Planning a fiber backbone riser in a multi-family building
A fiber backbone riser in a multi-family building is one of the few parts of the low-voltage package that cannot be redone a piece at a time. A suite drop can be re-pulled; a slab penetration cannot be moved once the concrete cures. Most riser problems that surface in year eight trace back to a decision made in design development.
Copper's 100 m limit is what forces a distributed topology
Balanced twisted-pair horizontal cabling is limited to a 100 m channel, and that single number decides where the telecom rooms go. The working budget is about 90 m of fixed cable plus patch cords at each end, measured along the real route — up the wall, across the corridor ceiling, around the shaft — not off the floor plan.
That is why one room in the parkade cannot serve a tower. The building gets a distributed topology: a main telecom room and floor rooms placed so every outlet is within copper reach, each linked back on fiber.
The number of floor rooms follows the floor plate, not the storey count. A compact plate may serve two or three floors from one room; a long-wing plate can need one on every floor. Measure the longest real route to the farthest outlet before a room count goes on a drawing.
Stack the rooms vertically so the riser is a straight shot; every offset adds a horizontal transition, more penetrations and more firestop detail. Those rooms also hold the floor switch and the PoE feeding corridor access points, cameras and door controllers, so size them for year ten.
What a riser actually is, and why pathway capacity is sized for the building's life
A riser is the pathway, not the cable in it, and the pathway is the part that cannot be added cheaply later. Physically it is sleeves or cored openings through each floor slab in the stacked telecom rooms, sometimes with conduit, sometimes a framed shaft with a tray. The cable is a consumable. The hole is permanent.
Size it for what the building will hold across decades and treat day-one cable as a small fraction of capacity. Several smaller sleeves beat one large one: building systems stay separated from carrier plant, and a sleeve can be left empty, capped and firestopped for a pull nobody has scoped yet.
Adding a penetration later in an occupied concrete building means structural review, scanning for post-tension tendons and rebar before coring, water control, noise and resident access coordination — none of it comparable to a sleeve set in formwork before the pour.
Plan the mechanical detail with it: support so the cable's weight lands on the structure rather than an enclosure, bend radius top and bottom, separation from electrical, and bonding of metallic pathway. Room sizing and rough-in sequencing are covered in the multi-family pre-wire guide.
Firestopping at every floor penetration
Each penetration through a rated floor assembly has to be closed with a firestop system tested for that assembly, sleeve, penetrant and fill — not with whatever sealant is on the truck. The tested system is the whole combination, so changing the cable type or the fill can put the seal outside what was tested.
Re-enterable firestop devices earn their price wherever cable will be added later: the installer pulls new cable through and the device reseals, instead of cutting out a cured seal.
Assign the responsibility explicitly, because several trades pass through the same riser rooms. What is acceptable, who inspects it and what records are required all vary, so confirm requirements and accepted systems with the authority having jurisdiction before rough-in.
Singlemode or multimode fiber in a new building
For a new multi-family backbone, OS2 singlemode is the low-regret specification, and the case for multimode has narrowed to conditions most residential buildings do not have. Cable prices are close, and multimode's old advantage was cheaper short-reach transceivers — a gap that narrowed as singlemode optics at 10 Gb/s and above became commodity parts. Neither fiber runs out of distance in a riser of tens of metres; the question is whether the cable will constrain the switch bought fifteen years from now.
| OS2 singlemode | OM4 multimode | |
|---|---|---|
| Reach at 10 Gb/s | Kilometres | ~400 m |
| Reach at 40–100 Gb/s | Kilometres | Roughly 100–150 m, usually parallel optics |
| What limits an upgrade | The transceiver | The cable |
| Best fit | Building and campus backbones | Data centre rows, short high-count links |
Distances are approximate and vary with fiber grade and transceiver.
Multimode still answers one case: extending an existing multimode plant in a retrofit, where matching the installed base avoids conversion at every floor. Hedging with a hybrid cable carrying both types usually costs more on a new riser than singlemode with spare strands.
Fix connector type and polish per segment on the drawings, not on site. LC duplex is the default for a building backbone, MPO where pre-terminated trunks or parallel optics are used. Carrier and PON equipment often expects APC connectors while building equipment is usually UPC; mixing them costs loss and can damage endfaces.
Sizing the fiber backbone riser in a multi-family building
Spare strands are the cheapest insurance in the project: the glass costs less than the labour to pull it, and far less than pulling a second cable up a finished building. Count what the design needs on day one — an uplink to each floor room, any second path, building systems, the carrier hand-off — then buy the next standard count up.
Home-run each floor room back to the main room wherever the pathway allows. Independent paths mean a fault or a change on one floor does not touch another. A daisy chain with a splice at each floor uses less cable but couples the floors together.
Keep building systems and resident-facing service on separate strands. Cameras, access control, elevator and mechanical traffic should not share one uplink with a service network somebody else administers. Assign strands on the riser diagram and mark the spares — a spare nobody can identify is not a spare.
Terminating every strand costs a little more and makes a spare usable with a patch cord; strands left coiled in the enclosure need a technician and a fusion splicer before they carry anything. Terminate what is plausibly needed within a few years and document the rest. How we scope and document backbones is on the structured cabling page.
Splicing or pre-terminated assemblies
Pre-terminated assemblies buy schedule and factory test data; field termination buys tolerance for a building that is never exactly what the drawings said.
A pre-terminated trunk arrives tested and fitted with a pulling grip, and installs quickly with less specialised labour. It needs accurate routed lengths early, including slack for the enclosure and floor offsets: too short is scrap, too long is a coil that must be stored at full bend radius. Check that the pulling grip and connector end fit the sleeves and bends before ordering.
Field termination means pulling bulk cable and fusion splicing pigtails into an enclosure at each end. It is forgiving on length, which suits retrofits and routes that cannot be measured until late. Each splice adds a small loss and shows as an event on an OTDR trace.
Carrier entrance and demarcation: bring the providers in at design
Where a provider's network ends and the building's begins is a design decision with a lead time, and it is routinely settled late. The entrance facility is where the provider's plant terminates and the building's backbone starts. It needs a pathway from the property line — usually dedicated conduit with a pull rope, sized to the provider's requirement — plus space, power and bonding.
Start the applications early. Provider design and service orders can run months, and the entrance conduit has to be in the ground before backfill and hardscaping. More than one provider will usually want to serve a new building in Metro Vancouver and the Fraser Valley, which normally means separate pathways and space.
The distribution model changes the strand count, so settle it before the riser is sized. Some providers run their own fiber to every suite and use only the building's pathway; others hand off at the demarcation and the building's backbone carries service upward. Establish in writing where the demarcation sits, who maintains each side, and how a technician gets into a locked room after hours.
Labelling and test documentation at handover
A backbone that is not labelled and tested is one the next contractor re-traces with a light source, on the owner's time. Use a consistent scheme identifying room, enclosure, cassette and port, with the same identifier at both ends of every strand. Document the scheme itself; the logic matters as much as the labels.
Test insertion loss on every strand with a light source and power meter, against a loss budget calculated from length, connector count and splice count rather than a generic pass mark. On a backbone, ask for OTDR traces as well: they characterise each splice and connector and give a baseline for future troubleshooting.
Specify the handover package in the tender, because it is rarely produced if it was never priced — test results in native format plus readable PDF, a riser diagram showing sleeve locations, routes and strand assignments including spares, splice records, connector type and polish per segment, and firestop locations.
What to do next
Fix the telecom room locations and their vertical alignment before the structural set is issued, then size the sleeves for spare capacity rather than day-one fill. Specify OS2 singlemode, a strand count above what the design needs, and connector type and polish per segment. Open the carrier applications early enough that the entrance conduit goes in with the underground work, and write labelling, test results and an as-built riser diagram into the scope rather than leaving them for handover.
Orbit Automation is a licensed and insured low-voltage integrator in Surrey serving Metro Vancouver and the Fraser Valley, and can review a riser design, telecom rooms and backbone specification on site.