— Test lab

The MIPS helmet test laboratory, located 15km north of Stockholm in Sweden, is a product of twenty years of continual development. This hyper-modern, 300 sqm complex includes four experimental test machines designed to mimic real-world impact scenarios and state of the art tools to ensure we have full control over what we are measuring — tools like advanced, high-speed cameras that automatically save movies from each helmet test.

What do we measure?

MIPS measures 6DOF (six degrees of freedom) accelerations over time. To ensure that the MIPS BPS adds protection, we test the same model of helmet, with and without MIPS BPS, in all available sizes.


For each helmet, we focus our testing on three different areas: front, lateral and pitched. And because statistics tell us that the vast majority of accidents have angled impacts, we perform all helmet tests at an angle of 45 degrees.


But above all, the most important thing is to have control over the initial position of the head and the helmet during the vertical drop to the impact anvil.


It took 15 years to develop the complete MIPS test environment. Our first test machine came together back in 1997, and since then, it has evolved into the controlled test machine that we use today.

How do we test?

We developed our test methods to replicate real-life accident situations. For example, analysis of bike crash statistics shows that impacts most commonly occur at 6.0-6.5m/s at an angle of 45 degrees (Richter et al. 2007, Verschueren 2009, Bourdet et al. 2013). Therefore, we test the helmets at a speed of 6.2m/s with a similar impact angle.


As a substitute for the human head, we use an experimental system called the Hybrid III. Inside the head form, a set of accelerometers measures 6DoF accelerations over time. All helmets are impacted at three different points in order to control the MIPS BPS function around all three anatomical axis (X, Y and Z).


After the test, data from the impact is gathered and analysed. These thorough tests are performed to ensure that the MIPS BPS version of a helmet adds protection and that it can be approved according to the MIPS standard.

Test rig

We conducted our first test in 1996, using a pendulum to impact a spherical representation of a helmet, Figure A. Naturally, this was a very premature design, but it did give us the information to show that the MIPS BPS could work to reduce the forces, or torque to the head.

Figure A: Pendulum test rig measuring the torque. B and C: Movable plate, where the plate is accelerated by a pneumatic cylinder. D: Helmeted head form dropped to a 45 degree impact surface. In test B-D, the head form is equipped with a nine-accelerometer system [1, 2].


The design shown in Figure B and C is used in Halldin et al 2001 and Aare et al 2003. The test method shown in Figure D was used in Halldin et al 2015, and also the one proposed in CEN TC158 as a future test method.

Moving surface

Several publications detail how to design a method to measure the energy absorption in an oblique impact with a significant tangential force acting on the helmet. Aldman et al. 1968 presented a method that uses a spinning concrete wheel onto which helmets are dropped. Halldin et al. (2001) and Mills and Gilchrist (2008) presented methods where the head was dropped onto a sliding steel plate to create an oblique impact. Pang et al. (2008) presented a method similar to that employed by Halldin, but with the addition of an HIII neck and the option of measuring the force on the plate.

Drop to angled surface

Another way of testing helmets for oblique impacts is to drop them onto an angled surface. Deck et al. proposed to impact the helmet at 6.5 m/s at an angle of 60 degrees (90 degrees defines a pure vertical drop as done in current helmet test methods. It was also proposed to use the HIII head as it has more human-like inertial properties than the head form used today (EN960). Further, Willinger et al. (2014) suggested progress in helmet testing method by introducing the Hybrid III head form, tangential tests and model-based brain injury criteria.


The main difference between impacting the movable plate and dropping the helmet onto an angled surface is the difference in the gravitation vector in relation to the normal force vector against the helmet, which could result in different outcomes for the two methods, even when testing identical helmets with the same impact speed and impact angle. The movable plate could, therefore, be more realistic in simulating a fall from a bike or a horse to the ground. However, the movable plate method has drawbacks compared to the angled surface technique since it is more complex and because it can be challenging to maintain a constant speed of the plate during the impact.

Linear impact

A third potential test method is the NOCSAE (2006) pneumatic linear impactor, which was first developed by Biokinetics in Canada. The linear impactor is equipped with a curved plastic surface attached to a disc made of vinyl nitrile foam, to mimic a helmet-to-helmet hit (designed for American football or ice hockey helmets).


In this test, the head form is attached to an HIII neck and a sled moving horizontally. The test method specifies different impact locations on the helmet, all of which result in impacts to the centre of gravity in the dummy head (NOCSAE, 2006). Rousseau et al. (2011) have proposed a modification of the test method by hitting the helmet at directions that are offset from the centre of gravity of the head to simulate real impacts as seen in football and ice hockey games. However, this test is not taken into account as the impacts result in less tangential force and are not realistic for bicycle accidents.

Which testing method to use?

Due to its simplicity and robustness, MIPS uses the vertical drop to an angled impact surface method.

Why don’t you include a neck in your tests?

In current test methods, the head either falls unrestrained onto the impact surface (European test standards) or is constrained to a monorail via a rigid arm attached to the head (US test standards). These two methods can be said to represent the extremes of the scale. Sitting somewhere in the middle is the normal situation, in which the neck constrains the head.


One option instead of designing a test without a neck could be to attach an experimental neck (like the HIII neck) to the vertical sled in a monorail helmet test rig. But to design an oblique test method, questions remain as to whether the neck will affect the measured translational and angular accelerations in the dummy head. It is clear that the neck restrains the head and that it will, at some time, rotate around a point in the neck, or even lower down in the thoracic region.


Earlier studies like the COST 327 study, have shown that the amplitude of the angular acceleration is affected by the neck. Helmeted full body Hybrid III dummies were dropped onto an angled surface and compared to free-falling helmeted head forms. The results showed that the angular acceleration differed in amplitude by about 20%.



Studies with and
without neck

Beusenberg et al. (2001) presented a numerical study on helmet-to-helmet impacts simulating American football accidents. It concluded that the neck did indeed change the characteristics of the angular acceleration comparing impacts with and without a neck. In the study by Beusenberg, however, the impacts were close to a radial impact to the helmet, where the neck is the only cause for the rotation of the head, i.e. there was no or little tangential component in the impact.


Ghajari et al. (2012) showed that the angular acceleration components could differ by as much as 40% when comparing a helmet impact with the full body and the head only. In this study, Ghajari used the THUMS finite element model and simulated an oblique impact on the lateral (temporal) portion of the helmet. Ghajari proposed changing the inertial properties of the head to compensate for the neck and the body if using only the head in an oblique impact test.


Forero (2009) reconstructed 12 jockey accidents using MADYMO. Two of these were studied in detail in simulations with and without the body in a helmet-to-racetrack turf impact. The angular acceleration was increased from 6462rad/s2 to 10104rad/s2 in one case and from 5141rad/s2 to 6444rad/s2 in the second case, comparing the simulation with a complete body and a simulation with the head only. Forero also mentioned that the absence of the neck and the body might cause the direction of the acceleration to alter. This study stated that the MADYMO human body model provides an unrealistic representation of the flexibility in the vertebral joint that could have resulted in this large discrepancy.


Verschueren et al. 2009 performed reconstructions of 22 bike accidents using MADYMO. Nine of the accidents were simulated both with the head only and with the entire body. The results of this study showed that the correlation of the angular acceleration between the head-only simulation and the simulation with the complete body was good for four out of nine reconstructions. The correlation was defined as medium for three, while two out of nine were defined as poor, with a difference of about 30% for one of those examples defined as poor.


Forero 2009 discussed the duration of the impact pulse, pointing out that it is different in a jockey accident against racetrack turf (8-20ms) compared to bike accidents against a hard road (5-10ms). Therefore, when designing a test with a surface mimicking racetrack turf for jockey helmets, a neck might prove necessary.


Also, Fahlstedt et al. (2015) and Klug et al. (2015) have investigated the effect of the neck in helmeted impacts. The conclusion from both studies is that the neck is not needed, especially when comparing the change of the angular velocity.


The conclusion that can be drawn here is that, in general, the neck affects the motion of the head. It can also be argued that a test method could be defined with impact angles where the effect of the neck is small during the short time (5-10ms) during which the helmet comes into contact with the impacting surface.


Experimental tests on human cadavers show that the upper part of the human neck is flexible and could be seen as decoupled from the head for a certain amount of displacement or rotation. Motion in a human joint that does not result in a force or moment is defined as the neutral zone. The neutral zone in which the upper part of the neck allows the head to rotate without extensive load is in the range of 10 degrees, depending on the axis of rotation.


Thus, when we do not take muscle activity into account, the head can rotate around 10 degrees without having any effect on the kinematics in horizontal loading of the head. In a typical helmet to asphalt impact, the free-falling head rotates about 10 degrees during the first 10 ms of impact. Based on this, one could argue that there would not be sufficient time for the neck to affect the head in this specific impact direction significantly. However, neither Ivancic nor Camacho et al. analysed a helmet impact situation with a vertical compression force to the neck.


In order to define the importance of the neck in a typical helmet to ground impact situation, the partners in this COST action performed FE studies on how the neck could affect the head in a helmeted impact situation. The conclusion was that the neck affects the kinematics of the head, but that it is dependent on the impact point and direction.


Questions could also be raised on how human like the HIII neck is in helmet impact situations. The HIII neck is designed and validated only for frontal car collisions at speeds of around 11 m/s resulting in a flexion motion of the neck. Thus, the HIII dummy neck is not validated for compression loading, lateral bending or rotation around the vertical axis, also shown by Myers et al. (1989) and DiSantis (1991).


Other aspects of the neck/no neck questions that would need to be taken into consideration when designing a new test method are:

  • Assuming that the human neck does not affect the head during the first 10ms in most impact situations, is the result the same when the musculature is tensed to a theoretical maximum contraction?
  • There are disadvantages to the use of a neck, like the cost involved and the need for calibration.
  • There are advantages to use an experimental neck in a helmet test as it makes the positioning of the helmet easier, as the neck keeps the head in position.

The conclusion is that the human neck as a boundary condition to the human head will affect the head kinematics. However, for bicycle accidents most frequently impacting the ground resulting in short (5-10ms) impacts, the neck will not affect the head kinematics so much that the neck is essential. Taking all known aspects into account, it is proposed to design the new test method without the neck.