Arena DAS: OM4 Bid Fails on Reach and Loss
A 22,000-seat arena is adding a neutral-host digital DAS. The BTS hotel sits in the north service corridor, and 36 remote units (RUs) serve 12 sectors. Fiber home runs range from 90 m to 640 m; the longest passes through two telecom rooms with patch panels. Each RU needs a 10 Gb/s CPRI link and 1+1 fiber redundancy. The general contractor allows six weeks for cabling, and a sold-out concert is ten weeks out. The low bidder proposes OM4 for all 36 links, citing 10GBASE-SR optics at $120 versus $240 for 10GBASE-LR. Worst-case link: 640 m, six connectors, two splices. Fiber attenuation is 3.5 dB/km at 850 nm and 0.4 dB/km at 1310 nm; connector loss 0.3 dB, splice loss 0.1 dB. The 10GBASE-SR channel budget is 2.6 dB; 10GBASE-LR budget is 6.2 dB. The arena's pathways are 60% full.
1. Is OM4 acceptable for all 36 RU links? Show the worst-case loss calculation and state the maximum OM4 distance.
Read the worked answer
No. The OM4 proposal fails on both reach and loss.
- Reach: 10GBASE-SR is limited to 400 m on OM4. The 640 m run is 240 m beyond that before any loss calculation.
- Loss: fiber 0.64 km × 3.5 dB/km = 2.24 dB; connectors 6 × 0.3 = 1.80 dB; splices 2 × 0.1 = 0.20 dB; total 4.24 dB against a 2.6 dB channel budget.
Use OS2 for every RU, not just the long ones. At 1310 nm: 0.64 × 0.4 = 0.26 dB + 1.80 + 0.20 = 2.26 dB, leaving 3.94 dB against the 6.2 dB LR budget. Standardizing on OS2 also avoids two optic spares pools and prevents a future RU relocation from becoming a reach problem. The apparent optic savings from SR are small compared with the cost of a failed link or a second media type in the spares cabinet.
2. Recommend a fiber count and topology that satisfy 1+1 redundancy, and explain the operational trade-off of star versus ring.
Read the worked answer
Provide four fibers per RU: two working and two protect, giving 144 fibers total. Build this as six 24-strand trunks or twelve 12-strand trunks, terminated at the headend on LC panels. Where pathways allow, route the A and B fibers through different cable trays or risers.
Use a star, point-to-point topology from the BTS hotel to each RU. CPRI is a point-to-point link, and star keeps latency and jitter predictable, simplifies timing, and isolates a single RU failure. A ring can save home-run fiber and provide path diversity, but it usually means daisy-chaining RUs or adding add/drop nodes. In a DAS, a failed RU or a ring break can then affect downstream sectors, and troubleshooting becomes harder.
The trade-off is fiber count and headend rack space versus operational simplicity. For 36 RUs, the extra strands are cheap; a sector outage during a concert is not.
3. The cabling window is six weeks. Identify the schedule constraint and propose a work breakdown that protects the concert date.
Read the worked answer
The constraint is not pulling cable; it is terminating and testing 144 fibers in six weeks while the headend and RU power are still being completed. At roughly 24 terminations per day, field-install connectors would consume the window and leave no float for retests.
Proposed breakdown:
- Weeks 1–2: install pathways, pull pre-terminated trunk cables, and land the headend panels. Pre-terminated LC harnesses move the loss-critical work to the factory.
- Weeks 3–4: pull and dress RU drops, label, and complete sector-by-sector continuity checks.
- Weeks 5–6: bi-directional OTDR and power-meter testing, RU activation, and sector optimization.
The concert is ten weeks out, which leaves four weeks of float for RF optimization and retests. Protect that float by ordering 10% spare fiber and spare LR optics, and by refusing any change that consumes the last two weeks before the event.