340 m Hospital Backbone: OM3 Proposal Under Review
A 1970s hospital pavilion is adding a 10GBASE-SR backbone from the main cross-connect to a new imaging suite. The route is 340 m through existing 1.25 in conduit with two sweeping 90° bends and one older 90° pull elbow. The existing cable is 62.5/125 µm OM1 installed in 1998, with an orange jacket and no documentation. The IT team proposes replacing it with OM3 because 'OM3 is rated for 10G.' The link will serve a 48-port switch stack, and the longest reach is the full 340 m. The electrical contractor can open one wall but cannot enlarge the conduit without a fire-rated penetration change and a two-week inspection. The hospital requires a 15-year life and may later add 40GBASE-SR4. The project manager asks you to review the fiber choice, the pathway constraint, and the interim option of running 1000BASE-SX on the existing OM1 while the new cable is procured.
1. Evaluate the OM3 proposal and recommend a fiber type. Support the recommendation with reach limits, a loss-budget check, and at least one requirement beyond immediate 10G operation.
Read the worked answer
Reach check at 850 nm: 10GBASE-SR is specified to 300 m on OM3 and 400 m on OM4. The 340 m route therefore fails OM3 outright and leaves only 60 m of margin on OM4. Future check: 40GBASE-SR4 is 150 m on OM4, so an OM4 backbone would not support the stated future upgrade on these runs. Loss check: the 10GBASE-SR channel budget is 2.6 dB. Four LC connectors at the 0.75 dB allowance already consume 3.0 dB before fiber and any splices, so a 340 m channel with multiple panels is not a comfortable design even if the distance table is met. Recommendation: specify OS2 single-mode. 10GBASE-LR reaches 10 km over OS2 with a 6.2 dB budget, and 40GBASE-LR4 and 100GBASE-LR4 also reach 10 km. Use LC UPC unless the selected future optics require APC, and test at 1310/1550 nm. OM3 is not a valid substitution at 340 m. OM4's 400 m rating is a maximum, not a design target when four connectors, an aged pathway, and a 15-year life are involved. The decision should be driven by reach, loss margin, and future optics, not by the fact that a cable is labelled '10G.'
2. Can the existing OM1 support 1000BASE-SX at 340 m? What would you tell the project manager about using it as an interim link, and what measurements would you require before relying on it?
Read the worked answer
Do not treat the existing OM1 as a 340 m 1000BASE-SX link without verification. For 62.5/125 µm multimode, 1000BASE-SX reach is typically 220 m for 160 MHz·km fiber and 275 m for 200 MHz·km fiber. At 340 m, the link is beyond both typical ratings, so 1000BASE-SX should be expected to fail. The fact that an old cable 'worked at 100M' says nothing about gigabit operation. If an interim link is essential, consider 1000BASE-LX at 1310 nm, which has a longer reach on 62.5 µm fiber, but confirm the fiber's bandwidth and use the mode-conditioning patch cord required by the optics vendor. Before relying on the old cable, require: endface inspection and cleaning of every connector; an OLTS insertion-loss test at the actual operating wavelength; a bidirectional OTDR trace to locate bends, tight ties, and water damage; and a documented loss budget with margin. Attenuation testing cannot prove modal bandwidth, so even a passing loss test does not guarantee 1000BASE-SX at 340 m. Present the interim link as a temporary risk with a defined cutover date, not as an equivalent replacement.
3. The contractor says the existing conduit has 38% fill and can accept a 12-strand OS2 cable. What installation and documentation conditions would you attach to that plan?
Read the worked answer
Accept the plan only with verifiable conditions. First, confirm fill by calculation using the actual cable outside diameter and the conduit inside diameter, not by a verbal percentage; include the existing cable and any innerduct. Second, require a pre-installation mandrel and swab of the conduit, and replace the older 90° pull elbow with a sweeping bend if the cable manufacturer's minimum bend radius cannot be maintained. Third, set a pulling-tension limit at or below the cable's short-term rating, use a breakaway swivel and a dynamometer, and stop the pull if tension rises abnormally. Fourth, require compatible lubricant, a pulling eye that grips the strength members rather than the jacket, and firestopping at the rated penetration. In the telecom rooms, support the cable at the manufacturer's interval, use Velcro rather than narrow nylon ties, and maintain bend radius at slack loops. Documentation should include the as-built route, cable reel and lot numbers, connector and splice loss results at 1310 and 1550 nm, bidirectional OTDR traces, endface inspection records, TIA-606 labeling, and a warranty package. Do not accept '38% fill' as sufficient evidence that the aged pathway is suitable.