PULSCAR ANALYSIS OF NHTSA PUBLIC DATA — BRAKING VIBRATION ("WARPED ROTORS") Published September 2026. License: CC BY 4.0 (credit "PulsCar analysis of NHTSA public data, September 2026"). ------------------------------------------------------------------------------- 1. SOURCE FILES ------------------------------------------------------------------------------- NHTSA public ODI flat files, downloaded 17 September 2026 from https://static.nhtsa.gov/odi/ffdd/ (field definitions: cmpl/CMPL.txt, rcl/RCL.txt, ftsb/FLAT_TSBS.txt), loaded into a local SQLite database. complaints 777,040 rows, received 1 January 2015 - 8 September 2026 recalls one row per make/model/year per campaign (campaigns counted DISTINCT) tsbs 187,827 distinct manufacturer communications Owner complaints are self-reported, unverified, and have no denominator. Nothing here is a failure rate. No "most reliable" or "least reliable" claim can be made from these counts, and none is made. ------------------------------------------------------------------------------- 2. HOW THE COMPLAINT SET WAS BUILT — AND WHY THE OBVIOUS QUERY FAILED ------------------------------------------------------------------------------- ROUND 1 (rejected). The obvious query is the word ROTOR in the narrative plus any brake context: 3,844 complaints. We drew 30 at random (Python random.seed(2026)) and read every one. Only about half were complaints about a brake disc: in 9 of the 30 the rotor was named only as a part that had been replaced along the way while the real fault turned out to be a brake booster, an ABS module or a master cylinder; in 5 more the rotor appeared in a list of repairs or in a dealer upsell; one was a motorcycle. A bare "rotor" keyword over-counts badly, so we did not publish any number from it. ROUND 2 (rejected). Narrowing to complaints where a vibration word sits within 70 characters of a braking action gave 2,172. Reading 30 of those showed the set was full of steering "death wobble" on heavy trucks, cruise-control events and transmission shudder that merely mentioned braking. ROUND 3 (published). The set used in the article is: a) NHTSA's own component tag contains SERVICE BRAKES b) AND the narrative contains one of VIBRAT, PULSAT, SHUDDER, JUDDER, SHAK, SHIMM, SHUTTER, WOBBL c) AND that word occurs within 70 characters of a braking action: BRAKING, BRAKE PEDAL, APPLY/APPLYING/APPLIED THE BRAKE, PRESSING/DEPRESSING THE BRAKE, WHEN I BRAKE, HIT THE BRAKE, COMING TO A STOP, SLOWING DOWN, DECELERAT, or the words ROTOR / BRAKE DISC d) AND the narrative names brake hardware: ROTOR, BRAKE DISC, BRAKE PAD, PADS, CALIPER, WARP, RESURFAC, MACHIN, LATHE e) AND NOT an anti-lock-only narrative: dropped if the text names ABS, ANTI-LOCK, BOOSTER, MASTER CYLINDER, VSA, VSC, STABILITRAK, ACTUATOR or BRAKE FLUID and never names a rotor, a brake disc, warping, resurfacing or a lathe f) AND the phrase DEATH WOBBLE does not appear (47 complaints dropped) g) AND the model is not a motorcycle (pattern ^(F|R|K|S|G)\d{3,4}$ or HARLEY, GOLD WING, CAN-AM, KAWASAKI, YAMAHA, SUZUKI, DUCATI, TRIUMPH, INDIAN, VICTORY, MOTORCYCLE, SCOOTER) Result: 556 complaints, received 1 January 2015 - 8 September 2026. SAMPLE READING OF THE PUBLISHED SET. We drew 30 at random (random.seed(7)) and read every one. 22 of 30 were genuine reports of a vibration felt while braking. The 8 that were not were all the same family: pedal pulsation produced by the anti-lock system activating, by a failing brake booster, or by a hybrid's regenerative braking — plus one whole-car vibration that was not brake-specific. Observed false-positive rate: 8/30 = 27%. The independent reading of 110 complaints described in section 4 gives a consistent figure: 27 of 110 (25%) turned out not to be a disc problem or not a brake problem at all. WHAT THAT MEANS FOR EVERY COUNT BELOW: roughly a quarter of the 556 are not brake-disc reports. Percentages computed on the 556 are therefore indicative of the vocabulary and the experience of people who complain about braking vibration, not measurements of a defect rate. ------------------------------------------------------------------------------- 3. WHAT THE 556 SAY ------------------------------------------------------------------------------- name a rotor or a brake disc 517 93.0% use the word WARPED 189 34.0% use the phrase THICKNESS VARIATION 1 0.2% use the word RUNOUT 4 0.7% name resurfacing, machining or a lathe 115 20.7% say the problem returned after a repair 220 39.6% say the shop could not find or duplicate it 17 3.1% are told normal wear / not covered 70 12.6% name a downhill grade, a mountain or towing 68 12.2% cite a technical service bulletin 16 2.9% towed 4 0.7% (all-complaint baseline 5.1%) crash reported 7 1.3% (baseline 4.4%) fire reported 1 0.2% (baseline 1.8%) Failure mileage: 331 of the 556 report a usable figure (0 and >= 400,000 dropped). Median 32,000 miles; 25th percentile 16,000; 75th percentile 65,000. 46.8% under 30,000 miles and 13.6% under 10,000. For comparison, the same mileage statistic computed on all 3,844 complaints that merely mention a rotor is a median of 50,000 miles: braking vibration is reported markedly earlier than rotor complaints in general. By make: brake-judder-complaints-by-make.csv (makes with >= 5 complaints in the set). By model: brake-judder-top-models.csv (models with >= 4). Each carries the share using "warped", the share naming resurfacing, the share saying the symptom returned, and the median reported failure mileage. These are complaint counts, not failure rates. Honda has the most complaints in the set (123, 22%) and the Odyssey the most of any model (74, median reported mileage 29,006). That reflects how many of those vehicles are on the road, how long the model has been sold, how visible the problem is to a driver, and whether owners have read about it — not how the vehicle ranks against any other. ------------------------------------------------------------------------------- 4. 110 COMPLAINTS READ IN FULL — WHAT IT TURNED OUT TO BE ------------------------------------------------------------------------------- From the 556 we drew 110 at random (random.seed(2026)) and read every one, assigning a single primary outcome from the narrative: the rotor was replaced, or is named as the fault 36 32.7% the rotor was resurfaced, machined, turned, cut or reground 20 18.2% a caliper, slide pin, hose or wear sensor is named as the cause 7 6.4% brake pads are named as the diagnosis 4 3.6% no diagnosis: the shop found nothing, could not duplicate it, or called it normal 16 14.5% not the disc: anti-lock module, booster, master cylinder, suspension, steering or another system 17 15.5% off target: a motorcycle, or a complaint about something else that happens to contain the words 10 9.1% Derived figures used in the article: - 56 of 110 (51%) ended at the disc itself, replaced or machined. - Of the 20 that were machined, 14 say in the same complaint that the symptom came back, in some cases within a few thousand miles. That is the number behind the article's statement that machining bought time rather than a cure for most of the owners we read. It is a small sample and is labelled as one. - 17 of 110 (15%) were a different system entirely — the honest measure of how often "your rotors are shot" is the wrong answer. - 16 of 110 (15%) never got a diagnosis at all. - Off-target rate as read: (17 + 10) / 110 = 25%. Published as brake-judder-outcomes-sample110.csv: complaint id, make, model, model year, reported failure mileage, assigned category, and whether the narrative says the symptom returned after machining. No free text, no VIN, no dealer name, no city. ------------------------------------------------------------------------------- 5. MANUFACTURER COMMUNICATIONS (438) ------------------------------------------------------------------------------- From 187,827 distinct manufacturer communications: summary contains BRAK and one of JUDDER, PULSAT, VIBRAT, SHUDDER, SHIMM, with the vibration word within 80 characters of a braking action or of the words ROTOR / BRAKE DISC. Result: 438 distinct bulletins. name a rotor or a disc 147 33.6% name thickness variation, DTV, RTV or runout 17 3.9% use the word WARPED 1 0.2% prescribe resurfacing, machining, refinishing or grinding 56 12.8% name corrosion or rust 40 9.1% name a software update or calibration 33 7.5% name lug-nut torque 8 1.8% By make (top): Acura 68, Honda 32, Land Rover 32, Toyota 31, Chevrolet 26, Mazda 26, Bentley 20, Jaguar 17, Subaru 17, Audi 15, Dodge 15, Mercedes-Benz 14, BMW 12, Volkswagen 12. SAMPLE READING: we read 30 of the 438 at random (random.seed(11)). Roughly 6 of 30 describe a brake NOISE — squeal, groan, click — rather than a vibration you feel, which is the main impurity in this count; one was a mixed list of unrelated topics. We report the count as "bulletins that describe a braking vibration or noise condition" wherever precision matters, and the 1-in-438 "warped" figure is unaffected by that impurity in any direction that would help the opposite case. The central contrast in the article — owners write "warped" 189 times in 556 complaints, manufacturers write it once in 438 bulletins — is a vocabulary finding about two populations of documents. It is not a physical measurement of how many discs are bent. Published as brake-vibration-documents.csv together with the recalls in section 6. ------------------------------------------------------------------------------- 6. RECALLS ------------------------------------------------------------------------------- Campaigns whose defect description names a brake rotor: 32, covering 63,010 vehicles or units in total. The largest is 21V-390 (Ford, May 2021, 17,616 F-350 through F-600 trucks with front wheel hub extenders). The two largest after it are aftermarket parts: 12E-057 (Neotek, December 2012, 10,626 DURALAST replacement rotors sold through AutoZone) and 24E-001 (Spektrum Brakes, January 2024, 699 Dynamic Friction R1 rotors). 14V-240 (GM, May 2014, 8,208 Buick LaCrosse and Chevrolet Malibu) is a factory assembly error: rear rotors fitted to the front. Searching every defect description in the file for JUDDER, BRAKE VIBRAT or PULSATION returns two campaigns, and both are about fuel pulsation dampers (18V-107 and 18V-432, Lexus). No campaign in this file was opened because a vehicle's brakes shake. Caveat: "potentially affected" is a population, not a count of failures; the same vehicle can appear in more than one campaign; equipment (E) campaigns and vehicle (V) campaigns can describe overlapping populations. ------------------------------------------------------------------------------- 7. PRIMARY DOCUMENTS OPENED ------------------------------------------------------------------------------- Downloaded from static.nhtsa.gov and read in full before being cited: - BMW SI B34 01 14, "Vibration When Braking: Diagnosing and Repairing", March 2015, all models. MC-10147671-9999.pdf - MINI SI M34 02 15, "Vibration when Braking: Diagnosing and Repairing", May 2017, all models with non-drilled discs. MC-10149267-9999.pdf - Lexus L-SB-0151-17, "Brake Vibration and Lug Nut Torque", 8 September 2017, 2015-2018 Lexus models. MC-10119408-9999.pdf - Chrysler 05-005-14 Rev. A, "Steering Wheel And/Or Seat Vibration - Front Brake Rotors", 19 December 2014, 2014-2015 Town & Country / Caravan / Ram C/V. MC-10221084-9999.pdf Recall texts 12E-057, 24E-001 and 14V-240 were read in the recall file itself. ------------------------------------------------------------------------------- 8. WHAT WE DID NOT DO ------------------------------------------------------------------------------- - No normalization by vehicles on the road. No reliability ranking. - No claim about what share of braking vibration is thickness variation versus bent metal. The complaint file records what owners and service advisers SAID, not what a micrometer measured. Every number here is a vocabulary or an outcome count. - No price re-verification. The repair price ranges in the article are prior published estimates and are labelled as such; consumer estimator sites block automated access and we do not work around that. ------------------------------------------------------------------------------ CASE SENSITIVITY — SECTION ADDED 19 SEPTEMBER 2026 ------------------------------------------------------------------------------ NHTSA stores complaint narratives in ALL CAPS for records filed through roughly 2020 and in mixed case for records filed from 2021 on. A case-SENSITIVE search of the narrative field loses about 92% of the complaints filed after 2020, so this is the single largest way an analysis of this database can go wrong while still looking finished. Every query behind this file is case-insensitive. SQLite's LIKE and its FTS5 full-text index both ignore case by default, and no query in this work uses GLOB or any case-sensitive comparison. This was not taken on trust. On 19 September 2026 every published folder was re-checked against the same database, comparing the figures here with what a case-sensitive query would have returned. The result: the complaint set published here (3,844) is LARGER than the absolute ceiling a case-sensitive search of the same terms could have returned (2,482 narratives). A case-sensitive query cannot produce a bigger set than itself, so the bug is ruled out for this folder by arithmetic, not by assurance.