1×128 PLC Fiber Optic Splitter Guide: 23.5 dB Loss Budget, Real Deployments & When It Beats 1×64

I’ve recommended a 1×128 PLC splitter exactly three times in ten years.

Not because the product is bad — a well-made single-chip 1×128 is an impressive piece of planar waveguide engineering. But because most engineers who ask for one are trying to solve a problem that a pair of 1x64s handles better, with more margin and less risk.

The times I did recommend it:

• A 58-story residential tower in Dubai — 232 units, two XGS-PON ports, two 1x128s in a basement FDH
• A Singapore research campus — 128 lab buildings across 2 km², one OLT port, one splitter
• A Nairobi cable TV headend — 1550 nm broadcast to 128 neighborhood nodes via a single EDFA

Three very specific architectures. Three very tight link budgets. Three deployments where anything less than 128-way splitting would have meant adding OLT ports, rack space, and cost.

If your project doesn’t resemble one of these three, you probably don’t need a 1×128. But if it does — or if you’re designing for NG-PON2 and need to understand the practical ceiling of passive optical splitting — this guide covers everything I’ve learned.

Quick check — before you read further: Send your PON standard and feeder length to the loss budget table. If your numbers land in the ❌ column, skip straight to the decision framework for your alternatives. If they land in ✅, keep reading.

1. What Is a 1×128 PLC Splitter — and What’s the Physical Limit of Splitting?

A 1×128 PLC splitter takes one optical input and divides it across 128 output ports using a planar lightwave circuit chip — a silica-on-silicon wafer with waveguides etched via photolithography. One fiber in, 128 fibers out. Each output gets roughly 1/128th of the input power.

1×128 splitters are single-mode only. The PLC waveguide is designed for single-mode transmission at 1260–1650 nm (G.652.D or G.657.A2 fiber). There is no standard commercial multimode 1×128 splitter — multimode PLC splitters exist as specialty/custom items only up to ~1×16, and are rarely stocked due to modal noise accumulation at higher port counts. If you’re running multimode backbone (OM3/OM4/OM5), a 1×128 splitter isn’t part of your architecture — multimode PON splitting is not deployed beyond 1×16 in practice.

How the PLC Chip Works: Seven Stages to 128 Ports

The splitting happens in seven cascaded 1×2 Y-branch stages inside the chip: 1→2→4→8→16→32→64→128. Each stage is a waveguide junction where the optical power splits evenly in two directions. Seven stages accumulate 1.5–2.4 dB of excess loss above the theoretical minimum of 21.07 dB (10 × log₁₀(128)).

FBT (Fused Biconical Taper) technology is limited in practice — twisting and fusing fibers produces reliable single-stage splitters up to 1×8. Cascade FBT configurations (multiple 1×2 stages) can theoretically reach 1×16 or even 1×32, but optical performance degrades steeply compared to PLC at ratios above 1×8 — higher insertion loss, worse uniformity, and narrower wavelength range. For 1×128, the question doesn’t arise: virtually every commercially available 1×128 splitter uses PLC technology. All ratios above 1×8 — 1×16 through 1×128 — are effectively PLC-exclusive for any deployment you’d trust.

Why 1×128 Is the Practical Ceiling (Not 1×256)

1×128 isn’t the physical limit of the chip — 256-port and even 512-port PLC wafers have been demonstrated in labs. It’s the practical limit because the insertion loss at higher ratios leaves nothing usable for fiber plant. At 1×128, you’re already consuming 22.5–23.5 dB before connecting a single meter of fiber. A 1×256 splitter at ~25 dB theoretical minimum leaves less than 7 dB for your entire fiber plant — a lab curiosity, not something you’d deploy.

Long-term reliability data reinforces this ceiling. ITU-T G.Sup39 (Optical component reliability) and accumulated field data from 15+ years of GPON deployments show PLC splitters have a median failure rate below 0.1% per port per year when properly manufactured — making them among the most reliable elements in a PON. But failure rate increases with port count due to higher waveguide density and tighter tolerances. A 1×128 has roughly 2–3× the infant-mortality risk of a 1×32 from the same production line. Per-port testing is mandatory on 1×128 because you’re screening for the ~1–2% of ports that could pass a spot check but fail under full-spectrum measurement. If your supplier does spot-sampling instead of per-port testing on 1×128, that failure risk migrates from the factory QC bench to your network.

2. Can Your PON Budget Handle 23.5 dB? The Complete Loss Budget Table

A 1×128 splitter consumes 23.5 dB of your optical power budget before you’ve plugged in a single patch cord. Whether this is workable depends on your PON standard, your ODN class, and your fiber plant length.

The Quick Answer

The Full Truth Table

ODN Classes N1 (≤29 dB), N2 (≤31 dB), E1 (≤33 dB), and E2 (≤35 dB) are defined in ITU-T G.989.2 Amendment 1. XGS-PON’s 32 dB budget is specified in ITU-T G.9807.1.

GPON B+ with 1×128? Forget it. 4.3 km of total fiber plant including connector pairs (0.3 dB each per IEC 61753-1) and splices isn’t a deployment — it’s a lab bench. I’ve worked on rural builds where the feeder alone ran 8 km. If you’re locked into GPON B+, the link budget math eliminates 1×128 as an option. See the Wrong Tool checklist for your alternatives.

32 dB budgets are the minimum viable threshold. At 15.7 km total reach, a typical 5 km feeder + 1 km drops + connector losses leaves 2–3 dB of breathing room. Most urban and suburban deployments fit.

35 dB budgets are the sweet spot. 8.5 dB for fiber plant covers 20+ km of reach, or shorter distances with generous margin for aging, repairs, and coexistence.

The Coexistence Element Tax: 2 dB You Didn’t Budget For

When GPON, XGS-PON, and NG-PON2 share the same ODN — a common migration path — you need a coexistence element (CEx) , a WDM filter that combines/separates wavelengths per ITU-T G.989.2. It adds ~2 dB of insertion loss on the NG-PON2 path.

Still workable but it cuts 6 km off your reach. Calix’s NG-PON2 deployment white papers document operators using Raman amplification (pumped at 1535 nm via a dedicated pump laser in the OLT shelf) to recover margin in coexistence scenarios — effective, but it adds active equipment cost and complexity. This is an option only for carrier-grade deployments with the operational capability to manage active amplification in the outside plant.

The Loss Budget Calculator: 6 Numbers That Tell You If 1×128 Works

Skip the full truth table. Copy these six fields into a spreadsheet, a calculator, or a napkin. If your spreadsheet gives you a positive number in Step 4, your link budget passes. If it’s negative, see the alternatives in the decision framework.

Step 1: Your total PON budget
| Your PON Standard | Enter Budget (dB) |
|—|—|
| GPON B+ | 28 |
| GPON C+ / XGS-PON | 32 |
| XGS-PON Extended / NG-PON2 | 35 |
| 50G-PON | 29–31 |

→ A = ____ dB

Step 2: Subtract known losses

A (your PON budget):           ____ dB
Minus 1x128 splitter (max):   - 23.5 dB
Minus safety margin:           -  3.0 dB
Minus coexistence (CEx):       -  2.0 dB (enter 0 if single-PON)

→ B = ____ dB (remaining for fiber plant)

Step 3: Calculate your fiber plant loss

Feeder length (km)      ____ × 0.35 dB/km = ____ dB
Drop length (km)        ____ × 0.35 dB/km = ____ dB
Connector pairs         ____ × 0.3 dB     = ____ dB
Fusion splices          ____ × 0.1 dB     = ____ dB

→ C = ____ dB (total fiber plant loss)

Step 4: Your remaining margin

B (remaining budget)        ____ dB
Minus C (fiber plant loss)  - ____ dB

→ Margin = ____ dB

How to read the result:
– Margin ≥ 3.0 dB: Comfortable. Proceed to supplier evaluation.
– Margin 1.5–2.9 dB: Tight. Use conservatively budgeted worst-port IL (not typical). Consider premium single-chip for the extra 0.5 dB.
– Margin 0.1–1.4 dB: Too tight. Recalculate with actual fiber route distances, not map distances. If still <1.5 dB, spec 1×64.
– Margin ≤ 0 dB: Budget exceeded. Spec 1×64 or lower. Do not spec 1×128.

Real example (Dubai Case 1 from the deployments section):

A = 32 dB (XGS-PON)
B = 32 - 23.5 - 3.0 - 0 = 5.5 dB
C = (0.3 km feeder × 0.35) + (0.0 km drop × 0.35) + (4 × 0.3) + (1 × 0.1) = 1.8 dB
Margin = 5.5 - 1.8 = 3.7 dB → ✅ Comfortable

(The original case used measured IL of 23.2 dB, giving 4.0 dB margin. The calculator uses the conservative 23.5 dB max — your actual margin depends on your specific unit’s measured IL.)

Want us to check your numbers? Send the six values from this calculator to [email protected] with subject line “1×128 link budget check”. We’ll run the calculation on our end and tell you what we see. Takes 5 minutes, no strings attached.

Why Uniformity Matters More Than Typical Insertion Loss

Design for the weakest port, not the average. Uniformity is the maximum insertion loss difference between any two output ports per GR-1209-CORE — expressed as max-min, not ±. A standard-grade single-chip 1×128 has uniformity ≤2.0 dB. If your typical IL is 22.5 dB and the distribution is roughly symmetric, your best port lands near 21.5 dB and your worst near 23.5 dB. For conservative link budgeting, assume the worst port could reach typical IL + the full uniformity range (22.5 + 2.0 = 24.5 dB) — because you can’t guarantee symmetry.

Each 0.5 dB of tighter uniformity buys ~1.4 km of additional fiber (0.5 ÷ 0.35 dB/km), or one more connector pair. When your budget is tight, pay for better uniformity — not better typical IL. The symmetric estimate is what you’ll likely measure. The conservative budget is what you use when you can’t afford to be wrong.

3. Single-Chip vs Spliced 1×128: The Manufacturing Decision That Changes Your Link Budget by 1.5 dB

Not all 1×128 PLC splitters are made the same way. The manufacturing approach — single monolithic chip vs two spliced 1×64 chips — changes insertion loss by up to 1.5 dB and uniformity by up to 0.8 dB.

How Each Type Is Made

Single-chip (monolithic): One silica wafer contains all seven splitting stages — a continuous waveguide circuit with zero internal splices. Manufacturing yield is lower: if any of the 128 output waveguides is defective, the entire chip is scrapped.

Spliced (cascaded): Two pre-tested 1×64 PLC chips are fused together internally. Yield is higher because both chips are binned before splicing. But the splice adds loss, and the two chips’ individual uniformity variations stack rather than average out.

Performance Difference

The 1.5 dB difference equals 4.3 km of fiber on XGS-PON, or 4.3 km on NG-PON2. If your deployment is anywhere near the edge of the budget, ask your supplier: “Is this single-chip or spliced?”

How to Verify What Your Supplier Ships

Request the per-port test report. A spliced unit will typically show uniformity ≥2.5 dB spread, and a “step” in the loss distribution — ports 1–64 and 65–128 clustering at different levels because they originated from different chips. A single-chip unit shows a smooth loss curve across all 128 ports.

BWNFiber uses single-chip design exclusively for 1×128. No internal splices, no stacking two 1x64s. The datasheet below reflects this.

Competitive Specification Cross-Reference: 1×128 PLC Splitter Prices & Specs

Here’s how the major suppliers’ publicly listed specs compare as of June 2026. The 2.5 dB range in max insertion loss is real — it reflects different manufacturing approaches and spec philosophies:

† Fiber-Mart offers both spliced and single-chip 1×128; pricing differs. Verify chip type when quoting.

How to read this table: Lower max IL and tighter uniformity are better. But two caveats: (1) CommScope’s 26.0 dB is their EOL worst-case number including 20 years of aging per GR-1221 — they’re spec’ing what you’ll measure in 2046, not what leaves the factory. Most other suppliers quote T0 (fresh from QC). (2) A “24.0 dB” spliced splitter that isn’t per-port tested may have 4–5 ports at 25+ dB — they just weren’t the ports that got sampled.

The buyer’s checklist: ask every supplier three questions — single-chip or spliced? T0 or EOL spec? Per-port test report or sample only?

4. 1×128 Splitter Price & Total Cost of Ownership

Before you look at unit price, understand what a 1×128 actually costs over its life. The cheapest option is rarely the cheapest to own.

1×128 Splitter Price Range (June 2026)

Bulk pricing (MOQ 10+) is typically 30–50% lower. Premium single-chip costs 25–40% more than spliced because of manufacturing yield, not margin.

The Real Cost of a 1×128 Splitter (TCO, Not Unit Price)

A $90 bare-fiber splitter vs a $350 connectorized single-chip unit — the $260 difference feels significant on a PO. But the TCO math works differently:

The spread overlaps. The low-cost option can be cheaper — if your deployment is indoors, in a dry climate, without multi-PON coexistence, and you have in-house testing capability. The premium option’s TCO is more predictable. The low-cost option’s TCO has a wider range because the tail risk (early port failures in a humid deployment) is real and expensive.

The procurement rule I use: the cheaper the splitter, the more you need to budget for verification and replacement. The premium splitter’s TCO advantage isn’t the unit cost — it’s the narrower uncertainty band.

Need a project-specific TCO comparison? Send your quantity, deployment climate, and whether you have in-house test capability. We’ll model the 5-year cost for both spliced and single-chip options.

5. 1×128 PLC Splitter Technical Specifications (BWNFiber vs Industry)

We test every 1×128 splitter at 5 wavelengths (1310, 1490, 1550, 1577, 1625 nm) and ship the per-port test report with every unit. The table below puts our numbers next to what you’ll see from other manufacturers.

Optical Performance Parameters

Want the full datasheet? Download the BWNFiber 1×128 PLC splitter datasheet: insertion loss curves, uniformity distribution, WDL/PDL/TDL plots, package drawings, and connector options.

Download 1×128 Datasheet (PDF) →

Or request a de-identified per-port test report showing all 128 ports at 5 wavelengths — we’ll send one within one business day.

GR-1221-CORE reliability testing isn’t one test — it’s a battery. Here’s what “GR-1221 qualified” actually means for a 1×128 splitter, and why it matters for your deployment:

A splitter that passes thermal cycling but fails damp heat will work fine in Riyadh but fail in Jakarta within 18 months. When we see a competitor’s splitter priced 40% below ours, the first place we look for cost-cutting is the reliability test program — skipping damp heat or shortening thermal cycling saves manufacturing cost but guarantees field failures in humid climates.

I’ve personally investigated a field failure where 1×64 splitters from a low-cost supplier showed 1.5–2.0 dB of IL drift after two monsoon seasons in coastal India. The root cause: epoxy that passed dry heat but degraded under sustained humidity. The operator replaced 200+ splitters. The procurement savings were wiped out 3x over in truck rolls alone.

Deploying in a humid climate? Ask us for batch-specific damp heat test data (85°C/85% RH, 2,000 hours per GR-1221-CORE) before you order. We test every production batch, not just the qualification sample — and we’ll share the data.

Climate-specific 1×128 recommendations:

PDL and temperature — a non-linear problem most datasheets hide: PDL (Polarization Dependent Loss) is typically specified at room temperature. But PDL is temperature-dependent and the relationship isn’t linear — it can spike at temperature extremes where waveguide stress birefringence peaks. Standard-grade PDL ≤0.5 dB at 23°C can become 1.2+ dB at -40°C. Our premium-grade chip design minimizes waveguide stress asymmetry, keeping PDL below 0.5 dB across the full temperature range — but if you’re buying standard grade from any supplier, budget an extra 0.5 dB for PDL drift at temperature extremes.

Two Specs That Matter More Than the Datasheet Suggests

WDL (Wavelength Dependent Loss): A 1.5 dB WDL means port insertion loss varies by that much depending on wavelength. In a network running GPON (1490/1310 nm), XGS-PON (1577/1270 nm), and NG-PON2 (L-band) simultaneously — the port that’s fine at 1490 nm might be marginal at 1577 nm. Five-wavelength testing isn’t overkill; it’s the minimum for multi-PON coexistence.

What “Industry Range” actually means: CommScope specs their 1×128 at ≤26.0 dB — noticeably higher than our ≤23.5 dB. They’re quoting end-of-life worst-case including aging, temperature cycling, and humidity exposure per GR-1221-CORE reliability testing. Our ≤23.5 dB is the time-of-manufacture spec. Both numbers can be correct; they represent different confidence intervals. When comparing supplier specs, ask whether the number is T0 (fresh out of QC) or EOL (end-of-life worst-case).

How We Test vs What Most Suppliers Claim

Want to see what a real per-port test report looks like? Request a sample report at [email protected] — we’ll send you a de-identified 1×128 test sheet showing all 128 ports at 5 wavelengths. No obligation. It’s the fastest way to understand what you should be getting from any supplier.

Physical, Mechanical & Environmental

International certification by market: If you’re procuring for a specific region, the standard certifications required may differ. Below is what most tender documents ask for:

If your tender requires a certification not listed above, ask during RFQ — we maintain a library of third-party test reports and can support additional compliance documentation for project-specific requirements.

Export Packaging, Documentation & Logistics

For international buyers procuring from Chinese manufacturers, here’s what matters beyond the spec sheet:

Export packaging: 1×128 rack-mount trays ship in custom foam-lined flight cases rated for air and sea freight. Each output port has an individual dust cap and the fiber array is secured with releasable cable ties — not adhesive tape, which leaves residue. ABS box and LGX cassette packages ship in master cartons with 20 mm closed-cell foam lining and desiccant packs (10g silica gel per unit). We’ve shipped to 14 countries without a transport-damage claim.

Documentation supplied with every international shipment:
– Commercial invoice (with HS code 8517.62 — optical communication equipment)
– Packing list with serial numbers and port count
– Per-port test report (all 128 ports, 5 wavelengths, signed by QC engineer)
– Certificate of Conformity (GR-1209-CORE, GR-1221-CORE)
– RoHS 2.0 declaration
– Country of origin certificate (available on request; standard lead time + 2 working days)

Typical shipping lead times (from factory gate to destination port):
– Middle East / Gulf: 5–7 days air freight, 18–25 days sea freight
– Southeast Asia: 2–4 days air, 8–12 days sea
– Europe: 5–7 days air, 25–35 days sea
– Africa (major ports): 5–8 days air, 20–35 days sea
– South America: 6–9 days air, 30–40 days sea

Sea freight requires a Full Container Load (FCL) or Less than Container Load (LCL) arrangement unless ordering small quantities (≤20 units, which can ship via air freight courier). We coordinate with your freight forwarder or can recommend one if you don’t have a regular partner.

OEM/ODM for regional distributors: Available at MOQ 50 units. Options include:
– Custom branding (logo on tray faceplate, ABS box label, or LGX cassette)
– Custom packaging (your company’s box design, insert, and labeling)
– Custom test report format (your company letterhead, your internal part numbers)
– Regional-language documentation (Spanish, French, Arabic, Portuguese — technical translations reviewed by native-speaking engineers, not machine translation)

Regional procurement behaviors worth knowing:
– Middle East buyers typically request damp heat test data and dry heat aging data before price negotiation — send these proactively to shorten the evaluation cycle
– European buyers prioritize IEC standards over GR standards — reference IEC 61753-1 equivalency in your compliance documentation
– Southeast Asian government-funded FTTH projects often mandate local type-approval testing even if the product already carries international certifications — budget 4–6 weeks for this in your project timeline
– African ISP procurement often involves multiple layers of approval (technical → financial → board) — the per-port test report is the single most influential document in the technical review stage

Why the pigtail jacket matters for 1×128: Every output pigtail on a 1×128 splitter uses tight-buffered fiber (0.9 mm / 2.0 mm / 3.0 mm options), not loose-tube. Tight-buffer is needed because the individual fibers must be connectorized and handled individually in the rack — loose-tube cable requires breakout kits at every termination point, impossible at 128-port density. The jacket is LSZH (Low Smoke Zero Halogen) for indoor CO/headend deployment — building codes in most countries require halogen-free materials in occupied communications spaces (per IEC 60754-1/2 for halogen gas emission), with controlled smoke density (IEC 61034) and flame spread (IEC 60332-1). If you’re deploying the splitter outdoors in an FDH cabinet, confirm the enclosure provides UV protection — LSZH jacket alone does not provide long-term UV resistance, and prolonged UV exposure can cause jacket discoloration and embrittlement within 12–18 months.

TIA-568 structured cabling context: In a CO or headend, the 1×128 splitter sits in the Equipment Distribution Area (EDA) per TIA-568-C.0. The 128 output ports connect to the Horizontal Distribution Area (HDA) or directly to the Zone Distribution Area (ZDA) via fiber patch panels. The splitter itself is a passive device in the optical distribution network — it’s not a cross-connect — but the port numbering, labeling, and cable management should follow TIA-606-B administration standards. If your CO doesn’t have a coherent port labeling standard, 128 ports will become 128 mysteries within six months.

Ribbon Fiber Output: Two Factory Configurations

With 128 individual 0.9 mm fibers, cable management turns into a rat’s nest fast — especially during the first maintenance visit. Two ribbon options:

• 10 × 12-fiber ribbons + 8 individual fibers = 128 live outputs
• 11 × 12-fiber ribbons = 132 total (128 live + 4 dark spares for future activation)

Ribbon fiber enables mass-fusion splicing — terminating all 12 fibers in a ribbon in one splice cycle instead of twelve individual splices. When technician time is the dominant CO cost, ribbon pays for itself on the first move/add/change.

6. Where a 1×128 Splitter Actually Works: 3 Real Deployments

Case 1: Ultra-Dense Urban XGS-PON Tower — Dubai, 2024

What we faced: 58 floors, 4 units per floor = 232 residential units. The developer wanted fiber to every unit from a single basement communications room. Vertical riser space couldn’t physically accommodate 232 individual drop cables — a common constraint in towers built before FTTH mandates.

What we did: Two XGS-PON OLT ports, each feeding one 1×128 rack-mount splitter (single-chip, measured IL at 23.2 and 23.1 dB) in the basement FDH.

OLT Port 1 (XGS-PON 32 dB)
    │  50 m SC/APC patch
    ▼
Basement FDH
    ├─ 1x128 Splitter A (23.2 dB measured, ports 1–116 active)
    │    └─ 116 single fibers → vertical riser tray → floors 1–32
    │
    └─ 1x128 Splitter B (23.1 dB measured, ports 1–116 active)
         └─ 116 single fibers → vertical riser tray → floors 33–58

Link budget per subscriber: 23.2 dB (splitter) + 1.8 dB (300 m riser fiber at 0.35 dB/km + 4 connector pairs at 0.3 dB typical each per IEC 61300-3-34 + 1 fusion splice at ≤0.1 dB per IEC 61300-3-4) = 25.0 dB total. Remaining margin on XGS-PON: 32 – 25.0 – 3.0 (safety) = 4.0 dB.

What I’d do differently: Pre-terminate the 12 spare ports per splitter with pigtails coiled inside the tray. The developer added penthouse units a year later and we had to send a technician to terminate dark ports in a live FDH. Pre-terminating spares costs ~$15 per port at the factory vs $150–$300+ in the field (technician visit + splice + connector + OTDR verification). That ~$360 we “saved” cost us over $2,400.

Case 2: Campus Research Network on One OLT Port — Singapore, 2025

What we faced: 128 lab buildings across 2 km². The operator had one spare XGS-PON OLT slot. They couldn’t add another port without a full line card upgrade.

What we did: One OLT → 2 km campus backbone (G.652.D) → 1×128 splitter in a hub cabinet → individual fibers to each building’s ONU.

Link budget: 23.0 dB (splitter) + 1.2 dB (2 km fiber at 0.35 dB/km + 3 connector pairs at 0.3 dB typical each per IEC 61300-3-34) = 24.2 dB total. Margin: 32 – 24.2 – 3.0 = 4.8 dB.

The bandwidth reality — and why it worked anyway: 10 Gbps ÷ 128 buildings = 78 Mbps average. Sounds thin. But actual campus utilization data showed peak concurrent demand under 40% of buildings, with average usage below 100 Mbps per building during peak. Traffic was bursty — large simulation data transfers followed by hours of processing. When you’re moving a 50 GB dataset once and then computing for 3 hours, average throughput matters far less than peak transfer speed.

Research labs don’t stream 4K Netflix at 8 PM. If this were residential with 128 homes all streaming simultaneously, 78 Mbps average wouldn’t cut it. Traffic pattern matters as much as link budget. Both need to work for 1×128 to make sense.

Case 3: Cable TV RF Overlay at 1550 nm — Nairobi, 2023

What we faced: A cable operator distributing analog/digital video from one headend to 128 neighborhood optical nodes. Cable TV is broadcast-only (downstream, 1550 nm) — no upstream constraint. This changes the math entirely.

What we did: One 1550 nm EDFA (+10 dBm output) → 1×128 splitter → 128 distribution fibers across the city.

Link budget: +10 dBm – 23.5 dB (splitter) – 3.0 dB (15 km fiber at 0.20 dB/km at 1550 nm) = -16.5 dBm received. Against -20 dBm typical receiver sensitivity: 3.5 dB margin.

At 1550 nm, fiber attenuation is only 0.20 dB/km vs 0.35 dB/km at 1310 nm. Over 15 km, that saves 2.25 dB — the difference between a working node and a dark one. An EDFA outputs +7 to +13 dBm, far above a PON OLT’s +1.5 to +5 dBm. Video overlay works on 1×128 where PON wouldn’t — but only for downstream broadcast.

A Deployment I Stopped: When 1×128 Was the Wrong Answer

Not every near-miss becomes a deployment. In 2022, a Middle Eastern FTTH operator asked us to quote 1×128 splitters for a suburban GPON deployment — 25 km average reach, GPON B+ optics, outdoor pole-mounted closures. They’d seen a competitor’s marketing material claiming “1×128 ready” and assumed it applied to them.

I ran the numbers on the call: 28 dB GPON B+ budget minus 23.5 dB splitter = 4.5 dB left. Subtract 3 dB safety. At 0.35 dB/km, they’d have ~4 km of fiber plant. Their average feeder alone was 8 km. Plus, they wanted pole-mounted splitters — 128 pigtails in a street closure in 45°C summer heat, with dust and condensation.

I told them no. They pushed back. I sent the link budget calculation in writing. They went with 1x64s on a distributed splitting architecture — and the network’s been running for four years. Sometimes the best engineering decision is the sale you don’t make.

This is the scenario that plays out most often. Someone reads “1×128 PLC splitter” in a datasheet, assumes higher port count = better, and doesn’t check the budget math. Every time I write “no” to a 1×128 inquiry, I’m not losing a sale — I’m preventing a network that works on paper but fails in the field.

Got a link budget you want checked? If you’re unsure whether 1×128 fits your project, send the numbers — PON standard, feeder length, drop length, connector count — to [email protected]. I’ll run the calculation and tell you honestly if it works. This is not a sales pitch. It’s the same calculation I’ve used to say “no” more often than “yes.”

7. Where a 1×128 Is the Wrong Tool: A Pre-Spec Checklist

I’ve said “no” to more 1×128 requests than I’ve said “yes.” Before opening a datasheet, check these six items:

1. GPON B+ or ODN Class N1 (28 dB budget). ~4 km of usable fiber isn’t a network. Stay at 1×64 (≤20.5 dB max). The Loss Budget section has the full math.

2. Average fiber reach exceeds 8 km on a 32 dB budget. Use 1×32 (≤17.2 dB) or cascade 1×4 + 1×8 for distribution flexibility. Per the Fiber Broadband Association’s “FTTH Network Design: Splitter Architecture Best Practices” (FBA-050C FB101 Series), distributed splitting (1×4 in CO → 1×8 in field) is the recommended architecture for suburban and rural deployments — centralized splitting above 1×64 at the CO becomes power-budget constrained beyond 8 km of feeder reach.

3. Guaranteed bandwidth SLAs above 50 Mbps per subscriber. The bandwidth math is laid out in Case 2 above — 128 subs sharing 10 Gbps XGS-PON gives 78 Mbps average, and at 80% concurrency that drops to ~98 Mbps per active sub. Fine for residential best-effort. Not fine for business SLAs. For guaranteed bandwidth, use 1×32 (312 Mbps avg) or 1×64 (156 Mbps avg).

4. Cascade configurations. “Can I cascade two 1x64s?” 20.5 + 20.5 = 41.0 dB. No ITU-T PON standard supports 41 dB. A single 1×128 at 23.5 dB is your only practical path. For the correct parallel alternative, see the Decision Framework.

5. Pole-mounted or wall-box splitter locations. 128 pigtails don’t fit in a street closure. You need rack space, cable management, and climate control. If your architecture requires field-mounted splitters, use 1×32 or 1×64 at the field node and keep the 1×128 upstream.

6. Cannot get a per-port test report from the supplier. At 128 ports with ±2.0 dB uniformity, a 3-port sample has a 78% statistical chance of missing a marginal port. Per-port data at all operating wavelengths is table stakes. If the supplier won’t provide it, find another supplier.

7. Deploying ABS box 1×128 outdoors without IP68 protection. The ABS enclosure itself is not weatherproof — it’s designed for indoor or sheltered FDH cabinet use. If you’re deploying outdoors in an unmanaged enclosure, condensation cycles will degrade the epoxy bonds at fiber entry points within 18–24 months. Use IP68-rated splice closures with desiccant packs and pressure-equalization valves. Better yet, keep the 1×128 indoors and run distribution fibers to outdoor field nodes.

8. 1×128 Splitter Packaging, Connector Types & Cable Management for 128 Fibers

Package Options: Rack-Mount / ABS Box / LGX Cassette

1×128 splitters ship in three package types — not one, as some older guides claim:

The ABS box 1×128 is a standard catalog item from multiple manufacturers — available at prices ranging from ~$90 (bare fiber) to ~$565 (connectorized). The box routes all 128 pigtails out the side into an external splice tray. It’s not pole-mountable — the box is too large and 128 fibers need organized management — but it fits inside a fiber distribution hub cabinet. For field-mountable applications, use 1×32 or 1×64.

TIA-598-C Color Coding for 128-Port Organization

Without a system, 128 ports become unmanageable on the first maintenance visit. TIA-598-C color coding in banks of 16:

Every adapter position is laser-etched. Every fiber gets a matching numbered label at both ends. Troubleshooting a dark ONU on port 97 at midnight? Labeled fibers save hours.

Cable Management Rules (From Someone Who Learned the Hard Way)

• Vertical channels: ≥50 mm wide on both rack sides. 128 patch cords need volume — a 25 mm channel jams.
• Horizontal management bars: every 2U. Gravity works on fiber bundles; don’t skip intermediate supports.
• Service loops: ≥300 mm at both ends. Extra 0.9 mm G.657.A2 pigtail fiber costs ~$0.10–0.15 per meter. Re-terminating a connector with no slack costs a truck roll and a splice crew.
• Bend radius: G.657.A2 tolerates 7.5 mm minimum per ITU-T G.657 vs 30 mm for G.652.D. Design for 30 mm minimum — G.657.A2 is your safety net, not your design target.
• Never route 128 fibers through a single 25 mm ring. I untangled this exact mess — a contractor had stuffed a full 128-fiber bundle through one management ring. Replacing port 73 meant un-routing 55 fibers. Four hours.

9. How to Test All 128 Ports Without Spending 2 Hours With a Handheld Meter

Manual testing with a handheld power meter and notebook: ~120 minutes, ~5% transcription errors, 6–7 ports with wrong data.

Testing Methods Compared

Our production setup: automated optical switch → stabilized light source on the input → sequential port cycling → calibrated power meter → software-generated pass/fail report for all 128 ports at 5 wavelengths (1310, 1490, 1550, 1577, 1625 nm).

Our QC Acceptance Criteria for 1×128

• Per-port IL within ±0.5 dB of spec, measured at all 5 PON wavelengths per IEC 61300-3-7 (insertion loss test method for passive optical components)
• Uniformity ≤2.0 dB (standard) or ≤1.5 dB (premium) across all 128 ports
• Return loss ≥50 dB (APC) per IEC 61300-3-6, measured on every port
• Any single port exceeding max IL by >0.3 dB = unit rejected, not downgraded
• Per-port test report ships with every unit. Sample testing is statistically inadequate for 128 ports per IEC 61300-1 — which requires 100% testing when uniformity requirements are tighter than ±2.5 dB for devices with more than 64 ports

Field OTDR Testing With a 1×128: What Actually Works

Testing 128 ports with an OTDR in the field isn’t practical — you’re not hauling an OTDR through 128 ports at 3 AM during a maintenance window. The right field strategy depends on what you’re verifying:

Never use an OTDR from the OLT side without disconnecting the PON port first. The OTDR pulse will blind every ONU on that splitter for the duration of the test. If you’re testing in-service, use a PON OTDR at the monitoring wavelength (1625–1650 nm) through a dedicated test port on the WDM — this is what carrier-grade operators do, and it’s the only method that preserves live subscriber traffic.

10. Should You Spec a 1×128 or a 1×64? A Decision Framework

Most engineers I work with already know what a 1×128 is. What they actually need: “Should I spec it, or should I stay with 1×64?”

Five Questions Before You Spec

If your deployment satisfies ≥3 of these, a 1×128 may be viable. If not, stop and spec 1×64:

1. PON budget ≥32 dB? (XGS-PON C+ or higher — yes. GPON B+ — no.)

2. Total fiber distance (feeder + drops) under 15 km? (Urban/campus — yes. Suburban/rural — probably not.)

3. ≥100 subscribers concentrated in one physical location? (High-rise, campus, dense urban — yes. Scattered homes — no.)

4. No guaranteed bandwidth SLA above 50 Mbps per subscriber? (Best-effort residential — fine. Business SLA — no.)

5. Splitter location climate-controlled with rack space? (CO, headend, building basement — yes. Outdoor pole/wall — no.)

Decision Flowchart

Can your PON budget handle 23.5 dB splitter loss?
│
├─ NO → Spec 1x64 (≤20.5 dB). Stop.
│
└─ YES (≥32 dB)
    │
    ├─ Fiber distance >15 km?
    │   ├─ YES → Spec 1x64 or cascade 1x32 + 1x2. Stop.
    │   └─ NO → Continue.
    │
    ├─ ≥100 subs in one location?
    │   ├─ NO → Spec multiple 1x32s or 1x64s. Stop.
    │   └─ YES → Continue.
    │
    ├─ >50 Mbps guaranteed per sub?
    │   ├─ YES → Spec 1x32 (312 Mbps avg). Stop.
    │   └─ NO → Continue.
    │
    └─ Splitter indoors with rack space?
        ├─ NO → Spec 1x32 or 1x64 for field. Stop.
        └─ YES → 1x128 is viable. Proceed to supplier evaluation.

Landed on “1×128 is viable”? The next step is sending your link budget and deployment climate to a shortlist of suppliers for per-port test data review. The supplier evaluation framework and Pre-PO checklist below have the details. Or send the numbers to [email protected] — we’ll check them and tell you if we’re the right supplier for your project.

When to Spec 1×64 Instead

For most deployments where someone initially thinks they need 1×128, 1×64 is the right call — it gives you 3 dB more fiber budget (~8.5 km additional reach), commodity pricing and availability, and field-deployable form factors.

If you need more than 64 ports but can’t justify 1×128, the correct architecture is parallel splitting: a 1×2 primary splitter at the CO feeding two independent 1x64s — ~24 dB total loss per path vs 41 dB for a series cascade. Manageable on XGS-PON N2. The wrong way is series (one 1×64 feeding another 1×64) — that’s the 41 dB dead end. See the FAQ for the detailed explanation.

11. 1×128 Splitter FAQ: The Questions Engineers Actually Ask

Can I use a 1×128 PLC splitter with standard GPON?

No. GPON B+ (28 dB per ITU-T G.984.2) minus 23.5 dB leaves 4.5 dB — about 4 km of total fiber plant. That’s a proof-of-concept, not a deployment. GPON C+ (32 dB) makes it marginally viable at ~15 km. If you’re on GPON, 1×64 at 20.5 dB max is your practical ceiling.

What PON standard actually works with 1×128?

XGS-PON at 32 dB (ODN Class N2) gives ~15.7 km of fiber plant. NG-PON2 or XGS-PON Extended at 35 dB (ODN Class E1) gives ~24.3 km. Defined in ITU-T G.9807.1 and G.989.2 respectively. These are where 1×128 becomes comfortable, not marginal.

What’s the difference between 1×128 and 2×128 splitters?

A 2×128 contains two independent 1×128 PLC chips in one physical package — two input ports, each feeding 128 outputs. Used exclusively for PON port redundancy (Type B protection per ITU-T G.983.1) where two OLTs serve the same subscriber group. Insertion loss per path is identical to a 1×128; you’re paying for the second path and failover capability, not better optics. Unless you’re designing a protected PON architecture, you want 1×128.

What’s the difference between single-chip and spliced 1×128?

Single-chip: monolithic wafer with 7 splitting stages — ≤23.5 dB IL, ≤2.0 dB uniformity. Spliced: two 1×64 chips internally joined — ≤25.0 dB IL, ≤2.8 dB uniformity. The gap equals 4.3 km of fiber on XGS-PON. BWNFiber uses single-chip exclusively.

How much does a 1×128 fiber optic splitter cost?

Per unit: ~$90–$100 bare-fiber ABS box to $295–$565 connectorized SC/APC (Fiber-Mart, FS.com). Per port: $0.70–$4.40. Bulk pricing (MOQ 10+) is 30–50% lower. Premium single-chip costs 25–40% more than spliced — the difference is manufacturing yield, not margin. 1×64 comparison: $1.50–$3.00 per port connectorized.

Is 1×128 available in an ABS box package?

Yes. Multiple manufacturers supply 1×128 in ABS boxes (~140 × 115 × 18 mm) with pigtail outputs to an external splice tray. Fits FDH cabinets and splice enclosures, but not pole-mounted street closures. For field-mountable applications, use 1×32 or 1×64.

Can I cascade two 1×64 splitters instead of one 1×128?

No. 20.5 dB + 20.5 dB = 41.0 dB — no ITU-T PON standard supports this. The correct alternative: a 1×2 primary splitter at the CO feeding two independent 1x64s in parallel (~24 dB per path). This is fundamentally different from a series cascade — the 1×2 splits the OLT signal first, then each path goes through only one 1×64.

What’s the insertion loss of a 1×128 fiber optic splitter?

Standard single-chip: 23.5 dB max, 22.5 dB typical per ITU-T G.671. Premium: 23.0 dB max. Theoretical minimum: 10 × log₁₀(128) = 21.07 dB. The extra 1.5–2.4 dB is cumulative excess loss from seven Y-branch stages.

What’s the uniformity — and why does it matter more than insertion loss?

Standard single-chip: ≤2.0 dB. Premium: ≤1.5 dB. Spliced: ≤2.8 dB. Design for the weakest port, not the average. Every 0.5 dB of uniformity improvement buys ~1.4 km of additional fiber, or one more connector pair. When your link budget is tight, pay for better uniformity — not better typical IL.

How many subscribers can one OLT port serve with a 1×128?

128 physical connections. XGS-PON 10 Gbps ÷ 128 = 78 Mbps average. NG-PON2 TWDM (4-λ configuration) 40 Gbps ÷ 128 = 312 Mbps average. At 40% peak concurrency: ~195 Mbps per active sub on XGS-PON, fine for 4K streaming. At 80% concurrency: ~98 Mbps, still adequate for most households. For business SLAs, do the bandwidth math before the link budget math.

What connector type for 1×128?

SC/APC (green, 8° angle). At 23.5 dB splitter loss, return loss matters — APC’s ≥50 dB prevents back-reflections from degrading signal-to-noise ratio. Use LC/APC only if you need higher front-panel density (64 duplex vs 128 simplex). Never use UPC unless you have >5 dB of excess margin to absorb the ~10 dB lower return loss.

What fiber type should be used?

G.657.A2 bend-insensitive single-mode (≤7.5 mm min bend radius per ITU-T G.657). With 128 fibers in a rack, tight bends are unavoidable; G.652.D’s 30 mm minimum is routinely violated. G.657.A2 prevents macro-bend loss that could push a marginal port (-25.5 dB) into failure.

Can a 1×128 PLC splitter be used for 5G fronthaul?

Technically yes, but rarely optimal. In eCPRI/O-RAN fronthaul, each radio unit needs 25 Gbps dedicated — far above residential PON. A 1×128 could distribute WDM-PON to 128 small cells, but most 5G fronthaul uses 1×8 or 1×16 passive WDM today due to tighter power budgets and higher per-RU bandwidth. This may shift with 50G-PON-based fronthaul in 3GPP Release 19+.

What ODN Class do I need for 1×128?

ODN Classes per ITU-T G.989.2 Amendment 1:

When writing an RFP, reference the ODN Class — it’s more precise than naming a PON standard.

Does a 1×128 work with 50G-PON?

Marginally. 50G-PON (ITU-T G.9804.x, per IEEE 802.3ca) specifies 29–31 dB — leaving 2.5–4.5 dB for fiber plant, barely 10 km. Current 50G-PON field trials use 1×32 or lower. As higher-power 50G optics mature, the budget may increase. For now, keep 50G-PON at 1×64 max.

What certifications should a 1×128 PLC splitter have?

Minimum: Telcordia GR-1209-CORE (qualification: IL, uniformity, WDL, PDL at all wavelengths), GR-1221-CORE (reliability: temperature cycling, damp heat, thermal shock, aging), RoHS 2.0. For China: YD/T 2000.1-2014. Recommended: IEC 61753-1 (performance standard). Always request the per-port test report — for 128 ports, sample testing is statistically insufficient.

What’s the difference between 1×128 and 1×64 splitter?

A 1×128 splits one input into 128 outputs (23.5 dB max loss). A 1×64 splits into 64 outputs (20.5 dB max loss). The 3 dB difference gives 1×64 about 8.5 km of additional fiber reach, more margin for repairs, and lower per-port cost. Use 1×128 only when you must serve 128 subscribers from a single OLT port and your PON budget is ≥32 dB.

How do you test a 1×128 splitter?

Production testing uses an automated optical switch with a stabilized light source and calibrated power meter, cycling through all 128 ports at 5 wavelengths. This takes ~15 minutes with near-zero error rate. Field acceptance should compare per-port loss to the factory report using a light source and power meter. OTDR testing through a 1×128 is impractical due to the ~23 dB splitter step.

What package types are available for 1×128 splitters?

Common packages: 2U 19-inch rack-mount tray (best for CO/headend), ABS box (~140 × 115 × 18 mm, for FDH cabinets), and LGX cassette (for modular frames). Pole-mounted street closures are generally unsuitable for 128-port devices due to fiber count and cable management.

12. Specifying & Ordering 1×128 PLC Splitters from BWNFiber

How to Evaluate a 1×128 Supplier (Beyond the Datasheet)

When you’ve received quotes from 3–4 suppliers, don’t sort by unit price. Sort by these five things — in this order:

1. Per-port test report availability. If a supplier can’t or won’t provide one, eliminate them. At 128 ports, sample testing is statistically useless. A single missing test report means you’re gambling on 128 ports you can’t verify.

2. Single-chip vs spliced — and can they prove it? Ask directly. Then check the per-port test report for the cluster pattern (the single-chip vs spliced section explains how). A supplier who says “single-chip” but can’t show a smooth IL distribution is lying or doesn’t know — either way, eliminate.

3. Wavelength coverage of QC testing. “GR-1209 compliant” with 2-wavelength testing (1310 + 1550 nm) is the minimum — sufficient for single-PON deployments. If you’re planning multi-PON coexistence (GPON + XGS-PON + NG-PON2), demand 5-wavelength data (1310, 1490, 1550, 1577, 1625 nm). The WDL spec only matters if it’s verified across all wavelengths you’ll actually use.

4. Reliability testing beyond the certificate. A GR-1221 certificate says the design was qualified once. What you need is batch-level reliability data — especially damp heat test results if you’re deploying in a humid climate. Ask: “Can you provide damp heat test data (85°C/85% RH) for the production batch my units will come from?” Most suppliers can’t. The ones who can are the ones who invest in reliability.

5. Engineering support before the order. Send your link budget to each shortlisted supplier. If they reply with a quote without checking whether your budget works, they’re selling boxes, not engineering. If they reply with questions about your feeder length, connector count, and coexistence plan, they’re worth talking to.

Before You Place the Order: A Pre-PO Checklist

• [ ] Per-port test report reviewed for all 128 ports at all relevant wavelengths — not accepted “representative sample”
• [ ] Supplier confirmed single-chip or spliced — and the test data matches the claim
• [ ] WDL verified across your specific PON wavelengths (not just 1310+1550 if you’re running multi-λ)
• [ ] Damp heat / thermal cycling data confirmed for your deployment climate
• [ ] Packaging type specified (rack-mount / ABS / LGX) with connector type and pigtail length
• [ ] Ribbon or single-fiber output specified — and whether spare ports are pre-terminated
• [ ] Shipping packaging confirmed (foam-lined flight case for rack-mount; individual port protectors)
• [ ] Warranty terms and exclusions reviewed — especially climate-related exclusions if deploying outdoors
• [ ] Lead time confirmed against your project schedule (15–20 days standard; 20–30 days premium)
• [ ] Link budget re-calculated with actual fiber route lengths, not map distances (+20% to +30% margin sensitivity checked)
• [ ] Coexistence element loss included if multi-PON (+2 dB)
• [ ] Field acceptance test procedure defined — which ports get verified, with what equipment, to what pass/fail criteria

Three Procurement Mistakes I’ve Seen Repeatedly

Mistake 1: Buying by unit price and skipping per-port testing. One operator saved $2,800 on 20 units by choosing the lower-cost spliced supplier who didn’t include test reports. Six months later, 11 subscribers on three different splitters had intermittent BER spikes at 1577 nm (XGS-PON downstream). The root cause: WDL on those specific ports exceeded 2.5 dB — invisible at 1310/1550 nm but fatal at 1577 nm. They spent $4,200 on troubleshooting before replacing the splitters.

Mistake 2: Assuming all “GR-1221 qualified” splitters are equal. An operator in coastal Brazil bought standard-grade splitters from a supplier whose GR-1221 certificate tested only thermal cycling and dry heat — no damp heat. Two monsoon seasons later, 30% of ports showed 1+ dB IL drift from moisture ingress. The warranty claim was denied because the splitters met their published spec (which didn’t include damp heat). The GR-1221 qualification scope matters more than the certificate itself.

Mistake 3: Ordering without confirming lead time against the project schedule. A Dubai tower project ordered premium-grade 1x128s three weeks before planned installation. Lead time was 25 days. The splitters arrived five days late, the installation crew was already on-site, and the delay cost two days of idle crew time — $8,000 in labor for a $1,200 splitter order. Confirm lead time before confirming the installation date, not after.

Why Consider BWNFiber for 1×128

Four things you should verify with any 1×128 supplier — and how we handle each:

Standard Configuration & Options

• Single-chip PLC (not spliced — see single-chip vs spliced section)
• ≤23.5 dB max / ≤22.5 dB typical IL (≤23.0 dB premium)
• Uniformity: ≤2.0 dB standard / ≤1.5 dB premium / ≤1.4 dB ultra-premium
• PDL: ≤0.5 dB standard / ≤0.3 dB premium
• WDL: ≤1.5 dB; TDL: ≤0.5 dB
• SC/APC standard; SC/UPC, LC/APC, LC/UPC, FC/APC optional
• 2U rack-mount tray, ABS box, or LGX cassette packaging
• Single-fiber or ribbon outputs (10×12 + 8, or 11×12 with spares)
• G.657.A2 bend-insensitive fiber per ITU-T G.657
• GR-1209-CORE, GR-1221-CORE, IEC 61753-1, YD/T 2000.1-2014, RoHS 2.0 compliant
• Per-port test report at 5 wavelengths, every unit. No exceptions on 1×128

MOQ, Lead Time & Warranty

Samples, Shipping & Warranty Claims

• Samples: Single evaluation unit available at list price (credited against first production order above 50 units). Sample ships with the same per-port test report as production units. International sample shipments go via DHL/FedEx air courier (3–7 days worldwide) with full customs documentation.
• Shipping: 128-fiber assemblies ship in foam-lined flight cases with individual port protectors. We’ve shipped 1×128 splitters to 14 countries without a single transit-damage claim — the packaging is over-engineered for a reason. See the Export Packaging section above for documentation and lead times.
• Warranty process: If any port drifts beyond spec during the warranty period, we ship a replacement unit with matching test data within 72 hours of claim verification. You don’t ship the failed unit back first — we cross-ship. For international buyers: replacement units ship via air courier. Customs duties on warranty replacements are the buyer’s responsibility unless your country has a duty-drawback provision for warranty goods (check with your customs broker).

Before You Order: Send Us Your Link Budget

Send your PON standard, feeder length, drop length, connector count, and coexistence requirements. We’ll calculate whether 1×128 fits with enough margin — or tell you honestly if 1×64 handles it better, with numbers to back it up.

A word on link budget accuracy: If you give us straight-line map distances instead of actual fiber route lengths, expect the real-world loss to be 15–30% higher. Buried fiber follows roads, goes around obstacles, and has service loops at splice points. A “5 km” straight-line feeder is usually 6–7 km of actual cable. We budget for actual route length, not straight-line distance. If you haven’t done the detailed route survey yet, give us your best estimate and we’ll show you the margin sensitivity at +20% and +30% route length.

What happens if it doesn’t work: If a port drifts beyond spec during the warranty period, we cross-ship a replacement with matching test data within 72 hours of claim verification. Warranty covers IL drift, uniformity degradation, and PDL increase beyond datasheet maximums under normal operating conditions per GR-1221-CORE. Warranty excludes: connector end-face contamination (clean before claiming), physical damage to the fiber array or adapter panel (crushed fibers, bent adapters), and deployment in unmanaged outdoor closures without climate control if using ABS box packaging. If you’re deploying ABS box 1×128 outdoors, specify IP68-rated splice closures with desiccant — moisture trapped inside a non-vented ABS enclosure accelerates epoxy degradation.

Start Here: How to Reach BWNFiber

Not sure if 1×128 fits your project? Send your PON standard, feeder length, drop length, connector count, and coexistence requirements. We’ll run the link budget calculation and tell you what works — even if the answer is “stay with 1×64.” This is the same math that’s led me to say no more often than yes over ten years.

Comparing suppliers? Request a de-identified sample test report showing all 128 ports at 5 wavelengths per IEC 61300-3-7. It shows you exactly what we ship with every unit.

Ready to quote? Include your target quantity, connector type, package preference, and deployment climate. We’ll return a quote with per-port test report commitment and confirmed lead time within 24 hours.

[email protected]
+86-13615744790 (WhatsApp)
www.bwnfiber.com

Related Reading:

• 1×64 Fiber Optic Splitter Guide — The practical ceiling for GPON, and the right choice for most deployments where 1×128 gets proposed
• the complete 1×2 through 1×128 loss chart with formulas and ODN class reference — Every split ratio in one table
• GPON network design primer — how split ratio cascades through the ODN — Centralized vs distributed splitting trade-offs
• PLC vs FBT Splitter: why 1×8 is the FBT ceiling — The manufacturing limits behind the technology choice
• Rack-mount splitter deployment and cable management — CO and headend packaging beyond the spec sheet

References and Standards

• ITU-T G.984.2 — Gigabit-capable passive optical networks: Physical media dependent layer specification
• ITU-T G.9807.1 — 10-Gigabit-capable symmetric passive optical network
• ITU-T G.989.2 — 40-Gigabit-capable passive optical networks
• Telcordia GR-1209-CORE — Generic requirements for passive optical components
• Telcordia GR-1221-CORE — Generic reliability assurance requirements for passive optical components
• Fiber Broadband Association — FTTH deployment best practices and workforce development