3. Function analysis of the retractor pretensioner

Function analysis is the bridge between the system description in Chapter 2 and the failure analysis that follows. For the retractor pretensioner, the function is deceptively simple to state — "remove webbing slack from the occupant restraint system at the onset of a crash" — but the chain that delivers that function contains a pyrotechnic energy source, a gas-driven mechanical actuator, a rotating spool, and a webbing path whose geometry changes as the occupant loads the belt. DFMEA requires that each link in this chain be expressed as a verifiable function with a measurable performance requirement, because a failure mode can only be defined relative to a function that has a defined success criterion.

Decomposing the Function into a Signal-and-Energy Chain

The pretensioner does not act on a sensor signal in the way an airbag does; it acts on an energy release command. The correct functional decomposition therefore follows the energy and motion path rather than the electrical path. Four sub-functions are typically distinguished:

• F1 — Arm and trigger: receive the crash signal from the airbag ECU (typically a current pulse on the squib circuit) and initiate the pyrotechnic charge within the specified firing window.
• F2 — Generate gas and convert to mechanical work: burn the propellant, pressurize the pressure chamber, and drive the piston or turbine that converts gas enthalpy into linear or rotary motion.
• F3 — Rotate the spool: transmit the actuator motion to the spool through the drive interface (rack-and-pinion, ball-and-tube, or rotary turbine), producing a defined spool rotation angle within a defined time.
• F4 — Tension the webbing: wind webbing onto the spool so that slack is removed and belt force at the occupant rises to the target pretensioning level, without exceeding the webbing or occupant load limits.

Each sub-function must carry a requirement expressed in engineering units. A representative (illustrative, not measured) requirement set for a front-row retractor might read: F1 fires within 2 ms of command at −35 °C to +85 °C; F2 generates chamber pressure sufficient to complete the stroke in under 12 ms; F3 produces 120°–160° of spool rotation; F4 removes 60–120 mm of webbing and develops 2.0–3.5 kN of belt force at the shoulder, with webbing payout after pretensioning governed by the separate load limiter function.

Pyrotechnic Actuation: What the Function Actually Requires

The actuation sub-function has three separable requirements that are often collapsed into one and should not be. The first is response: the squib must fire when commanded, across the full temperature range and after the environmental aging defined in the design specification. The second is timing: the interval between command and the start of spool rotation must be short and repeatable, because pretensioning competes with occupant forward excursion. The third is containment: the pressure vessel and gas path must retain the combustion products and direct them to the actuator, with no leakage path that would reduce available work.

In practice the timing requirement is the one that most often drives design change. The pyrotechnic charge, the initiator, and the gas passage geometry together determine the pressure rise rate. A charge that is adequate at 23 °C may be marginal at −35 °C, where propellant burn rate drops and the gas is denser. Function analysis should therefore state the timing requirement as a function of temperature, not as a single nominal value, so that the subsequent failure analysis can ask the right question: at which corner of the operating envelope does the function degrade first?

Spool Rotation: Kinematics and the Drive Interface

The spool rotation sub-function converts actuator output into angular displacement. Two parameters dominate: the total rotation angle and the rotation rate profile. The angle determines how much webbing is taken up, through the simple kinematic relation

webbing taken up  =  θ × r_effective

where θ            = spool rotation angle (rad)
      r_effective  = effective winding radius (mm), which grows as
                     webbing layers accumulate on the spool

Because r_effective increases with each wrap, the same rotation angle removes progressively more webbing per radian. This is why the requirement is usually written as a webbing take-up target rather than a pure angle target, and why the spool's initial wrap count is a legitimate design parameter in the function analysis. A spool that starts with more pre-wrapped webbing reaches the take-up target with less rotation but develops belt force earlier in the stroke.

The drive interface is where the function is most vulnerable to loss. In a rack-and-pinion pretensioner, the piston drives a toothed rack that meshes with a pinion coupled to the spool through a one-way clutch. The clutch must engage in the tensioning direction and freewheel in the payout direction so that the load limiter can function afterward. In a ball-and-tube design, the piston pushes a column of steel balls through a curved tube, and the balls engage the spool hub directly. Each architecture has a distinct set of interface parameters — tooth engagement depth, clutch preload, ball column integrity — that belong in the function analysis because each is a place where the function can be lost.

Webbing Tensioning: The Occupant-Facing Output

The final sub-function is the one the occupant actually experiences. Tensioning has to accomplish two things at once: remove slack so that the occupant couples to the restraint early, and stop at a force level that does not itself injure the occupant. The second constraint is what separates pretensioning from a simple winch. Belt force during pretensioning is a function of the torque delivered to the spool, the effective winding radius, and the compliance of the webbing and the occupant's body:

F_belt  ≈  T_spool / r_effective  −  k_webbing × ΔL

where T_spool     = torque delivered through the drive interface (N·m)
      r_effective = effective winding radius (m)
      k_webbing   = webbing axial stiffness (N/mm)
      ΔL          = elastic stretch accumulated in the belt path (mm)

The subtraction term is not negligible. As the belt tightens against the occupant, the webbing stretches and the occupant's torso compresses, so the force at the shoulder is lower than the force implied by spool torque alone. Function analysis should state the tensioning requirement at the occupant interface — shoulder belt force and take-up distance — and treat spool torque as an intermediate parameter, not as the requirement itself.

Turning the Function Analysis into DFMEA-Ready Statements

The output of this chapter is not a narrative but a set of function statements that the failure analysis can operate on. A well-formed statement names the sub-function, the performance parameter, the operating condition, and the acceptance criterion. Compare a weak statement — "pretensioner shall tension the seatbelt in a crash" — with a usable one: "F4: with the retractor installed at the design belt geometry and the occupant represented by a 50th-percentile ATD, the pretensioner shall remove 60–120 mm of webbing and develop 2.0–3.5 kN shoulder belt force within 30 ms of command, at all temperatures from −35 °C to +85 °C."

The second statement can be failed in specific, testable ways. It can fail on take-up distance, on force level, on timing, or on temperature corner. Each of those becomes a failure mode with its own cause chain and its own detection method. The first statement can only be failed generically, which produces generic DFMEA rows that do not drive design action.

Where the Function Analysis Typically Goes Wrong

Three errors recur in pretensioner function analysis. The first is treating the pyrotechnic actuator as a black box and writing a single "fires / does not fire" function. This collapses timing, containment, and gas-path integrity into one row and loses the cold-temperature degradation mode that is often the dominant field concern. The second is specifying spool rotation in degrees without tying it to webbing take-up and effective radius, which makes the requirement insensitive to the initial wrap count and to webbing build-up during the stroke. The third is specifying belt force at the spool rather than at the occupant, which ignores webbing compliance and occupant compression and produces a requirement that can be met on the bench while the occupant-facing output is out of target.

With the four sub-functions defined, parameterized, and bounded as above, the next analytical step is to ask, for each one, how it can fail to deliver its stated performance — which is the subject of the failure mode and effect analysis that follows.