Impact Mitigation in Modern Football Helmets: Advances and Limitations of Position-Specific Designs

Summary

Modern football helmets underperform at the posterior aspect of the helmet. The measurement process described herein elucidates the sports helmet's ability to mitigate translational and rotational accelerations caused by impacts of various magnitudes. It robustly characterizes energy dissipation mechanisms and can be used to refine the helmet design process.

Abstract

It has been established that repetitive head impacts lead to the accrual of changes in brain function, structure, and chemistry. The helmet is the primary piece of protective equipment for participants in American football. The first helmets were constructed from leather. Plastic helmets with chin straps were developed in the 1950s, but testing standards were not developed until 1973. While those standards resulted in improved padding materials, previous work demonstrated that modern helmets mitigate translational accelerations better than rotational accelerations, especially on a helmet's Rear Aspect. The goal of this study was to detail the methodology used to examine the impact mitigation of football helmets with differing energy dissipation mechanisms, including position-specific helmets. The experiments were integrated within a dimensional analysis framework to yield a technique that makes it possible to simultaneously evaluate translational and rotational accelerations due to quantified impact loads at multiple locations around the helmet. Despite improvements in the overall impact mitigation of one of the helmets tested here, in general, they did not achieve their design goals.

Introduction

It has been established that there are 1.6-3.8 million sports-related traumatic brain injuries (TBIs) each year in the United States^1,2,3,4, resulting primarily from contact sports such as football, ice hockey, soccer, lacrosse, rugby, and wrestling. While football consistently exhibits the greatest rate of diagnosed concussion, it should be noted that these rates are dramatically underestimated due to the underreporting of concussion symptoms^5,6,7,8,9. With or without a diagnosed concussion, it has been established that repeated head acceleration events (rHAEs) can lead to marked changes in brain structure¹⁰, function, and chemistry^{11,12,13,14,15,16,17,18,19,20,21}. These rHAEs may be the result of direct impacts between players, a player and the ground, or whiplash events resulting from impacts to the legs or torso. Subsequent research illustrated the effects of rHAEs on cognitive impairment, depression, and anxiety²². It has even been suggested that many years of exposure to rHAEs may increase the risk for long-term neurodegenerative disorders such as Alzheimer's Disease (AD) and chronic traumatic encephalopathy (CTE)²³ in both soccer and football players^24,25,26. In order to protect participants in contact sports such as American football from the deleterious effects of rHAEs, it is critical to understand the physics of head impacts, starting with the helmet, which serves as the principal piece of protective equipment.

The first football helmets were made of leather and became a required piece of equipment in college football in 1939²⁷. Chin straps and facemasks were in use by the 1950s, but it was not until 1973 that a testing standard was developed by the National Operating Committee on Standards for Athletic Equipment (NOCSAE). This drop tower-based standard was responsible for numerous improvements, including new cushioning materials and novel impact mitigating systems²⁸. While this methodology provided an important baseline, Breedlove et al. were the first to point out that not including the facemask led to underprediction of the peak accelerations at the rear of the helmet²⁹. In order to mitigate this shortcoming in the testing process, researchers began using the Hybrid III head form, originally developed for use in automotive crash tests, to evaluate protective gear for contact sports^30,31,32,33. Recent work has demonstrated that modern helmets mitigate up to 83% of the translational acceleration from a given impact to the front of the helmet, 80% to the side of the helmet, and 75% to the rear of the helmet^34,35. Mitigation of the rotational acceleration is considerably more variable, despite the fact that it is thought to be the primary correlate with tissue damage^{36,37,38,39,40,41,42,43}. Modern helmets mitigate up to 70% of the rotational acceleration from a given impact to the front of the helmet, and 82% to the side of the helmet, but only 48% to the rear of the helmet^34,35.

Based on previous results, the Riddell SpeedFlex (Riddell; Rosemont, IL) and Vicis Zero 1 (Seattle, WA) were the best-performing helmet models³⁵ with respect to both the translational and rotational accelerations. In 2022, a new version of the Riddell SpeedFlex was developed, as well as three new models of the Vicis: Vicis Zero 2; Vicis Zero 2 Trench, specifically designed for linemen; and the Vicis Zero 2 QB, specifically designed for quarterbacks. The current study was intended to evaluate the helmets that previously performed the best³⁵ and determine whether or not subsequent design modifications, including those intended to optimize the helmet design for specific positions, substantially improved performance. The overall goal of this study was to test the hypothesis that the newest helmets would perform better than the previous generation for both translational and rotational impacts.

Protocol

The consumables and the equipment used are listed in the Table of Materials.

1. Overview of data collection

Herein, four different helmet models from two manufacturers were utilized. These included the 2022 edition of the SpeedFlex (Riddell, Rosemont, IL), denoted by H1; the Zero 2 (Vicis, Seattle, WA), denoted by H2; the Zero 2 Trench (Vicis, Seattle, WA), denoted by H2T, designed for offensive and defensive line players; and the Zero 2 Quarterback (Vicis, Seattle, WA), denoted by H2Q, designed for quarterbacks, although it should be noted that the differences between H2 and H2Q were not obvious. For each helmet model, three separate helmets were tested for a total of 12 helmets.

The current study followed the basic procedure developed by Cummiskey et al.^34,44, and McIver et al.³⁵. Consistent with these previous studies, a 50^th percentile Hybrid III head and neck assembly testing rig, secured to a steel baseplate, was used to quantify forces and accelerations resulting from direct impacts to the bare head form as well as each of three helmets of each tested model. The impact mitigation, as defined here, was simply a measure of the decrease in acceleration due to the presence of the helmet.

2. Detailed process of data collection

After ensuring the DAQ and the computer that runs the LabView code were properly connected to power, all the sensors were plugged into the DAQ.

3. Software initialization

The custom LabView program, entitled "Hammer_Time_2017.vi" was double-clicked to initiate the start of the data collection. Basic data were then entered into the appropriate fields, including the path for the final data file, number of hits per file (typically five), number of files per location (typically four), and the headgear profile. Once the data were entered, the LabView code was initiated by pressing Run.

As a preliminary evaluation of the entire system, the modal impulse hammer was used to strike the front, top, and side in order to confirm that the hammer and all nine accelerometers were successfully collecting data.

4. Impact delivery

During the delivery of each impact, resultant accelerations at the center of mass (CoM) were measured using a nine-accelerometer array in a 3-2-2-2 setup, as defined by Padgaonkar et al.⁴⁵. A 200 ms time series was collected for each normal and oblique impact with a 70 ms pre-trigger, providing 130 ms of acceleration and impact force measurements, with all signals collected at 5120 Hz.

5. Impact types

Fourteen impact types (Figure 1) were evaluated for all test cases (bare head form and each helmet model), representing two angles of incidence (normal and oblique) at each of seven impact locations (Front, Front Boss, Side, Rear Boss, Rear, Top, and the SpeedFlex Cut Out). Normal impacts were delivered perpendicular to the surface, and oblique impacts were delivered at an angle of approximately 45^o to the normal direction. These impact types are hereafter denoted by: (1) Front-Normal, (2) Front-Oblique, (3) Front Boss-Normal, (4) Front Boss-Oblique, (5) Side-Normal, (6) Side-Oblique, (7) Rear Boss-Normal, (8) Rear Boss-Oblique, (9) Rear-Normal, (10) Rear-Oblique, (11) Top-Normal, (12) Top-Oblique, and, at the location of the SpeedFlex Cut Out, both (13) Cut Out-Normal and (14) Cut Out-Oblique (Figure 1). To simplify the discussion of the position-specific impact mitigation characteristics, two aggregated regions were also defined: the Frontal Aspect of the helmet encompassed impact types 1, 2, 3, 4, 13, and 14; and the Rear Aspect of the helmet consisted of impact types 7, 8, 9, and 10.

Impacts were first administered to the bare head form using a modally tuned impulse hammer (PCB Piezotronics, Inc.; Depew, NY) as described by Cummiskey et al.^34,44, to record the applied force during impact.

For each hit, the impact of the modal impulse hammer triggered the 200 ms acquisition window. After the data were visually inspected to ensure that there were no recording errors, the Save button was pressed to write the data to a file. Due to the violence of the experiments, most recording errors result from loose wires, a broken hammer cable, or broken accelerometer wires. For five hits per file, and four files per location, a total of twenty impacts were recorded at each location at levels 1 (2-4 Ns), 2 (5-7 Ns), 3 (8-10 Ns), 4 (11-13 Ns), 5 (> 14 Ns), with three repeats.

Subsequent to the acquisition of a bare head form reference set, three examples of each helmet were serially fitted to the head form, according to manufacturer specifications. All 14 impact types were delivered to each of the three helmets (Figure 2). These procedures resulted in a total of 840 data points being collected for each helmet model. Note that for the collection of the last, largest hammer strike at each location, a high-speed camera was used to document the blow at a frame rate of 959 Hz (Figure 3).

6. Post-processing

This study focuses on two important output parameters that result from impacts delivered to a Hybrid III head form: the peak translational and peak rotational accelerations, a_p,and , respectively. For the purposes of comparison, these parameters were recast in dimensionless form, , and , respectively, as described by Cummiskey et al.³⁴,
     (1)
    (2)

where t* is the difference between a reference time (100 ms) and the impact duration as measured by the impulse hammer, and w_n is the width of the neck. The impulse delivered to the head form (bare or helmeted) was also converted into a dimensionless input variable, , and given by,

    (3)

where m_h is the mass of the head form, and F(t) is the impact force as a function of time, integrated over the duration of the impact.

Acceleration traces were collected and processed using a custom MATLAB program. After collection, all traces were passed through a fourth-order Butterworth low-pass filter (750 Hz cutoff frequency) to reduce noise. Kinematic equations were then used to calculate resultant translational and rotational acceleration at the CoM^44,45. These values were output as dimensionless quantities to assess the response of the head form³⁴. Peak Translational Acceleration (PTA) and Peak Rotational Acceleration (PRA) were output as trace data for each recorded impact.

7. Statistical analysis

The current study also followed the data reduction and statistical analysis pipeline developed previously^34,35. Briefly, both , and , were shown to be related to the dimensionless impulse, , by a power law relationship⁴⁶. For the translational acceleration, this model has the form,

    (4)

A similar approach was used to model the dimensionless rotational acceleration. A modified version of Grubb's method was then used to remove outliers, and a final curve fit was generated³⁴. An ANCOVA test with an α level of 0.05 was used to examine differences between the regression coefficients for each helmet, with a Tukey post-hoc test and Holm-Sidak p-value correction^34,47.

The final intermediate asymptotic curves generated by this process were utilized to determine the impact mitigation for each helmet. Once the area under each curve was calculated using a total of 100 evenly spaced values spanning the distance between the minimum and maximum values for , it was possible to determine the effectiveness of the helmet at each location (Figure 4),

    (5)

This process was repeated for the peak rotational accelerations. With a maximum value of one (representing 100% attenuation), the higher the impact attenuation, the better the helmet performed. A change in mitigation of 0.05 was used as a threshold value, representing a physically meaningful effect size. Such an increase corresponds to 5% of the maximum possible mitigation and would roughly eliminate the effects of 25-50 head impacts for a typical athlete participating in an entire season of contact and accumulating a typical number of 500-1,000 head impact exposures^1,12.

Results

The masses of the four helmets tested herein ranged from 1.986 kg for the H2Q to 2.243 kg for the H2T (Table 1), which is the most massive helmet reported to date.

Over the course of the experiment, the peak force generated by the modal impulse hammer was 5,307 N.

Translational accelerations

As demonstrated previously^34,35, the addition of any helmet to the bare head form caused a significant decrease in peak dimensionless acceleration and a corresponding decrease in the regression coefficients relating the peak accelerations to the impulse delivered. Among the four helmets tested for this study, the translational impact mitigation ranged from 0.643 to 0.872 (Table 2). For the base models, the H1 exhibited a reduction in translational impact mitigation of at least 0.05 for five impact types (impact types 5 - 8, and 10) and did not exhibit an increase of at least 0.05 for any impact types. In contrast, the H2 demonstrated an increase in translational impact mitigation of at least 0.05 for five impact types (impact types 5 - 8, and 10) and no decreases of that magnitude.

On an impact type basis, the H2T achieved the maximum mitigation for eight of the 14 impact location/incidence types (Table 2). The H2Q performed best for two types: impact type 7 - tied with the H2T - and impact type 9. Interestingly, two helmets from a previous study, the Vicis Zero 1 and Riddell SpeedFlex 18, each performed best for two impact types (impact types 1 and 2). The H2 achieved the maximum mitigation for impact type 6, while the H1 did not register any of the top performances.

Rotational accelerations

The impact mitigation for rotational accelerations ranged from 0.444 to 0.882 for impact type 14 (Table 3). For the base models, the H1 exhibited a reduction in rotational impact mitigation of at least 0.05 for five impact types (impact types 2, 5, 7, 10, and 13) and an increase of at least 0.05 only for impact type 12. In contrast, the H2 demonstrated an increase in rotational impact mitigation of at least 0.05 for nine impact types (impact types 3-6, 8, 9, and 12-14) and a decrease of that magnitude only for impact type 11.

The H2 achieved the maximum mitigation for seven of the 14 impact types (Table 3). The H2Q performed best for two impact types: (impact types 2 and 4). The H2T performed best for impact type 12. Interestingly, two previously tested models performed best for two impact types apiece: the Schutt Pro VTD II exhibited the highest impact mitigation for impact types 7 and 10, while the Vicis Zero 1 performed best for impact types 1 and 11³⁵.

This study represents the first time that any football helmet tested according to this protocol has demonstrated impact mitigation greater than 50% for impact type 9. Three helmets, the H1, H2, and H2T, all surpassed that mark, with the H2 demonstrating the best response at 0.529 (Table 3). Interestingly, the H2Q only reached 0.473 for this impact type, which is substantially lower than the value for the base model.

On the Frontal Aspect (impact types 1-4, 13, and 14), the chief region of import to offensive and defensive line players, the H2T was demonstrably the best at mitigating translational accelerations, with the highest mitigation scores at four of the six locations (Table 2). In contrast, the H2T was less effective at mitigating rotational accelerations on the Frontal Aspect, being outperformed for every associated impact type (Table 3) by the Vicis Zero 1, H2, or the H2Q.

Across the Rear Aspect (impact types 7-10), the primary region of import to quarterbacks and some receivers, the best performing helmets were the H2 and the previously tested Schutt Pro VTD II.

DATA AVAILABILITY:

All data used for the calculations described herein will be made available through UC Figshare (DOI: 10.60696/uc.30661718).

Figure 1: Fourteen different types of impacts were delivered via a modally tuned impulse hammer to a Hybrid III head form. In the front view (A) and the rear view (B), normal impacts are denoted by black arrows while the gray arrows represent oblique impacts, which were delivered at an angle of approximately 45 degrees to the surface normal. The impact locations and types were denoted by (1) Front-Normal, (2) Front-Oblique, (3) Front Boss-Normal, (4) Front Boss-Oblique, (5) Side-Normal, (6) Side-Oblique, (7) Rear Boss-Normal, (8) Rear Boss-Oblique, (9) Rear-Normal, (10) Rear-Oblique, (11) Top-Normal, (12) Top-Oblique, and, at the location of the SpeedFlex Cut Out, both (13) Cut Out-Normal and (14) Cut Out-Oblique. Please click here to view a larger version of this figure.

Figure 2: A Front-Normal impact delivered to a common football helmet that demonstrates the peak deformation, the corresponding force (A), translational acceleration (B), and rotational acceleration (C), shown as functions of time.  Please click here to view a larger version of this figure.

Figure 3: The peak deformation observed during a Front-Normal impact delivered to each of the four helmets studied herein. The images demonstrate the relative stiffness of the H1 helmet shell and the relative compliance of the shells used to fabricate the H2 and H2Q. For the same impact magnitude, the H2T engages less of the helmet shell than does the H2 or the H2Q. Please click here to view a larger version of this figure.

Figure 4: Translational impact mitigation was measured using the peak translational accelerations (A) for the bare head form and the corresponding translational accelerations (B) for the helmeted head form for each helmet type. The presence of the helmet dramatically decreased both the measured accelerations. In order to calculate the impact mitigation, the peak accelerations for the bare and helmeted head forms were measured and used to calculate the dimensionless impulse, , and the dimensionless translational acceleration, . Integrating the areas under the curve for the bare head form, subtracting the corresponding area for the helmeted head form, and normalizing by the area under the bare head form provided a value between zero and one. Rotational impact mitigation was calculated similarly. Please click here to view a larger version of this figure.

Figure 5: Summary of translational and rotational impact mitigation for two of the 14 different impact types studied herein across all the helmets tested using this method to date. Mitigation of translational accelerations for Front Normal (Type 1) impacts (A) illustrates a subtle improvement in the Riddell SpeedFlex and little change in the Vicis helmets. Mitigation of translational accelerations for Rear Normal (Type 9) impacts (B) exhibited similar changes for the Riddell and subtle improvements for the Vicis helmets. Interestingly, both helmet types regressed slightly when it came to mitigating rotational accelerations for frontal impacts (C). The most dramatic improvements in rotational impact mitigation were observed for rear impacts (D) with both the Riddell SpeedFlex and Vicis Zero 2 exhibiting the highest values to date. Please click here to view a larger version of this figure.

Helmet Type	Mass (kg)	Shell Material
H1	2.074	Polycarbonate
H2	2.015	Thermoplastic elastomer
H2T	2.243	Thermoplastic elastomer
H2Q	1.986	Thermoplastic elastomer

Table 1: Mass and shell material for each helmet tested.

Impact Type		SpeedFlex	H1	Vicis	H2	H2T	H2Q
	Aspect	18		Zero 1
1	Frontal	0.705	0.706	0.832	0.821	0.82	0.799
2	Frontal	0.744	0.733	0.817	0.803	0.804	0.791
3	Frontal	0.75	0.751	0.786	0.798	0.805	0.779
4	Frontal	0.741	0.756	0.759	0.792	0.804	0.78
5		0.804	0.703	0.726	0.781	0.773	0.801
6		0.808	0.756	0.713	0.811	0.81	0.807
7	Rear	0.747	0.643	0.698	0.748	0.754	0.754
8	Rear	0.779	0.703	0.699	0.758	0.766	0.762
9	Rear	0.747	0.699	0.741	0.777	0.764	0.795
10	Rear	0.756	0.7	0.715	0.787	0.788	0.775
11		0.701	0.697	0.832	0.804	0.832	0.783
12		0.753	0.748	0.823	0.866	0.872	0.848
13	Frontal	0.741	0.719	0.817	0.834	0.836	0.819
14	Frontal	0.767	0.77	0.808	0.846	0.853	0.824

Table 2: Mitigation of translational accelerations by impact type. Gray shading indicates helmets tested as part of the current study. Bold numbers indicate the best performance to date for a given impact type.

Impact Type		SpeedFlex	H1	Vicis	H2	H2T	H2Q
	Aspect	18		Zero 1
1	Frontal	0.479	0.472	0.698	0.669	0.669	0.653
2	Frontal	0.55	0.444	0.727	0.716	0.696	0.738
3	Frontal	0.684	0.676	0.769	0.842	0.832	0.814
4	Frontal	0.728	0.738	0.773	0.848	0.843	0.853
5		0.811	0.758	0.783	0.863	0.84	0.846
6		0.811	0.775	0.801	0.871	0.853	0.858
7	Rear	0.73	0.585	0.698	0.747	0.736	0.666
8	Rear	0.578	0.613	0.479	0.622	0.594	0.595
9	Rear	0.483	0.52	0.391	0.529	0.505	0.473
10	Rear	0.701	0.621	0.693	0.706	0.692	0.688
11		0.564	0.57	0.723	0.67	0.698	0.642
12		0.619	0.75	0.769	0.846	0.853	0.762
13	Frontal	0.53	0.477	0.755	0.811	0.789	0.795
14	Frontal	0.685	0.69	0.778	0.882	0.878	0.855

Table 3: Mitigation of rotational accelerations by impact type. Gray shading indicates helmets tested as part of the current study. Bold numbers indicate the best performance to date for a given impact type.

Discussion

While the latest generation of helmets was found to mitigate 64%-87% of the translational accelerations, and 44%-88% of the rotational accelerations, the Rear Aspect of modern football helmets continues to exhibit poor impact mitigation in general, especially for impact type 9. Further, while the newer H1 did not, in general, perform better than the SpeedFlex 18 according to this testing protocol, the newer Vicis helmets did exhibit generalized improvement, with the rotational impact mitigation at the back of the helmet surpassing 50% for the first time.

Critically, however, the continued discrepancy in impact mitigation scores between the Frontal Aspect and the Rear Aspect of the helmet should be recognized as obviating the use of a single metric to determine the quality of a given helmet^{34,35,48,49,50}. The H2T, which performed well in mitigating translational accelerations but poorly in mitigating rotational accelerations, provides a clear illustration of the pitfall of using a singular metric for such a complex behavioral profile. Historically, modern football helmets have had success mitigating translational accelerations, but considerably more variable performance when considering rotational accelerations^34,35 (Figure 5).

It should also be noted that the design goals of the position-specific helmets do not appear to be met by the current iteration of helmets. For position-specific helmets, it is important to distinguish between translational accelerations, for many years the only test mode required by NOCSAE, and rotational accelerations, often thought to be the most relevant to the development of brain trauma^{36,37,38,39,40,41}. One of the strengths of the methodology used herein is that the translational and rotational accelerations are characterized, normalized, and nondimensionalized separately, making it possible to consider both in the overall evaluation and design of helmet performance. Another strength is the use of the bare head impacts to normalize the data set, effectively eliminating differences in neck stiffness that may be observed due to asymmetries in the Hybrid III neck structure and providing simple resultant translational and rotational mitigations that vary from zero to one at each location. This analysis is feasible because of the simultaneous integration of an impulse hammer to measure the force as a function of time delivered to the head form and an accelerometer array to measure the resulting accelerations.

While there has been considerable effort to characterize head accelerations during practices and games at many different levels of play^{21,51,52,53,54,55,56,57,58,59,60,61,62}, the magnitude of the forces generated during head impacts has not been characterized in situ. In order to provide an estimate of the peak contact forces experienced in football, it was noted that high school football players and women's soccer players exhibit nearly identical distributions of head acceleration events⁶¹. A laboratory-based study of head impacts in soccer⁶³ in which the measured peak impact force was approximately 4,600 N. Accounting for the added mass of the helmet, one would expect the peak forces to be roughly 30% higher. The peak contact force generated during this study was 5,307 N, which is within the expected range of 6,000 N.

Linemen typically experience the highest percentage of impacts to the frontal region^{1,12,59,64,65}. The H2T was the best at mitigating Frontal Aspect translational accelerations, but was outperformed by other helmets when rotational accelerations were considered. The extra material added to the front of the helmet increased its mass, likely helping it mitigate the translational accelerations, but appeared to interfere with the ability of the shell to deform, causing higher rotational accelerations. This is a surprising design strategy since it is well-documented that the Vicis helmet's primary energy dissipation mechanism is the flexibility of the shell.

In contrast, quarterbacks, and many receivers, often experience severe impacts to the back of the head. The H2Q performed best for two of the four Rear Aspect impact types when translational accelerations were considered, but none of the Rear Aspect impact types when rotational accelerations were considered. It is not clear why the H2Q did not perform as well in the Rear Aspect as compared to the base H2. There were no obvious differences in the structure of the helmets, but the manufacturer notes a difference in padding material in some, unspecified locations. Future work should examine all of the H2 models in more detail with an emphasis on the mechano-chemical analysis of the shells⁶⁶ and padding, as well as the development of high-resolution computational models to analyze the dynamic structural characteristics^{67,68,69,70,71,72,73}. The data collected herein would ideally be combined with previous mechanical characterization of the various materials that make up the helmet⁶⁶ for the generation of finite element analyses because the use of the modally tuned impulse hammer provides a well-defined force input to accompany the resultant acceleration. Ideally, this analysis would be performed within the context of a sensitivity framework^74,75,76,77.

The current method is acknowledged to exhibit some variability due to the fact that a human delivers each blow over a pre-defined range. To compensate, it was ensured that each mitigation metric is the result of 60 independent blows distributed over an established range of impulses. Consequently, this can be viewed as a feature of the testing protocol since on-field impacts are delivered with an inherent level of randomness with regard to their magnitude, location, and angle of inclination to the surface of the helmet. Another consideration is the size of the hammer face. For helmet-to-helmet contact, the hammer face acts over a larger area than does the helmet. In contrast, for helmet-to-ground, it generates a smaller contact area than does the turf. Future work should examine the nature of these impacts more precisely. A more general weakness of any helmet testing protocol is the lack of data quantifying the actual impact forces experienced on the field. Future work needs to address this issue experimentally, as it will help refine the range of impulses over which impact mitigation should be calculated for each location on the helmet.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Accelerometer wire	PCB Piezotronics	Model 030C10	Six required.
Adult football helmet size large	Helmet Types Listed Below	NA	Three of each model used to maintain external validity.
Data acquisition module	National Instruments	9234
Hybrid III 50th percentile male head assembly	Humanetics	78051-61X-H
Hybrid III 50th percentile male neck assembly	Humanetics	78051-90-H
LabView Software	National Instruments	NA
Laptop	Dell	XPS 16	Other models acceptable.
Large sledge soft rubber tip	PCB Piezotronics	Model 084A30.
Low noise hammer cable	PCB Piezotronics	Model 003D20
Modally tuned impulse hammer	PCB Piezotronics	Model 086D20
Steel plates	Midwest Steel and Aluminum	12"x12"x1"	Five plates drilled and bolted to form base for neck.
Triaxial accelerometer	PCB Piezotronics	Model 356A03
Triaxial Accelerometer Cable	PCB Piezotronics	Model 034G05
Uniaxial accelerometer	PCB Piezotronics	Model 352C22/NC	Six required.
	Riddell Speedflex		Denoted by "H1"
	Vicis Zero 2		Denoted by "H2"
	Vicis Zero 2 Trench		Denoted by "H2T"
	Vicis Zero 2 QB		Denoted by "H2Q"