8.2 Mandrel Wrapping and Encircled Flux (EF) Compliance

Mandrel Wrapping and Encircled Flux (EF) Compliance

Optical transmission in multimode fibers is inherently complex because optical energy propagates across hundreds of spatial paths simultaneously. The distribution of optical power among these guided paths—termed the modal distribution—fundamentally dictates how much optical attenuation a test set measures across a link.

Historically, variations in launch conditions between different test sets produced dramatic measurement discrepancies on the exact same physical fiber link. For technicians preparing for the BICSI Installer 2 Optical Fiber (INSTF) certification, understanding the physics of modal power distribution, the mechanics and standard dimensions of mandrel wrapping, the operational limitations of mandrels when paired with laser sources, and the industry mandate for Encircled Flux (EF) compliance is vital for reliable Tier 1 field certification.

---

The Physics of Multimode Waveguides and Modal Power Distribution

A multimode optical fiber (such as 50/125 $\mu\text{m}$ OM2–OM5 or legacy 62.5/125 $\mu\text{m}$ OM1) features a core diameter substantially larger than the wavelength of transmitted light. Light entering the core excites hundreds of discrete electromagnetic spatial distributions known as modes.

The approximate number of guided mode groups ($M$) supported in a graded-index parabolic core fiber is determined by the normalized frequency parameter ($V$-number):

$M \approx \frac{V^2}{4} = \frac{1}{4} \left( \frac{2\pi a}{\lambda} NA \right)^2$

Where:

• $a$ is the core radius ($25\,\mu\text{m}$ for 50 $\mu\text{m}$ fiber; $31.25\,\mu\text{m}$ for 62.5 $\mu\text{m}$ fiber).
• $\lambda$ is the operating wavelength (e.g., $850\text{ nm} = 0.85\,\mu\text{m}$).
• $NA$ is the numerical aperture of the fiber ($0.200$ for 50 $\mu\text{m}$; $0.275$ for 62.5 $\mu\text{m}$).

In a 50/125 $\mu\text{m}$ fiber at 850 nm (NA 0.200), $V \approx 37$, so roughly 340 guided modes propagate simultaneously. These modes are spatially stratified into two broad classifications:

1. Low-Order Modes: Propagate near the central longitudinal axis of the core. They travel along relatively direct, shallow angles, exhibit low attenuation, and rarely interact with the outer core-cladding boundary.
2. High-Order Modes: Propagate along steep, helical trajectories that bounce near the outer periphery of the core. Because their electromagnetic fields penetrate deeply into the core-cladding boundary, high-order modes are highly sensitive to microbends, geometric core variations, and absorption in the cladding, causing them to attenuate at a much higher rate.

+-------------------------------------------------------------------------+
| High-Order Mode (Steep Angle, Interacts with Cladding Boundary - High Loss)|
| ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ |
| Low-Order Mode (Axial, Concentrated at Core Center - Low Loss)          |
| ----------------------------------------------------------------------- |
| Low-Order Mode (Axial, Concentrated at Core Center - Low Loss)          |
| ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ |
| High-Order Mode (Steep Angle, Interacts with Cladding Boundary - High Loss)|
+-------------------------------------------------------------------------+

Equilibrium Mode Distribution (EMD)

As optical energy travels down a continuous multimode fiber over long distances, high-order modes gradually radiate away, and mode coupling transfers power among the remaining modes. Eventually, the optical power reaches a stable, reproducible steady state known as Equilibrium Mode Distribution (EMD).

In long-haul systems, EMD naturally establishes itself after approximately 1 to 2 kilometers of continuous glass. However, structured premises cabling links in commercial buildings and data centers typically measure only 30 to 300 meters. Over these short distances, light never reaches natural EMD. Consequently, the measured insertion loss of a short premises link is completely dominated by the initial launch condition of the test instrument's optical light source.

---

Overfilled Launch (OFL) vs. Restricted Mode Launch (RML) / Underfilled Launch

Historically, optical light sources exhibited vastly divergent emission patterns:

1. Overfilled Launch (OFL) — Legacy LEDs

Legacy surface-emitting Light-Emitting Diodes (LEDs) have a broad, divergent radiation pattern and a wide spot size that exceeds the fiber core diameter. An LED completely floods both the core and cladding with optical energy, exciting all possible low-order and high-order modes (and even temporary cladding radiation modes).

When testing a short premises link with an overfilled LED:

• The high-order modes attenuate rapidly within the first 50 to 100 meters.
• The power meter detects a large drop in optical power.
• Measurement Impact: The test instrument overstates link attenuation, leading to overly pessimistic readings and false failures on high-quality cabling installations.

2. Underfilled Launch / Restricted Mode Launch (RML) — VCSELs

Modern gigabit and multi-gigabit transceivers (such as 10GBASE-SR, 25GBASE-SR, 40GBASE-SR4, and 100GBASE-SR4) utilize Vertical-Cavity Surface-Emitting Lasers (VCSELs) operating at 850 nm. A VCSEL has a very narrow beam divergence (approximately $7^\circ$) and a tiny emission spot size ($10\text{ to }15\,\mu\text{m}$), which is much smaller than the 50 $\mu\text{m}$ core.

A raw VCSEL injects light almost exclusively into the low-order axial modes at the center of the core, leaving intermediate and high-order modes completely unexcited.

When testing a link with an underfilled VCSEL:

• The light travels straight down the center axis, completely bypassing imperfections, minor bends, or core boundary irregularities at connector interfaces.
• The power meter detects virtually no attenuation.
• Measurement Impact: The test instrument understates link attenuation, producing artificially optimistic readings and false passes on marginal or poorly installed links that will subsequently fail in service.

The Historical Measurement Crisis

Before standardized launch conditioning, one technician testing a link with an overfilled LED source would measure a loss of $2.20\text{ dB}$ (FAIL), while another technician testing the identical link with an underfilled VCSEL source would measure $1.20\text{ dB}$ (PASS). This $\pm 40\%$ measurement uncertainty ($1.0\text{ dB}$ or more of variance) rendered field certification unreliable and created massive disputes between installers, cabling manufacturers, and network owners.

---

High-Order Mode Stripping via Mandrel Wrapping

To establish repeatable measurements with overfilled LED sources, standards bodies developed the mandrel wrap. A mandrel wrap serves as a mechanical high-order mode stripper.

Physical Mechanism of Mode Stripping

When an optical fiber is wrapped tightly around a smooth cylinder (mandrel) of calibrated diameter, the physical curvature alters the internal geometry of the core-cladding interface:

• For low-order axial modes, the angle of incidence remains well above the critical angle, allowing them to traverse the bend with negligible power loss.
• For high-order modes traveling near the core boundary, the bend forces their angle of incidence below the critical angle ($\theta_1 < \theta_c$).
• As a result, the high-order modes refract across the core-cladding boundary into the cladding glass, where they are attenuated and dissipated by the protective polymer coating.

By stripping out the transient, highly attenuating high-order modes, the mandrel wrap forces an overfilled LED launch to approximate steady-state Equilibrium Mode Distribution (EMD) before light enters the link under test.

Mandrel Wrapping Protocol

To apply a standards-compliant mandrel wrap in the field, technicians must follow three strict rules:

1. Location: Apply the mandrel wrap only to the launch Test Reference Cord (TRC) exiting the optical light source. Never apply a mandrel to the receive cord or to the link under test.
2. Number of Turns: Wrap the reference jumper exactly five (5) complete turns around the cylinder.
3. Turn Spacing: Turns must be wrapped smoothly and closely side-by-side. Never overlap turns. Overlapping turns creates localized pinch points that crush the fiber, causing uncontrolled microbending loss and potential fiber breakage.

Mandrel Sizing Standards (ANSI/TIA-526-14 and ANSI/TIA-568)

The diameter of the mandrel is critical. If the mandrel is too large, high-order modes are not fully stripped. If the mandrel is too small, excessive bending strips low-order modes, introducing severe macrobend attenuation.

ANSI/TIA-526-14 and ANSI/TIA-568 specify exact mandrel outer diameters based on the fiber core diameter (50 $\mu\text{m}$ vs. 62.5 $\mu\text{m}$) and the outer jacket construction of the test cord:

The Mechanics of Cable Jacket Compression

Notice that for both fiber types, the mandrel diameter for a 3.0 mm jacketed cable is 3 mm smaller than the mandrel for an unjacketed 900 $\mu\text{m}$ buffered fiber (22 mm vs. 25 mm for 50 $\mu\text{m}$; 17 mm vs. 20 mm for 62.5 $\mu\text{m}$).

This size reduction is intentional and mathematically derived:

• A 3.0 mm jacketed cable features a thick outer PVC/plenum jacket and a layer of aramid yarn (Kevlar) surrounding the inner 900 $\mu\text{m}$ buffer tube.
• When wrapped around a mandrel, the outer jacket and yarn create a physical cushion that holds the central glass core away from the mandrel surface.
• If a 3.0 mm jacketed cord were wrapped around a 25 mm mandrel, the fiber at the center of the cord would follow a bend diameter of about $(25\text{ mm} + 3\text{ mm}) = 28\text{ mm}$, too loose to strip the intended high-order modes.
• By reducing the mandrel diameter to 22 mm, the thickness of the jacket brings the effective bend diameter of the internal glass core back to the required 25 mm.

---

The Failure of Mandrels with Modern VCSEL Sources

While mandrel wrapping was an effective solution for legacy overfilled LED sources, it is completely ineffective for controlling modern laser-based launch conditions.

A mandrel wrap functions strictly as a spatial low-pass filter: it removes excess energy from high-order modes at the outer periphery of the core. However, a mandrel cannot add optical power to unexcited modes, nor can it redistribute energy outward from the center.

When an underfilled 850 nm VCSEL injects light exclusively into the central $10\text{ to }15\,\mu\text{m}$ of a 50 $\mu\text{m}$ core:

1. The light travels entirely through low-order axial modes.
2. When the jumper is wrapped around a 25 mm mandrel, the curvature strips energy from the outer core boundary—but there is zero optical power at the outer boundary to strip.
3. The light exiting the mandrel remains strictly underfilled.

As multi-gigabit premises networks transitioned entirely from LEDs to 850 nm VCSEL transceivers, mandrel-conditioned testing continued to produce underfilled launch conditions. Installers were certifying links that exhibited artificially low insertion loss, only to find that active network transceivers suffered high bit error rates upon deployment.

---

Encircled Flux (EF) Compliance (ANSI/TIA-526-14-C and IEC 61280-4-1)

To eliminate the measurement discrepancies inherent in both overfilled LEDs and underfilled VCSELs, international standards committees established Encircled Flux (EF) compliance under IEC 61280-4-1 and ANSI/TIA-526-14-C.

ANSI/TIA-568.3-D and -E require multimode loss testing to use an EF-compliant launch (per ANSI/TIA-526-14-C, which adopts IEC 61280-4-1).

Mathematical Definition of Encircled Flux

Encircled Flux defines the radial distribution of optical launch power exiting the launch cord into the link under test. Mathematically, Encircled Flux ($EF(r)$) represents the fraction of cumulative optical power emitted within a given radial distance ($r$) from the optical core center, normalized to the total optical power emitted across the entire core radius ($R$):

$EF(r) = \frac{\int_0^r I(\rho) \rho \, d\rho}{\int_0^R I(\rho) \rho \, d\rho}$

Where:

• $I(\rho)$ is the near-field optical intensity profile measured at radial distance $\rho$ from the core center.
• $r$ is the radial evaluation boundary.
• $R$ is the full radius of the fiber core ($25\,\mu\text{m}$ for 50 $\mu\text{m}$ fiber).

Encircled Flux Cumulative Power Profile at 850 nm (50/125 um Fiber):

100% |                                              .- - - - Upper EF Limit
     |                                       . - '
     |                                . - '
 50% |                         . - '
     |                  . - '
     |           . - '  _ _ _ _ _ _ _ _ _ _ _ _ _ _ Lower EF Limit
     |    . - '
  0% +-------------------|-------------------|-------------------|
     0                  10 um               15 um               20 um
                                Radial Distance (r)

The Standard EF Template Boundaries

The standard establishes strict upper and lower boundary curves that the measured cumulative optical power profile must fall within across specific radial benchmarks (evaluated at $r = 10\,\mu\text{m}$, $r = 15\,\mu\text{m}$, and $r = 20\,\mu\text{m}$ for 50/125 $\mu\text{m}$ fiber at 850 nm):

• Upper Boundary Curve: Prevents the launch from becoming overfilled. If the power curve rises above the upper limit, too much energy resides in high-order modes, which would cause false failures.
• Lower Boundary Curve: Prevents the launch from becoming underfilled. If the power curve falls below the lower limit, light is overly concentrated in the center core axis, which would mask connector misalignments and cause false passes.

Field Implementation: EF Launch Cords and Modal Controllers

Technicians do not need to calculate integrals in the field. Test equipment manufacturers incorporate Encircled Flux compliance into field certification systems using two primary methods:

1. Integrated EF Light Sources: The optical light source within the OLTS contains internal precision optical conditioning that directly emits an EF-compliant launch beam.
2. EF-Compliant Test Reference Cords (EF-TRCs): A factory-calibrated launch cord that incorporates an inline modal controller. The modal controller utilizes precision micro-optics or engineered mode-scrambling elements that convert whatever arbitrary launch profile the source generates into a certified EF-compliant profile at the connector output ferrule.

---

Preventing Multimode Measurement Errors in Field Testing

To maintain Encircled Flux compliance and avoid corrupted test data, technicians must observe four critical operational practices:

1. Never Mix 50 $\mu\text{m}$ and 62.5 $\mu\text{m}$ Cores

Launching light from a 62.5 $\mu\text{m}$ test reference cord into an installed 50 $\mu\text{m}$ permanent link introduces a massive, artificial insertion loss. Because light from the larger 62.5 $\mu\text{m}$ core spills into the 50 $\mu\text{m}$ cladding, the mismatch creates an immediate geometrical coupling loss of approximately $4.7\text{ dB}$:

$\text{Area term} = 10 \log_{10}\left(\frac{50^2}{62.5^2}\right) \approx -1.94\text{ dB}, \qquad \text{NA term} = 20 \log_{10}\left(\frac{0.200}{0.275}\right) \approx -2.77\text{ dB}$

Together the two terms predict about 4.7 dB of loss for a fully filled launch.

Conversely, launching from a 50 $\mu\text{m}$ TRC into a 62.5 $\mu\text{m}$ link creates a severely underfilled launch condition, invalidating EF compliance and masking connection losses.

2. Avoid Cross-Mandrel Application

When legacy mandrel testing is contractually required on older LED installations, never swap mandrels between fiber types:

• Wrapping 50 $\mu\text{m}$ fiber around a 20 mm (62.5 $\mu\text{m}$) mandrel bends the glass too sharply, inducing macrobending loss in low-order modes.
• Wrapping 62.5 $\mu\text{m}$ fiber around a 25 mm (50 $\mu\text{m}$) mandrel fails to strip high-order modes, leaving the launch overfilled.

3. Protect and Inspect EF Launch Cords

EF-compliant launch cords are precision optical instruments. Technicians must never clamp, sharply bend, or zip-tie an EF launch cord. If the internal modal controller is stressed or damaged, the launch condition will fall out of the EF template, corrupting all test results.