| Alcatel TL Ultra | — | 1.4692 | — | — |
| AT&T TrueWave 95 | 1.4738 | 1.4732 | — | — |
| Corning LEAF | 1.46400 | 1.4690 | 1.4690 | — |
| Corning SMF-28/e | 1.4677 | 1.4682 | 1.4685 | — |
| Corning SUB SMF-ls | — | 1.4700 | — | — |
| Lucent AllWave | 1.4660 | 1.4670 | — | — |
| Lucent TW 97 | 1.4710 | 1.4700 | — | — |
| Lucent TW RS 98 | 1.4710 | 1.4700 | 1.4700 | — |
| Pirelli | — | 1.4700 | — | — |
| SIECOR/Corning SMF-28e | 1.4677 | 1.4682 | — | — |
| Sumitomo PureGuide | — | 1.4700 | — | — |
| Sumitomo SM OF SE-3 | 1.4660 | 1.4670 | — | — |
| G.652 | 1.4670 | 1.4675 | — | 1.4678 |
| G.653 | — | 1.4666 | — | — |
| G.655 | — | 1.4680 | — | — |
| IOR = 1.48 | 1.4800 | 1.4800 | 1.4800 | — |
A dash is not a zero: at that wavelength the vendor gives no value for that fibre at all. The long-wavelength columns are filled in for a handful of fibres only, which is why they stand empty; nothing was lost in the transcribing.
Multimode
| Fibre | 850 nm | 1300 nm | 1550 nm |
|---|---|---|---|
| Alcatel MM 50/125 | 1.4820 | 1.4800 | — |
| Alcatel MM GL62.5/125 | 1.4970 | 1.4920 | — |
| Alcatel MM 62.5/125 | 1.4970 | 1.4920 | — |
| Corning 50/125 | 1.4900 | 1.4860 | — |
| Corning 62.5/125 | 1.4960 | 1.4910 | — |
| Corning Inf300mm 62.5 | 1.4960 | 1.4910 | — |
| Corning Inf600mm 50 | 1.4900 | 1.4910 | — |
| Corning Inf1000mm 62.5 | 1.4960 | 1.4910 | — |
| Corning Inf2000mm 50 | 1.4900 | 1.4860 | — |
| Corning Metro Core | — | — | 1.4690 |
The vendor files "Corning Metro Core" under multimode and gives it one single value, at 1550 nm. That is how it stands in their file; we transcribe the table rather than correct it.
The second figure in the same table
Next to the index of refraction the same file holds the backscatter coefficient, between −82.5 and −67.0 dB across the rows where the vendor filled it in. It has no bearing on distance whatsoever: it is the share of light that comes back, and it sets the level the whole curve sits at. It has a field of its own in the file; the app prints it under settings as Backscatter and lets you edit it. It is deliberately not a column above: the figures would double, and what people come to this page for is length. The difference between the coefficients of two spliced fibres is exactly what averaging two opposite measurements takes out.
What recomputing does
It does not change the shape of the curve, only its distance axis: every point (the events, the fibre end, the cursors) moves proportionally. Levels in decibels stay as they were. The original bytes of the measurement are kept whole, and the recalculation does not overwrite them.
What we do not fill in for you
There is no default index of refraction in the app, just as there are no acceptance thresholds. A plausible number in a field would read as a recommendation, and the correct value is in your cable's datasheet, and we have no way of knowing it. The recomputation is done with your number, and by hand.
The table above does not undo that: it says what somebody else's instrument sets for itself, and not one of its numbers reaches your file on its own.
Next
Optical length and the cable's physical length are different quantities: fibre is laid in a cable with slack, and the stranding factor comes from the datasheet, not from the trace.