# Anthropomorphic Test Devices; Q3s 3-Year-Old Child Side Impact Test Dummy; Incorporation by Reference > Briefs, arguments, decisions, and more. URL: https://www.frixlaw.com/law-library/documents/fr%3A2020-21478 ## Record - **Collection:** Federal Register - **Document type:** Rule - **Published:** November 3, 2020 - **Citation:** 85 FR 69898 ## Text DEPARTMENT OF TRANSPORTATION National Highway Traffic Safety Administration 49 CFR Part 572 [Docket No. NHTSA-2020-0088] RIN 2127-AL04 Anthropomorphic Test Devices; Q3s 3-Year-Old Child Side Impact Test Dummy; Incorporation by Reference AGENCY: National Highway Traffic Safety Administration (NHTSA), Department of Transportation (DOT). ACTION: Final rule. SUMMARY: This final rule amends NHTSA's regulation on anthropomorphic test devices (ATD) to add design and performance specifications for a test dummy representing a 3-year-old child, called the “Q3s” test dummy. The Q3s is an instrumented dummy that can assess the performance of child restraint systems in protecting small children in side impacts. Adding the Q3s provides NHTSA a new test device that can be used to improve side impact protection for children. DATES: The effective date of this final rule is: January 4, 2021. The incorporation by reference of the publications listed in the rule has been approved by the Director of the Federal Register as of January 4, 2021. Petitions for reconsideration: Petitions for reconsideration of this final rule must be received not later than December 18, 2020. The petition will be placed in the docket. Anyone is able to search the electronic form of all documents received into any of the agency's dockets by the name of the individual submitting the comment (or signing the comment, if submitted on behalf of an association, business, labor union, etc.). ADDRESSES: Petitions for reconsideration of this final rule must refer to the docket and regulatory information number (RIN) set forth above and be submitted to the Administrator, National Highway Traffic Safety Administration, 1200 New Jersey Avenue SE, Washington, DC 20590. Note that all petitions received will be posted without change to http://www.regulations.gov, including any personal information provided. To facilitate social distancing due to COVID-19, please email a copy of the petition to nhtsa.webmaster@dot.gov. Privacy Act: In accordance with 5 U.S.C. 553(c), DOT solicits comments from the public to better inform its rulemaking process. DOT posts these comments, without edit, to www.regulations.gov, as described in the system of records notice, DOT/ALL-14 FDMS, accessible through www.dot.gov/privacy . In order to facilitate comment tracking and response, the agency encourages commenters to provide their name, or the name of their organization; however, submission of names is completely optional. Whether or not commenters identify themselves, all timely comments will be fully considered. If you wish to provide comments containing proprietary or confidential information, please see below. Confidential Business Information: If you wish to submit any information under a claim of confidentiality, you should submit three copies of your complete submission, including the information you claim to be confidential business information, to the Chief Counsel, NHTSA, at the address given under FOR FURTHER INFORMATION CONTACT . In addition, you should submit a copy, from which you have deleted the claimed confidential business information, to Docket Management at the address given above. When you send a comment containing information claimed to be confidential business information, you should include a cover letter setting forth the information specified in NHTSA's confidential business information regulation (49 CFR part 512). To facilitate social distancing due to COVID-19, NHTSA is treating electronic submission as an acceptable method for submitting confidential business information (CBI) to the agency under 49 CFR part 512. https://www.nhtsa.gov/coronavirus. FOR FURTHER INFORMATION CONTACT: For technical issues: Peter Martin, NHTSA Office of Crashworthiness Standards (telephone 202-366-5668) (fax 202-493-2990), email Peter.Martin@dot.gov. For legal issues: Deirdre Fujita, NHTSA Office of Chief Counsel (telephone 202-366-2992) (fax 202-366-3820), email Dee.Fujita@dot.gov. Mailing address: National Highway Traffic Safety Administration, U.S. Department of Transportation, 1200 New Jersey Avenue SE, West Building, Washington, DC 20590. SUPPLEMENTARY INFORMATION: Table of Contents I. Executive Summary II. Background a. 2013 Part 572 NPRM and 2014 FMVSS No. 213 NPRM b. Comments on the 2013 Part 572 NPRM III. Summary of Differences Between the NPRM and This Final Rule a. Acceptance Criteria for the Qualification Tests b. Qualification Test Procedures c. Engineering Drawings and the Procedures for Assembly, Disassembly, and Inspection (PADI) IV. Response to Comments (Part I) on Developing the Regulation a. Copyright and Patent Issues b. Dummy Availability and Associated Data c. Developmental Stage of the Dummy d. Biofidelity e. Repeatability and Reproducibility (R&R) V. Post-NPRM Test Program Overview a. Test Locations b. Other Data c. Component Tests in the Post-NPRM Test Program d. Controlling Variability VI. Results of the Post-NPRM Test Program and the Final Acceptance Criteria for the Qualification Tests a. Background b. Process for Setting the Final Qualification Criteria c. Head d. Neck e. Lumbar Column f. Shoulder g. Thorax h. Pelvis VII. Response to Comments (Part II) on the Dummy Qualifications and Test Procedures a. Head Qualification b. Neck Qualification c. Arm Position VIII. Post-NPRM Data From Humanetics a. Qualification Tests b. Mass and Anthropometry Measurements IX. Drawing Package and PADI X. Other Issues a. Durability b. Consideration of Alternatives XI. Rulemaking Analyses and Notices I. Executive Summary This final rule amends NHTSA's regulation on anthropomorphic test devices (49 CFR part 572) by adding a new Subpart W that sets forth design and performance specifications and qualification tests for a test dummy representing a 3-year-old child, called the Q3s test dummy. The Q3s is an instrumented dummy that can assess the performance of child restraint systems in protecting small children in side impacts. The Q3s weighs 14.5 kilograms (kg) (32.0 pounds) and has a seated height of 556 millimeters (mm), and is representative of a 50th percentile 3-year-old child. The Q3s dummy's main parts (head, thorax, neck, shoulder, spine, abdomen, pelvis, and relevant instrumentation) and biofidelity are described in detail in a November 21, 2013 notice of proposed rulemaking (NPRM) preceding this final rule (78 FR 69944, 69946). NHTSA plans to use the Q3s test dummy in a proposed side impact test for child restraints. 1 1 NPRM to amend Federal Motor Vehicle Safety Standard (FMVSS) No. 213, “Child restraint systems,” January 28, 2014, 79 FR 4570. This final rule incorporating the Q3s into 49 CFR part 572 standardizes NHTSA's specifications on the dummy for testing and research purposes. Subpart W specifies a set of qualification tests and acceptance criteria for the Q3s's head, neck, shoulder, thorax, lumbar, and pelvis, assessing 35 response mechanisms for the dummy. 2 Additionally, Subpart W incorporates by reference a technical data package (TDP) for the Q3s consisting of a set of engineering drawings, a parts list, and a user's manual that has procedures for assembly, disassembly, and inspection (PADI) of the dummy. 3 Q3s dummies manufactured to meet the acceptance criteria for the qualification tests and the TDP will be uniform in their design, construction, and response to impact forces. 2 Test dummies specified in 49 CFR part 572 are subjected to a series of tests, called “qualification tests,” to ensure that their components are functioning properly. Conformity to the acceptance criteria for the qualification tests qualify the dummy as an objective and suitable test device for the assessment of occupant safety in compliance tests specified in the FMVSSs. Conformity assures that the dummy can respond properly in the compliance test, while non-conformance indicates the need for adjustment, repair or replacement. Qualification tests also monitor the response of components that may tend to deteriorate over time. For each test, certain dummy sensors and signal characteristics (such as the magnitude and timing) have been specified as qualification targets. By monitoring these sensors, the qualification tests assure that the dummy is functioning properly. Loose or damaged dummy hardware is often manifested in a signal that does not conform to the qualification targets, thus indicating that dummy maintenance may be needed. Conformity also assures that the sensors themselves are working properly. 3 The parts list, engineering drawings, and the PADI for the Q3s are available for examination in the docket for this final rule. As discussed in the November 21, 2013 NPRM, the Q3s was found to exhibit repeatable performance in CRS side impact sled testing and in component-level qualification testing. However, NHTSA acknowledged in the NPRM that the agency's findings in the proposed rule were based on only a few Q3s dummies then in existence. At the time of publication of the NPRM, the Q3s was a proprietary product owned by Humanetics Innovative Solutions Inc. (HIS), and HIS was the only source from which to obtain the dummy. NHTSA developed the Q3s NPRM based on NHTSA's testing experiences with four units that the agency had purchased from HIS. In the NPRM, the agency expressed a desire to examine more data on more dummies from multiple test labs and an expectation that it will “continue to collect qualification data” and “will examine all qualification data provided to us by commenters.” 78 FR at 69959. NHTSA received comments on the Q3s NPRM from the Juvenile Products Manufacturers Association (JPMA), Graco Children's Products, Inc. (Graco), Dorel Juvenile Group (Dorel), and HIS. Several commenters said they could not obtain the Q3s dummies from the dummy manufacturer HIS and so had little or no information about the ATD. Some expressed concern that the dummy's repeatability and reproducibility of performance were not assessed across various test facilities. Some asked for more data from tests with more dummies to round out the qualification corridors. In addition, the commenters made several technical comments relating to the ATD. Subsequently, in mid-2014, HIS began delivery of new Q3s dummies to end-users that included NHTSA, CRS manufacturers, and testing laboratories. In 2014 and 2015, to obtain more data on the Q3s, NHTSA undertook systematic testing of the new units from HIS, contracting with laboratories to carry out a full series of qualification tests with six Q3s dummies. The units included three of the agency's original four dummies together with new dummies manufactured in 2014. The agency set up a series of experiments designed to evaluate the performance of the Q3s in several different labs, examining the repeatability and reproducibility of the Q3s's performance. NHTSA designed the test program to assess all sources of variability, to quantify the degree of variability, determine its acceptability, and assess whether the underlying cause was a non-uniform test procedure at a lab (and among the labs), an aspect of dummy design, or the dummy manufacturer's production of Q3s units. Data from the tests were used to finalize the acceptance criteria for the qualification tests and ensure that a high level of repeatability and reproducibility (R&R) will be maintained henceforth. 4 4 The additional data also led to NHTSA's making some technical modifications to the proposed part 572 specifications, i.e., NHTSA removed the requirement for the pubic load in the pelvis impact test, revised aspects of the neck and lumbar tests, and corrected some of the drawings for the dummy. The agency discusses and lists the technical changes from the NPRM to this final rule below in this preamble. For this final rule, HIS has removed all proprietary rights to the Q3s. Single-source restrictions were in place during the NPRM stage (HIS retained rights to manufacture the dummy). However, the dummy drawings and designs are now free of any restrictions. This includes restrictions on their use in fabrication and in building computer simulation models of the dummy. Benefits and Costs The benefits associated with this rulemaking cannot be quantified. The incorporation of the test dummy into 49 CFR part 572, the first-ever child test dummy incorporated by NHTSA for use in side impacts, has the potential to significantly improve child passenger safety in motor vehicles. Adopting the Q3s gives NHTSA a tool to assess the performance of dynamic side impact protection requirements for child restraints using an ATD representative of children for whom the CRS is designed, and quantitatively evaluate the effectiveness of CRSs in preventing or attenuating head and chest impacts in side impacts. In addition, the availability of this dummy in a regulated format will provide a test tool that can potentially be used with other products designed to benefit children in side impacts. This final rule does not impose any requirements on anyone. NHTSA has proposed to use the Q3s in its compliance testing of the FMVSS No. 213 test under development, but even following adoption of the test, manufacturers would not be required to use the Q3s or assess the performance of their products in the manner specified in the standard. Child restraint manufacturers would be affected by this final rule only if they choose to use the Q3s to test their products. For entities choosing to own the Q3s, NHTSA estimates that the estimated cost of an uninstrumented Q3s dummy is approximately $50,000. Instrumentation installed within the dummy needed to perform the qualification in accordance with part 572, subpart W, adds approximately $20,000, for a total cost of about $70,000. Summary of Decision The data presented in the 2013 NPRM and obtained in NHTSA's post-NPRM test program demonstrate that the Q3s is a valuable tool for use in side impact testing. Adopting the Q3s into 49 CFR part 572 enhances NHTSA's efforts to reduce unreasonable risks posed by side crashes to children. II. Background a. 2013 Part 572 NPRM and 2014 FMVSS No. 213 NPRM On November 21, 2013, NHTSA published an NPRM proposing design and performance specifications and qualification tests for the Q3s, a new test dummy representative of a 3-year-old child for use in side impact testing (78 FR 69944). On January 28, 2014, NHTSA published an NPRM proposing to amend FMVSS No. 213 to add a new side impact test in which the Q3s would be used. The proposed side impact test applies to CRSs designed for children weighing up to 18 kg (40 pounds) (79 FR 4570). The proposal responds to a statutory mandate in the “Moving Ahead for Progress in the 21st Century Act” (MAP-21), 5 that NHTSA “issue a final rule amending Federal Motor Vehicle Safety Standard Number 213 to improve the protection of children seated in child restraint systems during side impact crashes.” These two NPRMs are referred to herein as the part 572 NPRM and the FMVSS No. 213 NPRM, respectively. 5 Section 31501(a) of Subtitle E, “Child Safety Standards,” of MAP-21 (July 6, 2012) (Pub. L. 112-141). b. Comments on the 2013 Part 572 NPRM NHTSA received comments on the part 572 NPRM from HIS, Graco Children's Products, Inc. (Graco), Dorel Juvenile Group, Inc. (Dorel), and the Juvenile Products Manufacturers Association (JPMA). Some of the comments on the FMVSS No. 213 NPRM discussed subjects pertaining to the part 572 NPRM, which NHTSA discusses in this document as appropriate. The commenters on the FMVSS No. 213 NPRM include Evenflo Company, Inc. (Evenflo), Britax Child Safety, Inc. (Britax), Consumers Union, Advocates for Highway and Auto Safety (Advocates), and Transport Research Laboratory, UK (TRL). Commenters on the part 572 NPRM discussed issues related to the following main areas: single source and patents; dummy and qualification data availability, biofidelity; repeatability and reproducibility of results (R&R); qualification test corridors, drawing errors; and test procedure protocols. These issues and NHTSA's responses to the comments are discussed below in this preamble. III. Summary of Differences Between the NPRM and This Final Rule a. Acceptance Criteria for the Qualification Tests A comparison of the acceptance criteria for the qualification tests (or “qualification limits”) in the NPRM versus the final rule is summarized in Table 1. All changes from the NPRM are discussed below in this preamble. The velocities and acceleration pulses of the impacting pendulums, which ensure that qualification test conditions are uniform, are unchanged from the NPRM. Table 1—Q3s Qualification Limits [NPRM vs. final rule] Test Measurement Units NPRM Final rule Head—Frontal Resultant acceleration G 250-297 255-300. Off-axis acceleration (Ay) G −20 to +20 −15 to +15. Head—Lateral Resultant acceleration G 113-140 114-140. Off-axis acceleration (Ax) G −20 to +20 −15 to +15. Neck—Flexion Maximum rotation deg 70-82 69.5-81.0. Time of max rotation msec 55-63 no req. Peak moment (My) N-m 41-51 41.5-50.7. Time of peak My msec 49-62 note 1. Decay time to 0 from peak angle msec 50-54 45-55. Neck—Lateral Maximum rotation deg 77-88 76.5-87.5. Time of max rotation msec 65-72 no req. Peak moment (Mx) N-m 25-32 25.3-32.0. Time of peak Mx msec 66-73 note 1. Decay time to 0 from peak angle msec 63-69 61-71. Neck—Torsion Maximum rotation deg 75-93 74.5-91.0. Time of max rotation msec 91-113 no req. Peak moment (Mz) N-m 8-10 8.0-10.0. Time of peak Mz msec 85-105 note 1. Decay time to 0 from peak angle msec 84-103 85-102. Shoulder Lateral displacement mm 16-21 17.0-22.0. Peak probe force N 1240-1350 1123-1437. Thorax with Arm Lateral displacement mm 23-28 22.5-27.5. Peak probe force N 1380-1690 1360-1695. Thorax without Arm Lateral displacement mm 24-31 24.5-30.5. Peak probe force N 620-770 610-754. Lumbar—Flexion Maximum rotation deg 48-57 47.0-58.5. Time of max rotation msec 52-59 no req. Peak moment (My) N-m 78-94 78.2-96.2. Time of peak My msec 46-57 note 1. Decay time to 0 from peak angle msec 50-56 49-59. Lumbar—Lateral Maximum rotation deg 47-59 46.1-58.2. Time of max rotation msec 50-59 no req. Peak moment (Mx) N-m 78-97 79.4-98.1. Time of peak Mx msec 46-57 note 1. Decay time to 0 from peak angle msec 47-59 48-59. Pelvis Peak pubic load N 700-870 no req. Peak probe force N 1570-1810 1587-1901. 1 Maximum moment occurs during the time interval while the rotation is within the specified interval. b. Qualification Test Procedures The agency made a few adjustments to the proposed qualification test procedures, which are summarized in Table 2 below. (Noteworthy changes are discussed in this preamble.) For simplicity, the English units that were shown in parentheses in the regulatory text of the NPRM are omitted. The qualification tests themselves are essentially unchanged from the NPRM. 6 6 The qualification tests have proven reliable and sound in qualifying the Q3s throughout the dummy's developmental stages and in qualifying virtually all other test dummies specified in part 572. Table 2—Summary of Revisions to Procedures Reg. text affected section Description of change § 572.212(c)(1) Head drop test Ambient temp. now 20.6-22.2 deg C. § 572.213(c)(1)(i) Neck flexion test, § 572.213(c)(2)(i) Neck lateral flexion test, § 572.213(c)(3)(i) Neck torsion test, § 572.217(c)(1)(i) Lumbar flexion test, § 572.217(c)(2)(i) Lumbar lateral flexion test Maximum moment now occurs when rotation is within the specified range. § 572.213(b)(3)(ii) Neck torsion test Correction on time = 0 definition. § 572.213(c)(2)(ii) Neck lateral flexion test, § 572.217(c)(2)(ii) Lumbar lateral flexion test Correction on specifying left vs. right mirroring in test setup figures. § 572.214(c)(4) Shoulder test, § 572.215(c)(4) Thorax with arm tests New steps to position arm against thorax. § 572.218(a) Pelvis assembly and test procedure, § 572.219 Test conditions and instrumentation Pubic load cell now optional since pubic criterion has been omitted. § 572.212(c)(4) Head drop test, Figures W1, W2 Surface finish: 0.2-2.0 microns RMS. § 572.212(c)(2)(ii) Lumbar lateral flexion test Headform sagittal plane perpendicular (not parallel) to the motion of the pendulum. Figures W6, W7, W8, W11 Correction on probe mass: Now 3.81 kg. Throughout regulatory text English units omitted. This final rule also corrects the following errors. The surface finish of the steel plate used in the head qualification test was not specified correctly in the NPRM. The correct specification is 0.2-2.0 microns root mean square (RMS). In the lateral lumbar qualification test, the proposed regulatory text was unclear in how it described the orientation of the headform, so it has been clarified. In Figures W6, W7, W8, and W11 of the proposed regulatory text, the probe mass was labeled incorrectly as 3.85 kg. The correct value is 3.81 kg. c. Engineering Drawings and the Procedures for Assembly, Disassembly, and Inspection (PADI) For this final rule, NHTSA has revised some of the engineering drawings to address discrepancies between the PADI and the engineering drawings, and some inconsistencies HIS noticed between the drawings it provided NHTSA for development of the NPRM and the dummies HIS produced. The changes are all valued-added revisions that either correct errors or provide missing information. They are not alterations that would change the dummy in any meaningful way or alter the dummy's response in either pre-test qualification testing or dynamic sled testing with CRSs. The changes to the drawings and the PADI are discussed in detail in Section IX below. A comprehensive listing of changes is described in the document, “Q3s Engineering Drawing Changes, Rev. J, May 2016.” 7 The design of the Q3s is essentially unchanged. 7 This document can be found in the docket for this final rule. IV. Response to Comments (Part I) on Developing the Regulation a. Copyright and Patent Issues HIS had certain property rights in the Q3s engineering drawings during the notice and comment period of this rulemaking. As discussed in the NPRM (78 FR at 69965-69966), during the notice and comment period, the Q3s engineering drawings used to fabricate the dummy were available in the docket for public review and comment, but most displayed the HIS name in the title block with a note restricting copying of or using the drawings other than for commenting purposes. NHTSA stated in the NPRM that the name, note, and all restrictions associated with the drawings will be removed at the final rule stage. Separately, in the NPRM, NHTSA noted its awareness that a patent application filed by HIS may cover certain parts of the Q3s dummy. Comments Received NHTSA received several comments expressing concern about the intellectual property restrictions on the dummy. JPMA and Dorel expressed concern that manufacturers will be bound to purchase a single-sourced dummy that is subject to patents and unregulated price points. NHTSA Response The Q3s specified in this final rule is free of any known copyright or patent restrictions. Although copyright restrictions were in place during the NPRM stage for the Q3s engineering drawings, all restrictions are removed for this final rule. The HIS name and the copyright note have been removed from all of the drawings. The dummy drawings are free of any restrictions and can be used in dummy fabrication and in building computer simulation models of the dummy. Moreover, there are no patents associated with the Q3s adopted by this final rule. 8 8 The patent issue was discussed in the NPRM (78 FR at 69965). Around the time of the NPRM, NHTSA became aware that HIS had filed a patent application with the United States Patent and Trademark Office potentially covering certain parts of the Q3s dummy. However, the patent eventually issued—for a rib cage incorporating a polyurethane material with a type of metal insert—is not used in the current design. ( See U.S. Patent No. 8,840,404 B2, “Rib cage for assembly for crash test dummy,” September 23, 2014.) Accordingly, the patent does not apply to the version of the Q3s specified in this final rule. b. Dummy Availability and Associated Data The difficulty in obtaining the Q3s was brought up in comments to both the part 572 and the FMVSS No. 213 NPRMs by several commenters. JPMA indicated it was not possible to learn of the strengths and limitations of the Q3s, particularly regarding its repeatability, reproducibility, and reliability. Graco, Britax and Evenflo indicated that the lack of availability of the dummy to the CRS industry and outside test facilities has prevented a more complete evaluation of the dummy across various test facilities and multiple CRS manufacturers. Dorel and HIS commented that more data from more dummies are needed to round out the qualification corridors. NHTSA Response It is true that the Q3s was generally unavailable from HIS during the original comment period which ended April 28, 2014. Because of that unavailability, on June 4, 2014, NHTSA reopened the comment period for the FMVSS No. 213 NPRM, granting a petition from JPMA (79 FR 32211). NHTSA agreed at that time to reopen the comment period until October 2, 2014, because the Q3s was slated to become widely available from HIS to CRS manufacturers around mid-2014. 9 Since mid-2014, the dummy has been available, as HIS has filled many orders for the Q3s since then. 9 Graco apparently was able to obtain and assess a new Q3s unit during the reopened comment period. In a comment on the FMVSS No. 213 NPRM, Graco states that it “supports the use of the Q3s ATD for side impact testing based on NHTSA's data and confirmed structural performance during the developmental testing period. Graco has been using the Q3s in our internal lab for about 6 months and we are satisfied with the overall performance of the ATD.” “Feedback Document for FMVSS 213 Side Impact [NPRM], Oct. 1, 2014, p.10. Regarding the qualification corridors, NHTSA concurs that development of qualification corridors is benefitted when more data are available on the ATD's performance in the qualification tests. In the NPRM for this final rule (78 FR 69959), the agency acknowledged that there was a limited amount of qualification data available to NHTSA for use in setting the proposed qualification limits. 10 NHTSA stated in the NPRM that the agency expected to receive qualification data from end-user commenters on the dummies tested at their own laboratories, and that, with those data, the agency would adjust the qualification limits to account for a greater population of dummies, and modify the test procedures as needed. 11 10 For the NPRM, NHTSA established qualification requirements based on replicate trials conducted sequentially on the four NHTSA-owned Q3s units at VRTC. These tests were used to set the upper and lower limits of the qualification corridors. They were initially set as follows: Either ±3 standard deviations from the mean or ten percent from the mean, whichever was narrower. Upper and lower bounds were then rounded to the next whole number away from the mean using three significant digits such that the final bounds were slightly wider than the initial bounds. NHTSA expected to refine and narrow the corridors when additional data was received on other Q3s units. 11 The adjustments made to the limits and procedures are listed Tables 1 and 2, supra. When data from users were not forthcoming because of the unavailability of the Q3s, NHTSA designed a test program to obtain the desired data once the dummy became available. In mid-2014, NHTSA borrowed three new Q3s units from existing owners (manufactured by HIS and delivered to end-users in mid-2014) to collect comparative qualification data with their new units. The agency systematically tested the three new units, as well as three of the agency's older units (manufactured in 2012 or before and used to develop the 2013 part 572 NPRM). NHTSA hired test labs to carry out a full series of qualification tests with the six Q3s dummies. The agency's design of experiments allowed NHTSA to assess the reproducibility and repeatability of the dummy and sort out sources of variability. NHTSA examined variability due to any non-uniform test procedure at each lab (and among the labs), variability in the dummy design, and variability in HIS's production of multiple Q3s units. Using this systematic process, NHTSA compiled the additional test data, and those submitted by other end-users, to set the acceptance criteria for the qualification tests for the Q3s. The post-NPRM test program is discussed at length in this preamble in Sections V and VI. c. Developmental Stage of the Dummy Comment Received The NPRM referred to the Q3s as the “build level D” iteration of the dummy (Build D). “Build level” is a term used by HIS to describe a specific revision level of the dummy relative to previous versions it sold. The Q3s drawings that HIS provided NHTSA prior to the publication of the NPRM were marked as revision level D. 12 12 In the TDP drawings placed in the NPRM docket, the HIS build level that HIS identified for the ATD is reflected in the top level assembly drawing of the Q3s, 020-0100 (sheet 1). This drawing shows that HIS marked revision level D in the title block. In its comment, HIS states that it considers the build level D dummy to be out of date, and that the dummy specified in a final rule should be referred to as “Build E.” HIS states that not using the “Build E” designation could cause hardship to its customers who might not know which version of the dummy they own, or who might erroneously assume that their build level D dummy is up to date when in fact the ATD “may be missing key updates.” NHTSA Response For the reasons set forth below, NHTSA declines to make the change. NHTSA does not believe that using the HIS naming conventions for this final rule is necessary or warranted. For the final rule, the agency has adopted a drawing package that has been periodically fine-tuned since publication of the NPRM in 2013 (discussed in sections below), so the revision level of the Technical Data Package had been updated from Revision (Rev.) D to Rev. J. We do not believe that NHTSA has to name the Q3s “Build E” to enable HIS to notify customers who bought Build D units built between December 2010 and November 2013 that their units may be missing key updates. HIS can use its sales records and customer outreach to determine which Q3s units its customers bought and which need updating. With those records and outreach, HIS can determine the type of conversion needed to bring the units up to date and facilitate their customers' updates of the previously-purchased ATDs. Comment Received Dorel believed that many aspects of the Q3s, such as the fixture used to run the neck torsion qualification tests, were not fully engineered, and are thus not finalized and ready for sale. Dorel also cited unavailability of specialized Q3s signal processing software as a hold-up to its dummy evaluation. NHTSA Response Dorel is mistaken in believing that the Q3s and its complementary fixtures used in qualification testing were not fully engineered. The NPRM for the Q3s provided all the information needed to assess the dummy in qualification tests, including complete engineering drawings of the neck torsion fixture. The neck torsion fixtures were not rights-protected in the NPRM for the Q3s. The agency knows of at least two other labs in addition to the agency's Vehicle Research and Test Center (VRTC) that have built them on their own (MGA Research Corporation (MGA)) and Calspan). With regard to Dorel's software concern, NHTSA has not developed specific software for the express purpose of processing qualification data for the Q3s or any other dummy. NHTSA does not provide software that would fully automate the processing of raw signals to determine the PASS/FAIL outcomes in each of the eleven Q3s qualification tests. Such software is a third-party product. As with all part 572 regulations, NHTSA specifies the test procedures, the test equipment, the instrumentation, and the filter frequencies of the test signals. The means to process the signals (in accordance with the part 572 specifications) is left to the discretion of each test lab. NHTSA does maintain a library of software tools that aid in the processing of raw signal data. 13 This includes a collection of Microsoft Windows graphical applications for analysis and processing of signal data. Core algorithms in this package include minimum/maximum applications, signal scaling, numerical integration, and digital filtering as specified by many FMVSS and part 572 standards (including Subpart W for the Q3s.) These tools may be used to process data generated in Q3s qualification tests. 13 http://www.nhtsa.gov/Research/Databases+and+Software. d. Biofidelity The part 572 NPRM discussed NHTSA's findings that the Q3s is suitably biofidelic overall and especially in the head, thorax and neck which are the body segments most critical for the intended use of the dummy in side impact testing. (78 FR at 69947-69950.) Comment Received In its comment, JPMA stated its belief that the Q3s's biofidelity is not representative of a 3-year-old, living child. JPMA stated 14 — 14 See Docket NHTSA-2013-0118-0008, page 2. As the agency is aware, its assessment of the Q3s focused on (1) a scaled-down version of post mortem adult human subject data, and (2) cadaver testing under dynamic loading. Unfortunately, the scaled-down adult data presumes incorrectly that adults and children are the same internally, which is simply not the case. For example, children's bones and bodies in general are much more flexible than their adult counterparts. Merely scaling adult data on the basis of mass, geometric and stiffness ratios will not represent accurate child-centered data. Therefore, while appropriate in size and weight to a live 3-year-old, the Q3s is not representative of live, reactive 3-year-old children. Due to the known differences between the Q3s and the children the ATD is supposed to represent, the developing side impact test standard carries with it a certain level of inherent risk — that child restraints built to comply with the new standard will be moving away from real-world effectiveness. NHTSA Response NHTSA's biofidelity assessment of the Q3s (provided in a report in the docket for the NPRM 15 ) compared the responses of the dummy to targets previously established for a three-year-old child. The targets themselves were published in a Stapp Journal article by the SAE Hybrid III Dummy Family Task Group. 16 15 See Docket NHTSA-2013-0118, “Biofidelity Assessment of the Q3s Three Year-Old Child Side Impact Dummy,” July 2012. 16 Irwin AL, Mertz HJ, Elhagediab AM, Moss S (2002). Guidelines for Assessing Biofidelity of Side Impact Dummies of Various Sizes and Ages. Stapp Car Crash Journal V46: 297-319, SAE International, Warrendale, PA. For ethical reasons, biomechanical response data on children under impact loading are very limited. Therefore, scaling techniques are necessary to derive the child impact response targets from laboratory tests on adult post-mortem human subjects (PMHS). 17 17 Aside from its response to impact, the size and shape of the Q3s is based on child anthropometry. The size and shape of the ATD is not scaled from an adult model or other dummy size. The SAE scaling procedure followed an impulse-momentum approach to derive response targets for a three-year-old from targets established previously for adults. The procedure made use of adult-to-child ratios of mass, anthropometry, and bone stiffness. In its comments, JPMA implied that this procedure does not account for differences in bone flexibility between adults and children. This is not the case. Differences in bone flexibility are integral to the scaling process, which employs adult-to-child bone stiffness ratios. For three-year-old vs. adult scaling, a bone stiffness ratio of 0.475 was applied. This ratio was derived using measurements of the elastic modulus of human bone samples from actual children as explained in the Stapp article. The scaling ratios were all applied to a lumped mass and spring model to arrive at biomechanical corridors for a three-year-old. Stated differently, the scaling theory used to establish the impact response of a human three-year-old does account for differences in flexibility and stiffness between adults and children. Details on the derivation of the scaling model and its application may be found in Mertz (1984) 18 and Mertz, et al. (1989). 19 NHTSA notes that the impulse-momentum approach was used for other part 572 child dummies, including the CRABI infant dummy 20 and the Hybrid III family of child dummies. 21 22 Thus, the biomechanical targets used to assess the Q3s were derived the same way as the targets for all other child dummies. Given the limitations on pediatric data, NHTSA believes the scaling process represents an appropriate, best available method of estimating the living, human child's response characteristics. 18 Mertz HJ (1984), “A procedure for normalizing impact response data,” Paper No. SAE 840884, Biomechanics of Impact Injury and Injury Tolerances of the Thorax-Shoulder Complex—PT-45, SAE International, Warrendale, PA. 19 Mertz HJ, Irwin AL, Melvin JW, Stalnaker RL, Beebe MS (1989), “Size, weight, and biomechanical impact response requirements for adult size small female and large dummies,” Paper No. SAE 890756, Automotive Frontal Impacts, SP-782, pp 133-144, SAE International, Warrendale PA. 20 Melvin JW (1995), “Injury assessment reference values for the CRABI 6-month infant dummy in a rear-facing infant restraint with airbag deployment,” Paper No. SAE 950872, SAE Congress and Exposition, Detroit, pp 1-12, SAE International, Warrendale PA. 21 Kleinberger M, Yoganandan N, Kumaresan S (1998), “Biomechanical considerations for child occupant protection,” 42nd Annual Proceedings for the Association for the Advancement of Automotive Medicine, pp 115-136, Charlottesville, VA. 22 Mertz HJ, Jarrett K, Moss S, Salloum M, Zhao Y (2001), “The Hybrid III 10-year-old dummy,” Paper No. 2001-22-0014, Stapp Car Crash Journal, V45, SAE International, Warrendale, PA. To summarize, NHTSA believes that the scaling process used to derive biomechanical response targets for the Q3s is well-founded and reasonable. The scaling process does not presume that adults and children are the same internally. The process assumes that the response of the targeted subject depends on its internal stiffness, and that internal stiffness varies by the age of the subject. The agency is satisfied with the overall biofidelity of the Q3s and is convinced that CRSs built to comply with the new side impact standard using the Q3s will be effective in the real world. Q3s Shoulder NHTSA evaluated the biofidelity of the Q3s shoulder in component testing under the loading of a pendulum. In the NPRM, NHTSA described an “unpadded” test conducted involving an SAE International protocol (Irwin, 2002) that uses a rigid pendulum in a pure lateral direction. In the test, the Q3s shoulder showed high stiffness with respect to lateral shoulder displacement and probe force under this test protocol. NHTSA later reexamined shoulder biofidelity under “padded” conditions that the agency believed corresponded more closely to the planned use of the Q3s in the proposed FMVSS No. 213 test than the unpadded condition. In the latter test, NHTSA used the Ohio State protocol (Bolte et al., 2003), which utilizes the same impactor mass and speed as the SAE International test but with foam padding attached to the impactor face. NHTSA determined that the latter condition was particularly relevant because the Q3s would most likely be exposed to a padded side structure (“wing”) of the child restraint in the test. 23 The striking surface, like the probe in the Ohio State test, would be padded. 23 CRSs subject to a side impact test would likely use padded side wings as one of the main countermeasures to meet side impact protection requirements. Under the Ohio State protocol, the shoulder of the Q3s was also stiff when assessed for biofidelity as measured by its deflection (about 10 mm below the nominal biofidelity target). However, NHTSA found that the magnitude of the force applied by the padded probe (about 400 N) was well within the upper and lower limits of biofidelity. Therefore, NHTSA believed that the Q3s's shoulder loading of the child restraint, which could affect the overall motion of the dummy's upper torso and head (relevant for the measurement of injury criteria under consideration), was representative of an actual human. (78 FR at 69949-69950.) Comment Received JPMA commented that it believed the shoulder of the Q3s is too stiff relative to a human child. The commenter stated that, because the shoulder is too stiff, the trajectory of the head during a compliance test will be unrealistic such that it could register artificially high HIC values. JPMA asserted that child restraint designs will thus need to be ultra-conservative in their ability to keep HIC low, and that this, in turn, could necessitate a seat design that is uncomfortable for children. JPMA was concerned that, to get comfortable, children may take on seating postures that could ultimately put the child at higher risk than when seated in a current CRS ( i.e., one that is not designed to meet a new side impact requirement). The commenter did not did not provide any data or analysis supporting these views. NHTSA Response It is important to highlight the point made in the NPRM that, under conditions that correspond closest to the intended use of the Q3s in the proposed FMVSS No. 213 side impact test ( i.e., using a foam-covered probe that is more akin to the shoulder interaction with a CRS “wing”), the force response of the padded probe (external biofidelity 24 ) nearly matches the target. 25 With the magnitude of the force generated by the padded probe well within the envelope for a biofidelic response, these data show that the Q3s shoulder is biofidelic in the manner in which it will exert force on the CRS. Thus, this loading of the child restraint, which would affect the overall motion of the dummy's upper torso and head (through which the FMVSS No. 213 injury criteria under consideration would be measured), is representative of an actual human. JPMA did not provide any analysis or rationale supporting its conclusions that the Q3s shoulder will cause artificially high HIC values and that uncomfortable seat designs will result. Given all available data and information about the test dummy, NHTSA is satisfied with the biofidelity of the Q3s shoulder and how the ATD's shoulder, head and torso will interact when the dummy is restrained in a child restraint in the side impact test. 24 For pendulum impacts, biofidelity is generally assessed as “external” or “internal.” External biofidelity is related to the force generated on the face of a pendulum impact probe upon striking a subject. In other words, probe forces generated by dummies are compared against probe forces generated by PMHS. Internal biofidelity is related to a measurement on or within the subject itself, such as shoulder deflection or spine acceleration, for which corresponding measurements are made on both the PMHS and the test dummy. 25 78 FR at 69949. “Biofidelity Assessment of the Q3s Three-Year-Old Child Side Impact Dummy,” July 2012, Docket No. NHTSA-2013-0118. e. Repeatability and Reproducibility (R&R) A test dummy's R&R may be assessed in sled tests and component tests. “Repeatability” is defined here as the similarity of responses from a single dummy when subjected to multiple repeats of a given test condition. “Reproducibility” is defined as the similarity of test responses from multiple dummies when subjected to multiple repeats of a given test condition. Sled tests establish the consistency of the dummy's kinematics, its impact response as an assembly, and the integrity of the dummy's structure and instrumentation under controlled and representative crash test conditions. In component tests, the test conditions as well as the test equipment are carefully controlled to assure the dummy is subjected to a tightly controlled impulse and to minimize external effects on the dummy's responses. Assessment of R&R NHTSA's assessment of R&R was based on a statistical analysis of variance. The percent coefficient of variation (CV) is a measure of variability expressed as a percentage of the mean. The CV is calculated as follows: 26 Standard deviations are based on a sample and calculated using the “n-1” method. ER03NO20.021 NHTSA has used CVs to assess the repeatability and reproducibility of ATDs throughout the history of part 572, starting in 1975. 27 Separate CVs for repeatability and reproducibility, by labs and by dummies, were computed. The CVs were used to assess the degree to which the current population of Q3s dummies were able to attain targeted responses. In the NPRM, we described how provisional upper and lower limits for all qualification requirements were set at a maximum of 10% (before rounding) from a nominal response target. For any particular requirement, the 10% condition was always met in our post-NPRM testing when the CVs were all below 5% for repeatability and 6% for reproducibility. Under these circumstances, there is a high degree of uniformity in the construction of the dummy components being tested and in the procedures followed by the labs for that test requirement. 27 See NPRM for the original subpart B Hybrid II 50th percentile male ATD (40 FR 33466; August 8, 1975). For example, in the post-NPRM test series for neck flexion, neck moments from 81 trials were recorded. In all 81 trials, the neck moment was well within 10% of the nominal target and the CVs were all below 5% for repeatability and below 6% for reproducibility. Thus, in our post-NPRM assessments, when the CVs for a particular test condition were below 5% and 6% for repeatability and reproducibility, respectively, no further examination of the data or test condition was carried out. On the other hand, when a test condition produced a CV above 5% for repeatability or 6% for reproducibility, a response in at least one trial was usually beyond 10% of the nominal target. When a CV exceeded 10%, several trials were beyond 10% of the target. In these instances, a close examination of the data, dummies, and procedure was performed to pinpoint the source of the variability. Corrective actions were taken in most cases. Our investigative criteria for repeatability uses a slightly lower CV than for reproducibility (5% vs. 6%) as shown in Table 3. Since repeatability is an assessment of the same dummy by the same test laboratory, whereas reproducibility is an assessment of multiple dummies at more than one lab, reproducibility assessments include many more sources of variability. Hence, repeatability CVs are generally lower than reproducibility CVs. Table 3—CV Score Categorization for Repeatability and Reproducibility 28 Repeatability CV score Reproducibility CV score Assessment <5% <6% No further investigation; all trials within ±10% of the target response. 5%-10% 6%-10% Sources of variability investigated. One or more trials beyond ±10% of target response. ≥10% ≥10% Corrective actions considered for revisions to test procedure or dummy design. Several trials beyond ±10% of target response. R&R in Sled Tests Since the Q3s dummy is being considered as a measurement device for a proposed regulatory test that would evaluate CRS performance in side impact crashes, NHTSA assessed the R&R of the dummy in actual CRS side impact sled tests. This assessment was discussed in the NPRM (78 FR at 69951-69953), where two Q3s units were tested five times each. Of the greatest importance to the assessment were the two measurements associated with injury assessment reference values for CRS requirements under the proposed side impact upgrade to FMVSS No. 213. These were the response of the head 29 and the lateral thorax displacement. 28 The assessment categories in Table 3 differ slightly from those applied during the NPRM stage. In the NPRM R/R analysis, a similar Table 3 categorized the CV ranges as either “Excellent,” “Good,” “Marginal,” or “Poor.” For this final rule, we do not use these terms in the table to describe the CV ranges. Rather, the new Table 3 provides further explanation of the action taken by the agency when the CV for a particular test condition was in a specified range, which, we believe, is more informative and helpful to the reader. Also, although the previous nomenclature for the CV ranges provided a convenient shorthand, we believe the terms it used could be misconstrued by the reader as reflective of a final assessment of the qualities of the ATD being tested. 29 The response of the head was measured by the acceleration of the head. Additionally, R&R of the head was also assessed via its injury correlate, the head impact criterion (HIC). HIC is computed from the head acceleration measurements. The CVs for the response of the head were less than 3% for all measures of R&R. For the lateral thorax displacement, the CV for reproducibility was also under 6%, and CV for repeatability was under 5% for one of the two Q3s units. For the other unit, the data in one of the tests was quite different from the others. This discrepancy was traced to an inconsistency in the pre-test position of the dummy's elbow in one of the tests which had resulted in a CV for repeatability of 9% for that unit. In consideration of the elevated CVs, NHTSA ran another (“supplemental”) series of sled tests with an improved arm-positioning protocol. This was also described in the NPRM (78 FR at 69952-69953). Five trials were run with a single unit. The repeatability for the thorax displacement in this series had a CV of 4%. The response of the head again was highly uniform, with a CV of 3%. Given this high degree of uniformity in those tests and since the design of the dummy was essentially unchanged, NHTSA was satisfied with the R&R of the Q3s in sled testing and determined there was no need to perform additional sled testing for a final rule. Comment Received In its comments, Dorel said that it computed a CV of 32.6% for HIC results from ten tests in the supplemental series. NHTSA Response The agency believes that Dorel may have misread the results of this series of tests. There were only five tests in this series, not ten as suggested by Dorel. None of the HIC values listed by Dorel correspond with those in NHTSA's test series, so it is unclear where Dorel's data were derived. The agency's test data are available to the public in NHTSA's Biomechanics Data Base (BIODB). 30 The CV in sled testing was only 3% for the HIC values. Given these data, Dorel's comment appears to be mistaken. In view of this high degree of uniformity, NHTSA is satisfied with the R&R of the Q3s in sled tests. 30 The Biomechanics data base may be accessed at: http://www.nhtsa.gov/research-data/databases-and-software. R&R in Component Qualification Tests In the NPRM, acceptance criteria for the qualification tests were proposed to assure that the high level of R&R exhibited in the sled tests would be preserved in any dummy presented for compliance testing. In other words, the qualifications would serve to weed out any dummy that had a substantially different response from the uniformity of the original four units. The proposed acceptance criteria were based on a series of eleven component tests with multiple Q3s units in replicate trials. An upper limit and lower limit for an acceptable response were set for each test. The limits were chosen to be wide enough to account for normal variations in dummy and laboratory differences, and narrow enough to assure consistent and repeatable measurements in compliance testing. As part of this analysis, R&R was assessed for each set of qualification test outcomes. As discussed in the NPRM, most CVs were well under 5% and all were under 10%. The agency was aware, however, that for the NPRM the assessment was carried out using only four units, with all tests run at a single laboratory (VRTC). NHTSA explained in the NPRM that the agency anticipated finalizing the Q3s limits based on additional qualification data we would receive subsequent to the NPRM (78 FR at 69959). Various commenters responding to the NPRM expressed the view that the repeatability and reproducibility assessment of the Q3s ought to be assessed across various test facilities. Some asked for more data from tests with more dummies to round out the qualification corridors. After the NPRM was published, NHTSA proceeded to obtain more qualification test data as it had planned. NHTSA investigated whether newer dummies tested at different labs exhibited the same level of R&R as NHTSA's original units. In the test program NHTSA designed in mid-2014, the agency used different labs to test both newer Q3s units and the original dummies, and obtained data that could be compared to the existing NPRM data from the original four units. In 2014 and 2015, NHTSA systematically tested three new units that HIS delivered to end-users and three of the agency's original four dummies. NHTSA examined the R&R of the Q3s's performance to assess all sources of variability so as to identify the degree of variability and whether it was due to a non-uniform test procedure at a lab (and among the labs), an aspect of dummy design, or the dummy manufacturer's production of Q3s units. This systematic approach enabled NHTSA to assess the potential to which factors resulting in the variability could be remedied, adopt measures to mitigate the variances where possible, and assess the quality of the data on the Q3s. The testing also provided data that helped round out the qualification corridors. The program is discussed below. Test results and analyses are discussed in detail in a NHTSA report entitled, “NHTSA's Q3s Qualification Testing, 2014-2015, May 2016.” 31 31 A copy of the report has been placed in the docket for this final rule. V. Post-NPRM Test Program Overview a. Test Locations NHTSA collected data from tests run at three different laboratories (Calspan, MGA and HIS) independent of NHTSA, and conducted additional tests at NHTSA's VRTC. At each independent lab, a full set of qualification tests were run (consisting of 11 different types of tests) on two NHTSA-owned units and a new unit. Several trials, or repeat tests, were carried out on each dummy for each of the 11 qualification tests. Tests were done using qualification test equipment owned by each laboratory. Tests were run in strict accord with the procedures described in the NPRM. The input parameters for each test had to conform to the specifications set forth in the proposed qualification procedures. For example, a test in which the probe impact speed did not meet the required parameters did not count toward the total test repetitions. After each test, a post-test inspection of the dummy was carried out to determine if the ATD incurred any damage resulting from the test. NHTSA Tests at Outside Labs—Calspan and MGA NHTSA contracted the services of Calspan and MGA to perform the series of qualification tests. The test series are summarized in Table 3. All tests were carried out between January through March 2015. NHTSA In-House Tests (VRTC) Prior to shipping NHTSA's two dummies to Calspan and MGA, NHTSA tested the ATDs to the qualification tests at VRTC, but only one trial per test condition was carried out. These results (in addition to those provided in the NPRM) served as a comparative baseline for subsequent tests on the same units at the outside labs. Also, the agency arranged with Britax to test its new Q3s dummy that Britax had received from HIS in 2014. The tests were conducted at VRTC, and the results were added to the data pool. Tests at HIS In addition to the data NHTSA itself collected, the agency was also given data by HIS. In 2014, NHTSA lent HIS two of NHTSA's Q3s dummies for HIS to use to compare its qualification procedures and equipment to that described in the NPRM. HIS ran the qualification tests and provided NHTSA with the data from the tests. The agency also obtained from Calspan, MGA and Britax the qualification results performed by HIS on the new Q3s units sold to those end-users. These data were supplied by HIS to each respective purchaser of the dummy at the time of delivery. The owners, in turn, provided the data to NHTSA. The test results were added to the data pool. Table 4, below, provides an overview of the qualification testing conducted at each lab. Table 4—Overview of Q3s Qualification Testing Lab Q3s serial No. Dummy owner Number of trials Year of tests Note VRTC 004 NHTSA 5 2012 Results shown in NPRM. 006 NHTSA 5 2012 Results shown in NPRM. 007 NHTSA 5 2012 Results shown in NPRM. 007 NHTSA 1 2014 Prior to HIS testing. 007 NHTSA 1 2015 Prior to MGA testing. 008 NHTSA 5 2012 Results shown in NPRM. 008 NHTSA 1 2015 Prior to MGA testing. 3538 Britax 5 2015 Leased from Britax. HIS 004 NHTSA 3 2014 Leased from NHTSA. 007 NHTSA 3 2014 Leased from NHTSA. 3538 Britax 2 2014 Pre-delivery to Britax. 5860 MGA 2 2014 Pre-delivery to MGA. 059 Calspan 2 2014 Pre-delivery to Calspan. MGA 007 NHTSA 5 2015 Contract with NHTSA. 008 NHTSA 5 2015 Contract with NHTSA. 5860 MGA 5 2015 Contract with NHTSA. Calspan 007 NHTSA 5 2015 Contract with NHTSA. 008 NHTSA 5 2015 Contract with NHTSA. 059 Calspan 5 2015 Contract with NHTSA. b. Component Tests in the Post-NPRM Test Program The component tests were the 11 qualification tests proposed for the Q3s. For each test, there were at least 2 dummy responses for a total of 35 in all. Of the 35 responses, 20 were derived from peak values (such as the peak resultant acceleration for the head drop test or maximum probe force for the pendulum tests). Those 20 were assessed for R&R. 32 The 20 measurements that NHTSA assessed for R&R encompassed each of the eleven types of qualification tests. 32 The other 15 were time-related criteria (such as the time peak at which the maximum neck rotation occurs) or criteria that contained zero in their intervals (such as the peak off-axis acceleration in the head drop test). NHTSA did not include these measurements in the R&R assessment because the CV statistical measure is not a good indicator of variability in these instances. c. Controlling Variability An assessment of dummy R&R is dependent on controlling variability within and among test labs in conducting the qualification tests. A dummy must provide repeatable and reproducible results in the tests, but a qualification test must be repeatable and reproducible to serve its purpose to either qualify or disqualify a dummy. 33 Controlling variability within and among test labs is important for assuring the qualification tests fulfill their purpose. 33 If a dummy is qualified, it can act as an objective device in compliance tests such as those proposed in the FMVSS No. 213 NPRM. If disqualified, a dummy must be replaced or repaired. With this in mind, when NHTSA collected post-NPRM data and observed variability in the test results, the agency closely analyzed any effect a test lab's internal practices, protocols and procedures might have had on the results. Variability caused by a lab's not being able to run a test repeatedly (“test repeatability”) is discussed in each section below. In addition, NHTSA assessed the objectivity of the test methods themselves, or “test reproducibility,” to assure that tests with the Q3s at different labs would produce reproducible results. NHTSA also identified instances in which repeatability was compromised due to a discernable problem with the dummy, such as variability in a particular dummy's responses over time (“dummy repeatability”). The agency also assessed “dummy reproducibility,” i.e., the uniformity of the dummies themselves. This is partly a function of how well HIS was able to manufacture dummies that behave uniformly. Thus, NHTSA was especially interested in comparing the responses of older versus newer units. The agency only used the results from the same lab for this assessment. Summary of Test Repeatability Assessment NHTSA assessed the ability of each of the three outside labs (Calspan, MGA and HIS) to attain a repeatable response by analyzing the effect test lab practices, protocols and procedures might have had on the results. Test repeatability was based on same-lab trials with the same dummy: Serial no. 007 (owned by NHTSA), the only dummy tested by all three labs. Thirty-five responses were assessed at each lab. Additionally, NHTSA performed a separate assessment at Calspan and MGA based on tests with NHTSA-owned dummy serial no. 008. (HIS did not test serial no. 008.) At Calspan, all test repeatability CVs were below 5% for all tests and for both dummies (serial nos. 007 and 008). At MGA, the CVs were below 5% except in two instances: The Mz measurement in the “Neck Torsion” test (5.9%) and in the resultant head acceleration in the “Lateral Head Drop” test (10.0%). Both occurred with dummy serial no. 007. All tests at MGA on serial no. 008 yielded CVs below 5% for test repeatability. At HIS (with serial no. 007 only), the CVs where below 5% in all but two instances: The “Lateral Head Drop” test (5.6%) and the “Thorax With Arm” test (9.3%). 34 34 The few instances where CVs for test repeatability were greater than 5% are discussed in greater detail below in this preamble. These findings demonstrate a high level of test repeatability and the ability of the three outside labs to carry out the qualification tests. In summary, NHTSA is confident in the data generated by the test labs in this test program. Summary of Test Reproducibility Assessment NHTSA assessed the objectivity of the test methods to provide consistent results at different labs. The agency evaluated test results from replicate tests on the same dummy (Q3s serial no. 007) at different labs (this ATD was the only unit tested at all four labs). NHTSA also assessed test reproducibility with Q3s serial no. 008, which was tested at VRTC, MGA, and Calspan (but not HIS). For all 35 sets of measurements, all but three had test reproducibility CVs under 6%. The three sets of tests that had CVs over 6% were: The resultant head acceleration in the lateral head drop test; the Mx component in the lateral neck test; and the pubic force in the pelvis test. 35 The results are discussed in greater depth in a later section below. 35 As will be discussed later in this document, NHTSA has corrected aspects of the lateral head drop and lateral neck test procedures that had contributed to the elevated variability in the results. Further, the agency has decided not to adopt the pubic force limit in the pelvis test. Summary of Dummy Repeatability Dummy repeatability is a measure of how much the response of a given dummy changes during the course of testing. One with a high degree of repeatability exhibits little change from one qualification trial to the next. A change in response could be caused by a hardening or softening of polymeric components over time or the propagation of cracks and other defects that occur over repeated impacts. Repeatability could also be affected by loose assembly tolerances. Dummies are routinely disassembled and re-assembled, and wide allowances for settings (such as the joint torques) could result in poor repeatability. During the course of the qualification testing of the Q3s, NHTSA closely examined the root cause of any variability in trial-by-trial test results that might reveal a problem with the dummy ( i.e., a problem with dummy repeatability) rather than simple test variability. There was only one instance where repeatability was compromised due to a discernable problem with the dummy. 36 This instance, which affected the uniformity of the lumbar spine, is discussed below, along with NHTSA's simple fix to the problem. Aside from that, there were no other problems with dummy repeatability in any of the tests. Once the fix to the lumbar was implemented, it was demonstrated to have a highly uniform response. NHTSA also examined changes in the response of the dummy over time and found that such changes had only a negligible effect on dummy repeatability. This is also discussed below. 36 Torn lumbar column. Throughout NHTSA's test experience with the Q3s, dating back to the NPRM, there was only one instance where dummy durability was an issue. In the very last series of tests on serial no. 008 run at Calspan in March 2015, a tear in the rubber column within the lumbar assembly was observed after the first lumbar qualification trial. In subsequent tests, the tear became visibly worse and the lumbar moment and rotation both increased with each successive impact. The biggest jump occurred between trials 1 and 2, where the maximum neck rotation jumped from being centered within the limits of acceptability to just outside the limits. The agency views this instance as a successful demonstration of the ability of the qualification test to weed out a damaged unit. Loosening of lumbar cable. NHTSA observed that in the lumbar flexion tests, the first trial tended to register a lower moment that subsequent trials. This was consistent with all dummies at all labs. NHTSA examined the wire cable that runs through the center of the rubber column, which was initially placed under tension by tightening a lock nut with a nylon insert 37 prior to the first trial. After the first trial, it was apparent that the nut did not stay in its set position. It could be loosened by hand. 37 A nut with a nylon collar insert, often referred to by its tradename, NYLOC, is a nut that resists turning. This affected the response of the lumbar spine, as the tension on the cable governs the response of the lumbar column. NHTSA controls this in the PADI by prescribing the torque for the nut on the center cable. However, the torque on a nut with a nylon insert is partly dependent on the condition of the nut itself. A newer nut can resist more torque without affecting the cable tension than a worn nut. In other words, the tension on the cable (and the moment) can vary depending on the condition of the nylon insert of the nut. To alleviate this situation, NHTSA has replaced the nut with two jam nuts, i.e., two standard nuts twisted against each other. No pronounced changes in response over time. NHTSA assessed also the agency's older unit, serial no. 007, for signs that one or more responses was exhibiting a definitive change during the course of testing due to any sort of deterioration. This unit was tested repeatedly over the course of many years, with the initial tests pre-dating the NPRM. NHTSA examined data from 2012 to 2015 to see if there were any definitive trends in response changes. To avoid any lab-to-lab variability that could act as a confounder, NHTSA assessed the results from a single lab, VRTC. Data were collected in three separate periods: In 2012 (five trials for the NPRM), in 2014 (one trial prior to sending it to HIS), and in 2015 (one trial just prior to the MGA/Calspan series). Of all the responses, only two had a definitive change in response over the three test periods: Lumbar moment and shoulder deflection. In these instances, the 2015 trial produced a lower/higher response than any of the previous trials (lower for the lumbar moment, higher for the shoulder deflection), while the 2014 trial produced a result that was between the 2015 and 2012 trials. Yet, even for these two instances, the change in response was negligible. For the lumbar moment, the change in moment was just 2 Nm: 82.6 Nm (lowest of the 2012 trials), 82.1 Nm (in 2014), and 80.6 Nm (in 2015). Similarly, the change in shoulder deflection was less than 1 mm: 19.0 mm (highest of the 2012 trials), 19.5 mm (in 2014), and 19.6 mm (in 2015). In both instances, all responses fell well within the qualification limits specified in this final rule. NHTSA observed no other problems with deterioration over time. In summary, NHTSA has determined that there are no problems with dummy repeatability that might compromise the overall uniformity of Q3s responses. The one problem with dummy repeatability has been resolved and there are no further concerns. Summary of Dummy Reproducibility Assessment In assessing dummy reproducibility, NHTSA examined the uniformity of the dummies themselves. This is partially a function of how well the manufacturer HIS produced dummies that behave uniformly. The agency was especially interested in comparing the responses of older vs. newer units. To eliminate the effects of lab-to-lab variability, NHTSA only used same-lab results for this assessment. NHTSA also combined results for left and right aspects since the dummy was designed to yield the same response in impacts to both. Thus, four separate assessments of dummy reproducibility were carried out, one per lab, against the units referenced in Table 5 below. Table 5—Q3s Dummies Used in Reproducibility Assessments at Various Labs Lab Serial numbers of older NHTSA units Serial numbers of new units VRTC 004, 006, 007, 008 3538 (Britax-owned unit). HIS 004, 007 3538 (Britax-owned unit); 5860 (MGA-owned unit); 059 (Calspan-owned unit). MGA 007, 008 5860 (MGA-owned unit). Calspan 007, 008 059 (Calspan-owned unit). As a secondary assessment, NHTSA compared only the three new units against each other in tests at HIS (HIS was the only lab that tested all three new units). This gave the agency a better sense as to whether the newer units, when considered as a single lot, had more inter-dummy variability as compared to NHTSA's original lot of four units. (As a point of reference, NHTSA assessed dummy reproducibility in the NPRM based on tests with the agency's four units (serial nos. 004, 006, 007, and 008) at VRTC and the CVs were less than 6% in all eleven qualification tests.) The agency's ratings of dummy reproducibility of the new units in the secondary assessment produced CVs in the 6% to 10% range for about 25 percent of the qualifications. The CVs of the other 75 percent were all under 6%, and no further investigation was performed. NHTSA investigated any set of tests with a CV above 5% for repeatability and 6% for reproducibility to determine the source of the variability. Responses in the lateral head drop and thorax impact test were non-uniform. When units manufactured since 2014 were compared to older units as two separate sets, NHTSA observed differences in responses for several qualifications. In general, the newer Q3s units did not exhibit the same high level of dummy reproducibility observed in NHTSA's four older units. As explained later in sections below, in a few limited instances, values obtained from a qualification test of a newer ATD were too dissimilar to those from tests of other Q3s units to be included within a set of reasonable qualification limits. Including them would have unacceptably widened the limits, lessened the uniformity of the ATDs, and unacceptably reduced the biofidelity of the Q3s. In such instances, the agency considered the particular dummy part substandard and the values from tests of the part beyond the performance criteria for the qualification test. 38 38 The high CVs for dummy reproducibility indicates that some newer Q3s dummies in the field may have to have parts reworked or replaced to produce a “pass” in the head drop test and thorax without arm test. Going forward, this final rule's setting of the acceptance criteria for the qualification tests should help provide checks and controls in the ATD's manufacturing processes, which in turn should facilitate the production of ATDs that meet the acceptance criteria for the qualification tests. VI. Results of the Post-NPRM Test Program and the Final Acceptance Criteria for the Qualification Tests a. Background In the NPRM, NHTSA proposed acceptance criteria based on replicate trials conducted sequentially on four NHTSA-owned Q3s units at a single laboratory (VRTC). These tests were used to set the upper and lower limits of the qualification intervals and were used to assess the repeatability of the Q3s. Of the 35 measurements, the bounds of 21 measurements were proposed as ±3 standard deviations from the mean. Of the 14 other measurements that were set to ±10%, 12 were set at ±2 standard deviations from the mean or greater. Two had bounds that were less than ±2 standard deviations: Peak pubic load (1.9 standard deviations) and peak neck torsion moment (0.5 standard deviations). At the time of the NPRM, NHTSA recognized that 3 standard deviations comprised a wider-than-usual bound from a probabilistic standpoint. NHTSA regarded the bound as a starting point based wholly on the statistics of the measurements. Three standard deviations were wide enough to account for normal variations in dummy and laboratory differences and narrow enough to assure consistent and repeatable measurements in compliance testing. Moreover, many of the bounds were, in practice, extremely narrow from an operational standpoint owing to factors (equipment, set-ups, technicians) lending themselves to highly repeatable testing at a single lab (VRTC). 39 NHTSA anticipated finalizing the Q3s limits based on additional qualification data the agency would receive subsequent to the NPRM (78 FR at 69959). 39 For example, the NPRM's 3-standard-deviation interval for the time at which the peak neck moment occurs was only 7 ms. b. Process for Setting the Final Qualification Limits The data from the post-NPRM test program and other sources, discussed above, have helped NHTSA finalize the qualification test procedures and round out the qualification corridors. In specifying qualification tests and acceptance criteria for the qualification tests, NHTSA's goal is to assure that a “pass” is a true indicator of a dummy that is uniform in its design and performance. This goal is achieved by ensuring that the tests themselves are repeatable and reproducible, and by setting limits (or tolerances) on the qualification targets. As discussed in the previous section, test and dummy R&R have been demonstrated at four different labs. The proposed targets and acceptance criteria for the qualification tests in the NPRM were based entirely on the statistics of the agency's replicate tests. NHTSA considered those targets and limits as starting points, given that the agency did not have data from other labs. Since then, the agency has expanded the qualification database by adding much more data on tests with several dummies across four test labs. For this final rule, the qualification targets and limits are based on the statistics of the measurements, but also on the following factors. Other Part 572 ATDs. NHTSA considered the qualification limits of the other part 572 ATDs in use today in setting those for the Q3s. For example, the qualification bounds for the most recent dummy incorporated into part 572 (the Hybrid III 10-year-old child dummy (HIII-10C); see part 572, subpart T), are derived from tests on about 30 different dummies, with data supplied from about ten different laboratories. For the HIII-10C, there are nine qualifications based on a maximum measurement (such as a peak force), and the average limits ( i.e., the values defining the range of acceptable measurements) are 9.9% from the midpoint. The low is 8.4% (neck rotation in the neck extension test) and the high is 10.8% (seen in two qualifications: neck moment in the extension test and chest deflection in the thorax impact test). A limit of 11% from the midpoint is the average for all part 572 dummies and all qualifications. NHTSA has used this value as a benchmark for setting the limits for the Q3s in this final rule. The agency scrutinized any limit above 11% from the midpoint to ensure it could be justified. Biofidelity targets. In setting the qualification limits, the agency considered the biofidelity targets that were used as the basic design criteria of the Q3s during its development. The corridors surrounding biofidelity targets are generally wider than qualification limits owing to larger variances associated with tests with human subjects. In the NPRM, NHTSA compared the responses of various Q3s body regions against their respective human biofidelity corridors. For the most part, the responses of the body regions fell within the biofidelity corridors (including the responses for the head and thorax). For the final rule, NHTSA made sure that a contemplated qualification limit would not result in acceptance of a dummy response that is outside the biofidelity corridors. Some body regions, such as the shoulder, were shown in the NPRM to be stiff relative to the biofidelity targets. For these body regions, any shifts in the qualification limits for the final rule were generally made in a direction that was closer to the biofidelity target. In other words, NHTSA avoided moving the nominal qualification target further from the biofidelity target. Test input parameters. For this final rule, NHTSA has not changed the input parameters in any of the eleven qualification tests from those of the NPRM. The input parameters include impact speeds, probe masses, drop heights, and dimensional measurements related to dummy positioning. Tolerances on test inputs are also unchanged. For this final rule, nineteen Q3s qualifications are centered around a maxima. For these measurements, the limits proposed in the NPRM were spread around a nominal target response by plus or minus 9.9% (on average) of the target. The average spread in this final rule is slightly higher, at 10.1%. However, as seen in Table 1, supra, the limits are narrower for 11 of the nineteen qualifications, and only the shoulder has limits greater than 12%: Internal shoulder deflection (12.8%) and shoulder probe force (12.3%). Newer dummies and other test labs. NHTSA considered the population of all dummies tested—both old and new—and all four labs that were used. Recognizing that the newest dummies may be representative of the future population of Q3s dummies, steps were taken to be inclusive of them as reasonably possible. NHTSA also recognized that all four labs were highly experienced in dummy qualification testing, so in theory any dummy that qualified at one lab should have qualified at the others. When this was not the case, the situation was analyzed to determine the source of the problem. Balancing the factors. In setting the final qualification limits for the final rule, NHTSA examined the test data on a trial-by-trial basis and balanced all the factors discussed above. For example, for the lumbar flexion qualification, while keeping the 11% goal in mind NHTSA set the qualification limits such that serial no. 059 (a new unit owned and tested by Calspan) was just under the upper limit in four of five trials, while serial no. 5860 (a new unit owned by and tested by MGA) was just over the lower limit in four of five trials. Balancing the factors enabled NHTSA to set qualification limits spread 10.9% from the nominal target in a manner that included as many test trials from the new units as reasonable. In contrast, if the 10.9% limits were centered around the average of all responses, the Calspan unit would have failed to qualify in all trials. In summary, the agency analyzed the data from the testing of the seven Q3s units (the four NHTSA-owned units and the three new units) to the qualification tests proposed in the NPRM, assessing, among other matters, the measurements made by the units when tested to the qualification tests and the R&R of the dummies. Tests were run for both right and left side impacts. Average, standard deviation, and coefficient of variation were computed for each required measurement parameter of each qualification procedure. c. Head The head injury criterion (HIC), based on the Q3s's head acceleration, has been proposed as a criterion in the FMVSS No. 213 side impact NPRM and is important for assessing countermeasures that protect the child's head in side impacts. Thus, a uniform response of the dummy's head-neck system is important to achieve. Two qualification tests serve to assure the uniformity of the head response in an impact: A lateral head drop test and a frontal head drop test. In both qualification tests, the pass/fail specification is based on the resultant acceleration measured at the center of gravity (CG) of the head. Procedures for both tests also place limitations on the off-axis acceleration to assure that the free-fall of the head is uniform prior to impact. Lateral Head Drop The lateral head drop test is carried out by cradling the head within a looped wire rope, suspending the head 200 mm above a steel plate, and releasing the wire rope. The head is oriented within the cradle so that its lateral aspect strikes the plate. Lateral impacts are carried out on the left and right aspects of the head. 40 40 Cradling of the head is shown in the regulatory text figures, but specifics on how to release the head are left to the operator. The NPRM proposed that the head must respond with peak resultant acceleration between 113 g and 140 g when dropped from a 200-mm height such that the side of the head lands onto a flat rigid surface (lateral head drop). Off-axis acceleration was proposed to be +/−20 Gs. These values were based on tests of NHTSA's four Q3s dummies. For the final rule, NHTSA has set the lateral qualification limits as: Peak resultant acceleration is 114-140 Gs (spaced 10.2% from the range's midpoint of 127 Gs). Off-axis acceleration: +/−15 Gs. These values are based on tests of the seven Q3s dummies. Test Repeatability. Test repeatability problems became apparent once the agency began to assess lateral head drop data from the outside labs. NHTSA believes that the problem existed even at the time of the NPRM as many of the CVs reported in the NPRM were just under 5%, which, upon reexamination, were high for such a simple test. None of the CVs for the frontal head drop was over 2 percent. The problem was first discovered in the initial tests performed at MGA on serial no. 007. Fourteen trials were needed to attain the desired sample of ten trials (five left, five right) in which the off-axis acceleration was under the NPRM's requisite 20 Gs (and only three of those were under 15 Gs). The CV for the resultant head acceleration was over 8% in the trial tests, which is unacceptably high. The variability was eventually traced to MGA's head drop apparatus. MGA had used a one-piece cable loop to cradle the head, and the cradle was released via a magnetic actuator. Upon release, the head rotated slightly during its free-fall creating elevated off-axis accelerations and high variability in the resultant accelerations. For its subsequent series of tests on serial nos. 008 and 5680, MGA developed an improved test protocol that included a two-cable cradle that mitigated the problem. Off-axis acceleration was below 20 Gs in all twenty trials and below 15 Gs in sixteen of the trials. 41 41 The cradle problem at MGA highlighted the need for a drop test mechanism with a high degree of precision. Any slight deviation in the point of impact was shown to produce a large variation in both the resultant and off-axis acceleration. This was particularly true in the lateral head drop, where the curvature of the head at the point of impact contributes to the variation. Calspan had similar difficulty with its drop apparatus, which made use of a pneumatic actuator to release the cradle. In its initial tests, Calspan needed nineteen trials to attain the desired sample of 5 left and 5 right trials with an off-axis acceleration under 20 Gs. However, like MGA, Calspan could achieve the 20 G limit in their subsequent series (with ten trials each with serial nos. 008 and 059). At VRTC, the cradle was released by cutting the end of the cable. There were no problems with keeping the off-axis accelerations below 20 Gs, though in retrospect it was still unusually high for such a simple test (the average was 12 Gs, with a range of 7-18 Gs). High off-axis acceleration was particularly problematic for serial no. 007 (one of the older, NHTSA-owned units) at all four labs where it was tested (53 trials total). NHTSA observed that the flesh parting line 42 on the head coincided with the point of impact, causing added variability for that particular unit (the effect was more pronounced with serial no. 007 than with other dummies.) About half of the tests with no. 007 produced off-axis accelerations greater than 15 Gs, with 13 tests (21%) greater than 20 Gs. Just 14 tests were less than 10 Gs. 42 When two halves of a mold meet, the corresponding line or seam appearing on the molded object is referred to as the parting line. When data from VRTC, Calspan, and HIS were further examined, it became apparent that elevated off-axis acceleration was correlated with high variability in the resultant acceleration. The scatter in data is evident in Table 6 (which represents all dummy tests, not just serial nos. 007 and 008). The CV in the resultant acceleration is shown to increase when the off-axis acceleration falls in higher ranges. It is highest (10.24%) when the off-axis acceleration is above 15 Gs and it is lowest (4.04%) when under 10 Gs. In the ranges of 0-10 Gs, 0-15 Gs, and 10-15 Gs, the CVs are all about the same and all under the 5%. Thus, NHTSA concludes that 15 Gs is a more appropriate limit than 20 Gs. 43 43 All NPRM upper/lower limits, including 20 Gs, were derived from the statistics of the tests. With the further data obtained in the post-NPRM program, NHTSA has determined that 20 Gs was too broad. Table 6—Relationship Between Off-Axis Acceleration and Variability in Resultant Acceleration Off-axis acceleration, Gs Number of trials Resultant acceleration Limits, % of midpoint CV (%) 0-5 0 0-10 21 7.7 4.04 0-15 84 10.2 4.47 10-15 64 10.2 4.58 0-20 114 16.2 6.38 10-20 94 16.2 6.71 15-20 30 16.2 9.20 Over 15 34 18.4 10.24 All 118 18.4 7.34 For this final rule, NHTSA has set the limit for off-axis acceleration to +/−15 Gs. NHTSA notes that this limit is the same as those for the two other part 572 side impact dummies (Subpart U—ES-2re (50th percentile adult male) and Subpart V—SID-IIsD (small adult female)). NHTSA believes the 15 G limit (as opposed to an even lower limit) is sufficient to assure dummy uniformity, and that lowering it to a lesser value is needlessly onerous on test labs because it will likely require many more trials to achieve acceptable test results. Unlike a frontal drop, where the direction of the drop is symmetric with the sagittal plane of the head, the lateral drop is asymmetric, making it difficult to attain an off-axis acceleration below 10 Gs. When only those tests where the off-axis acceleration was under 15 Gs were included, the CVs for repeatability and test reproducibility for the peak resultant acceleration were all 5% or less at all labs with all Q3s dummies. The agency notes that attaining the requisite +/− 15 G may require multiple drop tests. Nonetheless, in NHTSA's test program all labs could eventually attain this limit with each dummy they tested. Moreover, NHTSA believes it would be a relatively simple matter for labs to come up with a way to run the test such that the head does not slip and turn during its free fall, which should enable them to meet the 15 G off-axis limit without difficulty. Dummy Reproducibility. When assessing dummy reproducibility in the lateral drop test, for the reasons stated above the agency also omitted drop tests where the off-axis head acceleration is greater than 15 Gs, and the tests at MGA on serial no. 007. There was still an ample number of trials (84) without those tests to make a reasonable assessment of dummy reproducibility. The CVs for dummy reproducibility in lateral head drop tests at the various labs ranged for 7.0% to 11.7%, which reflects a fairly wide range of head acceleration responses. Nonetheless, the qualification criteria are set at 114-140 Gs, which reflects the upper and lower limits spaced only 10.2% from the midpoint. NHTSA concludes that the qualification limit of 10.2% is appropriately balanced to accommodate dummy reproducibility without being unreasonably hard for test labs to attain. The narrowness of the final limits is also consistent with other part 572 dummies, as shown in Table 7 below, and is needed to assure a sufficient level of uniformity in head response. As stated above, the head's acceleration is an important criterion for assessment of head injury. Thus, the acceptance criteria should be narrow enough to achieve a uniform response of the head-neck system of the Q3s. Table 7—Acceptance Criteria for Resultant Head Accelerations in Head Drop Tests for Various ATDs Dummy Aspect Resultant head acceleration Lower limit, G Upper limit, G +/− % of midpoint Q3s (final rule) Lateral 114 140 10.2 Q3s (proposed) Lateral 113 140 10.7 Side Impact Dummy Crash Test Dummy, Small Adult Female (SID-IIsD) Lateral 115 137 8.7 Side Impact Crash Test Dummy 50th Percentile Adult Male (ES-2re) Lateral 125 155 10.7 Q3s (final rule) Anterior 255 300 8.1 Q3s (proposed) Anterior 250 297 8.6 Hybrid III (HIII) 3-Year-Old Child Crash Test Dummy (HIII-3C) Anterior 250 280 5.7 Six-year-old Child Test Dummy (HIII-6C) Anterior 245 300 10.1 HIII 10-Year-Old Child Test Dummy (HIII-10C) Anterior 250 300 9.1 HIII 5th Percentile Adult Female (frontal) Test Dummy Anterior 250 300 9.1 NHTSA observed that the envelope of 114-140 Gs reflects the data from all the considered tests of the Q3s, but that two of the three newest dummies, those owned by Calspan and Britax, registered high head acceleration responses relative to NHTSA's older units and the newer MGA unit. NHTSA had to decide how to set the qualification limits for the head given the differences in dummy head performance. If NHTSA had set qualification limits to include at least one test trial from all dummies tested (the NHTSA-owned units and the three newer units), limits greater than 13% would have resulted. The agency was concerned that such limits would be too wide for regulatory purposes, especially because the Q3s's head acceleration measurements would probably determine a pass or fail in any future application of the dummy. No other part 572 ATD has limits wider than 11% for a head drop test (anterior or lateral). The agency also considered the possibility of calibrating the limits around the new units (which generally produced higher head accelerations) even though one or more of the NHTSA-owned units may not be able to qualify. When only the three new units were considered (combining data from tests at VRTC, MGA, HIS, and Calspan), limits within 11% were possible. After further investigation, however, NHTSA decided against this alternative too. The agency's first step in assessing whether to use only the new units was to assess the biofidelity of the new Q3s units. When the agency assessed the head of the Britax unit (which produced the highest response) against the biofidelity targets to confirm that it was within the limits of acceptability, the agency found it was not. The limits of biofidelity acceptance are generally wider than qualification limits owing to the variability associated with human subjects. As explained in the NPRM, the test to assess lateral biofidelity is slightly different from the qualification test (78 FR at 69949). Derived by SAE (Irwin, et al, 2002), the target response is referenced from the non-fracture zone of the head (opposite the point of impact). For a 3-year-old, the target resultant acceleration is 114-171 Gs. The test results for the NHTSA-owned units fell squarely within these limits. For the Britax unit, however, the tests produced a resultant acceleration of 189 Gs, which is well beyond the limits of acceptability. Thus, if the qualification limits were recalibrated around the newer units, the limits would be set based on readings of a non-biofidelic dummy. NHTSA decided that such an approach would sacrifice dummy biofidelity and is unacceptable. Accordingly, NHTSA decided that the final acceptance criteria for the lateral head drop qualification test should be centered around essentially the same midpoint as the NPRM. Thus, all NHTSA-owned units remain centered within the limits of acceptability. There is no potential sacrifice in biofidelity, unlike the result if limits were established around non-biofidelic Q3s units. NHTSA notes that, under the qualification limits of this final rule, a “pass” was observed with the older NHTSA-owned units at all labs and in almost every trial. Newly-manufactured Q3s dummies, on the other hand, did not always qualify. Of the three new units tested, only the MGA unit consistently produced a passing result against the final qualification criteria. The Britax unit was well above the upper limit, a result that was observed repeatedly in all trials at both labs in which it was tested. The Calspan unit was borderline acceptable. HIS had reported responses within the limits, but Calspan was not able to consistently produce a passing result at its lab. Given these results, there is a possibility that some dummy heads of newer Q3s units in the field may need to be re-worked to pass the lateral head drop criterion of this final rule. Frontal Head Drop The NPRM proposed that the head must respond with peak resultant acceleration between 250-297 Gs (8.6% of the midpoint) when dropped from a 376 mm height. The head is oriented such that its sagittal plane is parallel with the direction of impact and the anterior-most aspect of the forehead strikes a steel plate. Off-axis acceleration was proposed to be +/−15 Gs. These values were set based on tests of NHTSA's 4 Q3s dummies. For the final rule, NHTSA has set the frontal qualification limits as: Peak resultant acceleration is 255-300 Gs (8.1% of the midpoint). Off-axis acceleration: +/−15 Gs (no change from NPRM). These values are based on tests of the seven Q3s dummies. Test R&R. The CVs for test R&R were universally low at all labs and for all dummies (all below 4%). Unlike a lateral drop, the motion in the head in the frontal drop is symmetric about the sagittal plane, i.e., rotation of the head during and after the impact takes place about the y-axis only. This makes it much easier to produce a repeatable response and to attain a low off-axis acceleration. In the NPRM, the off-axis limit for acceleration was only 15 Gs (vs. 20 Gs for the lateral drop). The 15 G off-axis limit was easily met at all labs with all dummies. NHTSA notes that the 15 G limit for frontal drops is also consistent with other part 572 dummies, as shown previously. Dummy Reproducibility. For the frontal drop test, the CVs for dummy reproducibility were under 6% for all but one dummy—serial no. 5860, the MGA-owned unit. Relative to the others, the MGA head registered low responses at both labs (HIS and MGA) where it was assessed, resulting in an elevated CV statistic of 8.0% at HIS and 5.4% at MGA. If only the new units are considered (combining data from tests at VRTC, MGA, HIS, and Calspan), the CV statistic is 6.8% for all three units vs. 3.4% when the MGA unit is excluded. The Britax and Calspan units had high responses in the lateral drop tests but were in line with each other and with NHTSA's older units in the frontal head drop test. The lower limit of 255 Gs coincides with the lower limit of an acceptable biofidelic response as described in the NPRM. 44 At this limit, the MGA unit did not qualify in any of its seven trials at either of the two labs where it was tested (HIS and MGA), as its response was too low. The highest response it produced in any of the trials was 242 G, well below the biofidelity target. This response is unacceptably low (non-biofidelic). Aside from the MGA unit, only the Calspan unit was at all marginal. Its response was borderline low in tests at HIS (253 Gs on average), but at Calspan it was squarely within the limits. 44 “Biofidelity Assessment of the Q3s Three Year-Old Child Side Impact Dummy,” supra. NHTSA's final upper limit of 300 Gs (raised from 290 Gs in the NPRM) is still well within the acceptable biofidelity limit of 315 G. There were no problems staying under the upper limit for any dummy in any trial at any lab. By raising the upper limit to 300 Gs, NHTSA is maintaining essentially the same limit widths (8.1% of the midpoint) as those proposed in the NPRM. As noted above, a uniform head response for the Q3s is particularly important to assess child side impact protection. Thus, NHTSA has set the resultant acceleration limits for the frontal head drop narrower than the 11% guideline target for all responses. This approach is consistent with other part 572 dummies. The Q3s width of 8.1% ( i.e., the +/− limits of the nominal qualification target) is roughly the equivalent to the average of the other dummies. d. Neck A biofidelic and repeatable kinematic response of the head-neck system is important to quantify the protection offered by CRSs in an impact. The acceptable criteria for the neck qualification test in this final rule consist of three test components: Lateral flexion, frontal flexion, and torsion neck pendulum tests. These tests serve to assure uniformity of the head kinematics in both a head impact and non-impact. In each test, the neck moment, the rotation of the neck, and the timing associated with the moment and rotation are assessed. All three use the conventional part 572 swinging pendulum to apply a prescribed impulse to the neck, with a headform designed to mimic the inertial properties of the head attached to it. Lateral Flexion The lateral flexion test specifies a 3.8 m/s impact speed with a prescribed deceleration pulse. A column of collapsible aluminum honeycomb is used to decelerate the pendulum at a relatively constant level of force. Part 572 specifications for almost all other dummies use the pendulum/honeycomb device for testing necks. Test labs generally adjust the honeycomb in some manner (for instance, by modifying the number of cells engaged by the impacting face of the pendulum) to attain the prescribed pulse. The NPRM proposed a maximum rotation of 77-88 degrees (6.7% from the midpoint). The maximum moment was proposed to be 25-32 Nm (12.3% of the midpoint). This final rule sets the maximum rotation at 76.5-87.5 degrees (6.7% of the midpoint). The maximum moment is set at 25.3-32.0 Nm (11.7% of the midpoint). Test R&R and Dummy Reproducibility. All four labs exhibited CVs below 5% for test repeatability in lateral flexion for both the rotation and the moment. NHTSA did, however, observe some lab-to-lab variability in the bending moment which resulted in CVs for test reproducibility that ranged from 6.3% to 7.2% for both Q3s units that were used in the assessment. This was not entirely unexpected. 45 The variability in test reproducibility is likely attributed to lab-to-lab differences in the aluminum honeycomb, such as the lab modifications of the number of honeycomb cells used in the qualification tests. Also, after impact, the trajectory of the headform does not occur within a single plane of motion because the neck bends along its non-symmetric axis. This generally reduces test reproducibility. 45 In the NPRM, the set of limits for the moment was constructed via the +/−10% rule rather than +/−3 standard deviations. The agency did not discern any trends that would indicate that the responses of the necks have changed over time. Also, the CVs were under 5% for test reproducibility and under 6% for dummy reproducibility for all measures of neck rotation and neck moment. This further suggests that the variability is due to the variability in test equipment ( i.e., honeycomb) among the various labs. In summary, all dummies and all labs could demonstrate a qualification pass for both rotation and moment. The results show that the necks themselves were highly uniform, but test labs may need to evaluate different honeycomb configurations to demonstrate a passing response. Experimenting with honeycomb is typical of the qualification process with all part 572 dummies. Frontal Flexion The NPRM proposed a maximum rotation of 70-82 degrees (7.9% of the midpoint), and a maximum moment of 41-51 Nm (10.9% of the midpoint). For the final rule, the acceptance criteria for the frontal flexion test are set as: Maximum rotation is 69.5-81.0 degrees (7.6% of the midpoint). The maximum moment is 41.5-50.7 Nm (10.0% of the midpoint). The frontal flexion test specifies a 4.7 m/s impact speed and its own deceleration pulse. Crushing of aluminum honeycomb is also used to generate the prescribed deceleration pulse. Test R&R and Dummy Reproducibility. The CVs for test R&R and dummy reproducibility were universally low at all labs and for all dummies and for both neck rotation and neck moment (all below 4%). Unlike the lateral and torsion tests, the motion in the headform in the frontal flexion test is symmetric about the sagittal plane. In other words, rotation of the headform during and after the impact takes place about the y-axis only. This makes it much easier to produce a repeatable response and to attain a low off-axis acceleration. For the neck flexion test, the wide intervals specified in the NPRM (built around 3 standard deviations) proved to be unnecessarily large, even with the latest results from the additional dummies tested at different labs added to the data pool. Therefore, NHTSA has narrowed the limits for the final rule from those of the NPRM. All dummies at all labs were demonstrated to pass at the narrower limits of the final rule. Torsion During CRS testing, the Q3s neck might flex with varying degrees of neck twist. The agency, therefore, proposed a procedure to assure that the neck is uniform under twist. The proposed neck torsion test uses a special test fixture attached to the part 572 pendulum, which imparts a pure torsion moment to the isolated neck. It specifies a 3.6 m/s impact speed with a defined deceleration pulse. Qualification is based on the rotation and moment about the long axis of the neck. The NPRM proposed that, for the neck torsion test, the maximum rotation must be 75-93 degrees (10.7% of the midpoint). The maximum moment is 8.0-10.0 Nm (11.1% of the midpoint). For this final rule, the final acceptance criteria for the qualification test are set as follows. The maximum rotation limits are 74.5-91.0 degrees (10.0% of the midpoint). The maximum moment limits are 8.0-10.0 Nm (11.1% of the midpoint) (unchanged from the NPRM). Test R&R and Dummy Reproducibility. All four labs exhibited low CVs for test repeatability and reproducibility for both the rotation and the moment, with one exception. At MGA, the variability in neck moments on serial no. 007 was slightly elevated (CV=5.9%) for the left aspect only. However, this elevation is mostly a function of the low moment generated by the test (only 9 Nm nominally), where variations as little as +/−1 Nm created a high CV. All moments were, in fact, within the prescribed, and narrow, 8-10 Nm range specified in the NPRM. The CVs for dummy reproducibility were universally low (below 6%) at all labs and for all dummies, and for both neck rotation and neck moment. In every trial, all dummies at all labs demonstrated a pass in accordance with the acceptance criteria of this final rule. e. Lumbar Column The Q3s's rubber lumbar column bends during a CRS side impact test. This bending can affect the overall kinematics of the dummy, including the excursion of the head. It can also affect lateral loads and the deflection of the thorax. Lumbar qualification consists of two types of pendulum tests: A lateral test and a frontal test. For both tests, the lumbar spine element containing the flexible column is removed from the dummy, like the neck qualification tests. The lumbar tests use the same part 572 swinging arm pendulum and the headform device used in the neck qualification tests. The headform is not intended to represent the inertial properties of a body region as it is with the neck tests. Rather, it merely provides an apparatus that helps to ensure a repeatable test condition. The lumbar tests also use crushable aluminum honeycomb to attain a prescribed deceleration pulse. In the case of the lumbar qualification, lateral and frontal tests are conducted at the same impact speed of 4.4 m/s and specify the same pendulum impulse. The rotation of the lumbar column, the lumbar moment, and the timing associated with the moment and rotation are set forth in this final rule. The agency notes that the lumbar qualifications for lateral and frontal tests are almost identical. This is to be expected since the lumbar element is a circular cylinder constructed from an isotropic material (rubber), and so, theoretically, the directional properties should be the same for lateral vs. frontal bending. However, the agency has established two separate sets of acceptance criteria owing to possible dissimilarities brought on by the molding and bonding processes and asymmetries of inertial influences due to differences in the configuration of mounting plates and headform. Further, the frontal flexion test helps assure that the metal-to-rubber bond of the lumbar is intact in a manner the lateral flexion test does not. This was demonstrated during the very last series of tests on NHTSA-owned serial no. 008 Q3s dummy, where NHTSA observed a slight separation after the first of five trials. The subsequent trials all produced a rotation failing the limits of the NPRM and the final rule, whereas lateral flexion tests performed on the damaged part resulted in passes. That is, the frontal test detected the tear in the part, whereas the lateral test did not. Lateral Flexion This test mimics the main bending direction of the Q3s's torso during a CRS side impact test as proposed in the FMVSS No. 213 upgrade. This test assures uniformity in such bending. The NPRM proposed a maximum rotation of 47-59 degrees (11.3% of the midpoint). The maximum moment was proposed to be 78-97 Nm (10.9% of the midpoint). This final rule sets the maximum rotation at 46.1-58.2 degrees (11.6% of the midpoint). The maximum moment is set at 79.4-98.1 Nm (10.5% of the midpoint). Test R&R and Dummy Reproducibility. At all four labs, the CVs for test repeatability and test reproducibility were below 5% and 6%, respectively, for both the rotation and the moment with all dummies. For dummy reproducibility, however, the CVs were above 6% at two of the labs. Tests revealed that two of the newer units, the Britax-owned unit (tested at VRTC) and the MGA-owned (tested at MGA), produced greater rotations than the older NHTSA-owned units. As a result, the CVs for dummy reproducibility in lumbar rotation at VRTC and MGA were 6.5% and 7.4%, respectively. All dummies at all labs were demonstrated to pass the qualification limits of this final rule. The margins for acceptance are essentially the same as those of the NPRM, but the midpoints for both rotation and moment have been shifted slightly downward for rotation and upward for moment. Frontal Flexion The proposed FMVSS No. 213 side impact test is carried out at a slight oblique angle. Typically, the torso of the Q3s bends laterally and slightly forward, so NHTSA has included a frontal (forward) component to the lumbar qualification. The NPRM proposed a maximum rotation of 48-57 degrees (8.6% of the midpoint) in the NPRM. The maximum moment was proposed to be 78-94 Nm (9.3% of the midpoint). This final rule sets the maximum rotation at 47.0-58.5 degrees (10.9% of the midpoint). NHTSA set the maximum moment at 78.2-96.2 Nm (10.3% of the midpoint). Test R&R and Dummy Reproducibility. The CVs for test repeatability and reproducibility were under 5% and 6%, respectively, at all labs and all dummies for both rotation and moment. However, the new MGA-owned unit produced consistently higher rotations than the two NHTSA-owned units, resulting in a CV of 8.0% for reproducibility of the dummy's lumbar in rotation. At VRTC, the new Britax-owned unit had rotations that were also high, resulting in a CV dummy reproducibility score of 6.6%. At Calspan, its new unit produced consistently higher lumbar moments than the two NHTSA-owned units. Thus, the Calspan CV score for dummy reproducibility of the lumbar moment was elevated (7.7%). All dummies at all labs were demonstrated to pass the qualification limits of this final rule. In setting the new limits, NHTSA has slightly widened the margins for acceptance relative to the NPRM for both rotation and moment to accommodate the newer units. In both instances, the margins are still under the 11% goal. Timing Specifications Associated With Lumbar Qualification All pendulum tests for the lumbar column have specifications on the time at which the maximum moment and maximum rotation occur. The agency has revised the way signal timing is assessed for the lumbar column and neck qualification tests and has slightly increased the time that it takes the lumbar column (or neck) to return from its position at peak rotation to the position of zero rotation. The discussion of those issues can be found in the section below. Timing Specifications Associated With Neck and Lumbar Qualification 46 46 The following discussion also applies to the timing specifications for the lumbar column qualification tests. 1. Maximum Moment and Rotation All pendulum tests for the neck and lumbar column place specifications on the time at which the maximum moment and maximum rotation occur. This final rule revises the way signal timing is assessed. The test data indicate that the proposed time specifications were generally met. There were only a few instances where the peak time was just under or just over the prescribed interval. All the tests would have met the time specifications if the intervals were expanded by just 1 ms, except for the time specification for the maximum moment in the neck lateral flexion test (see Table 8 below). Here, 60 trials (about half of all trials) were below the NPRM lower limit. However, for this test, the range of allowable times was only a 7 ms interval, whereas the intervals in the other four tests ranged from 11 to 20 ms. The 7 ms time interval was very narrow because, along with all qualification intervals proposed in the NPRM, it was derived solely from the statistics of the then-available test data. The interval of 7 ms represented three standard deviations from the mean of data gathered during the NPRM stage. The very narrow time interval was the result of running the tests at a single lab (VRTC) under highly similar impulses and using aluminum honeycomb from a common lot. Table 8—NPRM Time Specifications for Neck and Lumbar Qualification Tests Qualification test NPRM time specifications Max. rotation (ms) Max. moment (ms) Number of trials with a time that differed from the NPRM time specifications Max. rotation (ms) Max. moment (ms) Neck frontal flexion 55-63 49-62 1 0 Neck lateral flexion 65-72 66-73 0 60 Neck torsion 91-113 85-105 0 2 Lumbar frontal flexion 52-59 46-57 2 1 Lumbar lateral flexion 50-59 46-57 0 1 The agency's latest pooling of test data reveals that the timing disparity in the neck lateral flexion test is related to lab-to-lab variability, not to test repeatability or dummy repeatability. For any given lab, the times are clustered within a very narrow interval of about 6 ms for all trials of all dummies tested at that lab. Thus, the timing discrepancy appears to be related to the test protocol, not dummy reproducibility. Time specifications in final rule. NHTSA could have expanded this interval by 6 ms (which would have put it in line with the other intervals in part 572), which would have resulted in a pass for all trials. However, rather than adjusting the NPRM time interval in that way, the agency has adjusted the way signal timing is assessed. For the final rule, the agency has adopted the same performance specification that is used for other part 572 child dummies (Subpart N—HIII-6C; Subpart P—HIII-3C; Subpart T—HIII-10C). 47 Instead of using time t = 0 as a reference for the maximum moment, the final rule specifies a range for the peak moment during the time interval when the rotation is above a specified limit. For neck flexion, the regulatory text specifies that Plane D, referenced in Figure W3 of Part 572, shall rotate in the direction of pre-impact flight with respect to the pendulum's longitudinal centerline between 69.5 degrees and 81.0 degrees and that, during the time interval while the rotation is within these angles, the peak moment measured by the neck transducer shall have a value between 41.5 N-m and 50.7 N-m. 47 Subpart N, Six-Year-Old Child Test Dummy, Beta Version (HIII-6C); Subpart P, Hybrid III 3-Year-Old Child Crash Test Dummy, Alpha Version (HIII-3C); Subpart T, Hybrid III 10-Year-Old Child Test Dummy (HIII-10C). Similar wording is used for the neck lateral, neck torsion, lumbar frontal, and lumbar lateral tests. All dummies passed the time specifications at all labs in all trials using this approach. This revised specification for the timing is better than what was proposed in the NPRM because lab technicians following the procedure would not have to pinpoint time = 0 as specified in the NPRM. In the NPRM, time t = 0 is defined as: “All instrumentation data channels are defined to be zero when the longitudinal centerline of the neck and pendulum are parallel.” In practice, determining the instant at which the parallel alignment occurs can be challenging, and has a significant bearing on a pass vs. fail outcome (as shown by the post-NPRM data, where it was not unusual that a pass vs. fail outcome was determined by less than 1 ms). Referencing a particular data point (the point of maximum rotation) identifies the reference time with greater precision. 2. Decay Times The specification for decay time specifies the time that it takes the neck or lumbar column to return from its position at peak rotation to the position of zero rotation. This specification is included in all other part 572 dummies mentioned previously. It serves to assure uniformity of the hyperelastic material used to construct the neck (or lumbar column). It also ensures that the later part of the impulse brought on by the collapse of the aluminum honeycomb structure is uniform. The NPRM proposed decay times listed below in Table 9. In about 15% of the post-NPRM trials, the NPRM decay times were not met for neck and lumbar frontal flexion. Expanding the NPRM decay interval by only a few milliseconds results in PASS in all trials for all dummies at all labs. Table 9—Q3s Moment Decay Times for Neck and Lumbar Qualification Tests Test NPRM decay time, ms Final rule decay time, ms Number of trials that differed from the NPRM decay time specifications Neck frontal flexion 50-54 45-55 10 Neck lateral flexion 63-69 61-71 0 Neck torsion 84-103 85-102 0 Lumbar frontal flexion 50-56 49-59 11 Lumbar lateral flexion 47-59 48-59 0 Decay time in final rule. The decay intervals for the final rule are listed in Table 9. In qualification tests for other part 572 dummies, the intervals for neck decay times ranged from 10 to 35 ms. NHTSA considers 10 ms a practical lower limit on the interval, accounting for the precision of the measurement system of any given lab. Thus, the decay times have been adjusted so that the intervals are no narrower than 10 ms. With these time intervals, all dummies met the decay time interval at all labs in all trials. f. Shoulder This test assures that the shoulder acts uniformly in the way it deforms under load and distributes the load under a lateral impact during CRS testing, thus helping to ensure that whole-body kinematics are consistent. Shoulder qualification is accomplished with a lateral impact to the shoulder using a 3.8 kg probe at an impact speed of 3.6 m/s. Conformity is based on the maximum probe force and the maximum deflection of the shoulder, as measured by a potentiometer installed within the dummy. The NPRM proposed that the peak probe force must be 1240-1350 N (4.3% of the midpoint), and that maximum displacement of the shoulder must be 16-21 mm (13.5% of the midpoint). This final rule sets the peak probe force to be 1123-1437 N (12.3% of the midpoint). Maximum shoulder displacement is 17.0-22.0 mm (12.8% of the midpoint). Test R&R and Dummy Reproducibility. The CVs for test repeatability and reproducibility were below 5% and 6%, respectively, for the measurements of probe force and shoulder displacement with all dummies at all labs. However, compared to the other three labs, the probe forces in tests at HIS were consistently higher for the newer dummies, whereas for the older NHTSA units, test repeatability at HIS had noticeably more scatter. This trend may have been related to arm positioning. During the latest testing series, NHTSA realized that, contrary to the agency's intent, the Q3s's upper arm can meet the position setting described in the NPRM in both medial/lateral rotation and in ab/adduction. In other words, the NPRM did not specify a unique position for the upper arm. To address this, in the final rule, there are more instructions in the dummy positioning procedure for the shoulder test as to where to position the Q3s's elbows and arms. This simple step should result in better R&R of the qualification test. The CV for dummy reproducibility of the shoulder force was elevated in three of the assessments (ranging from 6.1% to 7.8%). Two of the newer units—5860 owned by MGA and 059 owned by Calspan—were different from the others in that they produced lower probe forces, particularly for the left aspect. This has resulted in slightly expanded qualification limits for the shoulder. While the limits for probe force have been widened, the midpoint is essentially the same. At 12.3%, the limits are now wider than the 11% goal, but still considerably narrower than those of other part 572 side impact dummies (the limits for the ES-2re and SID-IIsD are both 16%). 48 Also, there is no immediate injury reference value directly related to the shoulder in the proposed FMVSS No. 213 side impact test, so its uniformity is less important. 48 The NPRM limits for probe force were at 4.2%, but they were unusually narrow, even considering that all data was gathered at a single lab (VRTC). There is no limit narrower than 5% for any part 572 qualification requirement (displacement or otherwise). For shoulder deflection, the range of the limits is essentially the same as those of the NPRM, but they have been shifted upward to allow greater deflection. NHTSA considers this an improvement to the specification. From a biofidelity standpoint, the shoulder is stiff relative to a human. Shifting the deflection limits upward (rather than downward) is consistent with a more biofidelic response. The 12.8% shoulder deflection limits sound relatively wide, but are not of concern because they are a function of the low level of deflection seen in the test (only 17-22 mm). This 5 mm interval is lower than that of any deflection-based limit of any other part 572 d [Text truncated at 120,000 characters. The full text is on the page linked above.] --- Source: Frix Law Library, https://www.frixlaw.com/law-library/documents/fr%3A2020-21478. Public record. Not legal advice.