OTDR Studies First look with SM OTDR failed

OTDR Studies • First look with SM OTDR failed: not enough sensitivity to see RBS signal. • Nice clear signals visible with MM OTDR (on free loan until 11/05/07). • Next slides show typical traces – RBS signal clearly visible above noise floor can measure attenuation length of fibre in cable with laser at 856 nm – Higher resolution with 665 nm laser allows clear separation of the Fresnel peaks from the MT-12 at PPB 1 and the p-i-n diode on the dogleg. 1

OTDR Measurements APC ST MT MT MM v OTDR 3. 2 m or 2 m 5 m MT p-i-n diode ~ 2 m FUT (Fibre under test) 2

Connectors at start of cable and fan-out RBS Signal above background MT-12 at PPB 1 p-i-n diode on dogleg 3

Use delay setting to avoid peaks from start of cable & fan-out 4

Use 665 nm laser for higher resolution to see loss at MT 12 at PPB 1 5

Use 856 nm laser to measure attenuation length of fibre in cable 6

Measurements • Measured attenuation length of fibres in cable and loss at MT-12 at PPB 1 for – All fibres in 4 ribbons associated with “problematic” channels, all channels have < 90 m. A p-i-n current. – All fibres in 4 ribbons associated with “good” channels. All channels have >= 90 m. A p-i-n current. • Also checked that there was no excess loss near PPB 1 due to micro-bending. • Measured attenuation length of 4 ribbons in spare cable. 7

Summary • Attenuation lengths – <good channels> = 15. 1 +/- 0. 45 d. B/km – <bad channels> = 17. 1 +/- 0. 6 d. B/km – <fibres spare cable> = 8. 24 +/- 0. 28 d. B/km • Loss at MT-12 PPB 1 – <good channels> = 0. 13 +/- 0. 02 d. B – <bad channels> = 0. 17 +/- 0. 03 d. B 8

Discussion • There are systematic errors on the absolute values of these measurements but the relative measurements of good and problematic channels should be meaningful. • Losses at PPB 1 appear to be very small and no significant difference between problematic and good channels. • The attenuation length for problematic channels is slightly longer than for good channels. Marginally statistically significant (2. 7 s) • Attenuation length for fibes in spare cables is significantly longer. 9

Averages for ribbons • Look at average attenuation lengths for each ribbon • Good Crate/slot/Tx • Problematic • Spare Attenuation d. B/km Error 080 10. 55 0. 35 2 14 1 15. 06 0. 54 171 17. 58 0. 32 3 15 0 17. 14 0. 27 Crate/slot/Tx Attenuation d. B/km Error 060 15. 38 0. 60 070 12. 71 1. 34 2 12 0 18. 83 0. 15 2 12 1 21. 63 0. 50 10

Conclusions • Some evidence for non-statistical ribbon to ribbon variations. • Difference between “good” and “problematic” ribbons small, even for worst ribbon – extra attenuation compared to good ribbons is only ~ 0. 6 d. B. – For this ribbon <Expected/measured pin current> = 5. 9 d. B • Difference between fibre in spare cable and fibres in installed cables is much larger. • Different mode coupling for fibres in installed cables compared to fibres in spare cable ? ? ? • Still need “better” VCSELs … 11