Abstract from our ATZ professional article:

Noise of seatbelt retractors

Dr. Josef Girstmair
Group Leader Powertrain Dynamics & Acoustic
In a joint project, the VIRTUAL VEHICLE Research Center, the BMW Group and the two system suppliers Autoliv and ZF TRW are developing methods to prevent annoying noise from seatbelt retractors during early vehicle development.  The developed simulation methodology and the extended acoustic test procedure represent a promising solution in vehicle acoustics.

Belt retractor noise, which can be described as a rattle, chatter, buzz or knocking sounds, is a disturbing noise that occurs in the vehicle interior. It is mainly caused by road excitation, especially on rough roads made of paving stone. Passengers can perceive such disturbing noises from retractor components as annoying or distracting, especially in quiet vehicles with little acoustic masking from driving noise. The retractor noise is excited by impacts between internal parts of the retractor, which are connected via clearances that are necessary for operation.

Figure 1: Retrector rolling mode with spool assembly in phase


Supporting an NVH-optimised component integration

Based on the analysis of the contact between vehicle structure and belt retractor, as well as the experimentally evaluated eigenmodes, modelling guidelines are derived for the calculation of the eigenmodes of a reduced Finite Element submodel, which consists of the retractor, the mounting bracket and the local vehicle periphery. Comparisons between the full-vehicle and sub-model levels confirm equal eigenfrequencies (figure 1) and mode shapes.

For a reliable prediction of the retractor vibration behaviour, it is necessary to calculate the nonlinear contact, including the elastic behaviour and the pre-tensioning of the bolting (Figure 2). The calculation of the contact makes it possible to identify an insufficient contact situation, which would reduce the  local eigenfrequencies and thereby negatively affect the noise emission of the retractor.

Figure 2: Comparison of simulated (left) and experimentally evaluated (right) contact pattern


Improved acoustic testing

Based on measurements of the operational vibration and acoustic behaviour in test vehicles under characteristic road excitation, the requirements for the component test rig and the test procedure were derived. Figure 3 shows the vibration levels for a retractor mounted in the back of the car . Each vehicle has its own vibration spectrum. Besides the low-frequency road excitation (<15 Hz), the highest amplitudes can be found at the retractor’s eigenfrequencies (45 to 90 Hz). The excitation of the ball sensor (ball rattling), a very significant noise issue in the car, is mainly excited in the frequency range up to about 50 Hz. Comprehensive investigations for the development of the test rig and the procedure yielded the conclusions described below.

So far it has not been possible to reproduce the 3D vibration excitation of the retractor mounted in the car on a component test rig due to the combination of the complex motion in and extremely low noise levels. In addition, the acoustic boundary conditions are different in the car and in the laboratory. Therefore, a uniaxial excitation of the retractor under well-defined, reproducible conditions in an acoustic room is proposed, in order to analyse the acoustic behaviour.

For the component test rig, a stiff connection between the mounting plate and the retractor is needed to shift the eigenfrequencies of the retractor up and to obtain a well-defined vibration excitation without any disturbance caused by the dynamic behaviour. Furthermore, a low disturbance for the non-excited directions is required in order to keep all disturbances low and decrease the variability of results.

Figure 3: Retractor vibrations measured in a car during paving stone excitation (SB: rear suspension body side, MB: mounting bracket, SP: screw point retractor, RE1: measurement point 1 on retractor)


Round Robin

To analyse the reproducibility of the test procedure, acoustic testing was performed on different retractors in three different laboratories. In general, the results show a good agreement between the labs and low variance at each lab. The dynamic equivalent continuous sound power level can characterise the differences between the retractors considering the noise emission very well. It includes both the sound intensity and the duration and occurrence of the retractor noise.


Disturbance noises from belt retractors are a real challenge for vehicle manufacturers due to the high complexity of the problem, strict cost and weight requirements and, in particular, due to the need for decisionmaking at early development stages. The partners Autoliv, BMW Group, VIRTUAL VEHICLE Research Center and ZF TRW have developed specific, demand-oriented virtual and experimental methods to improve the vehicle body design in the regions of interest in order to support the target setting in the specification process and in monitoring the product quality.

The complete article can be read in the ATZ edition 7-8 / 2017.