PulsCar analysis of NHTSA public data, September 2026 Rattling noises: 5,789 owner complaints, 2,467 manufacturer bulletins, 2 recall campaigns Licensed CC BY 4.0. Queries run 18 September 2026. -------------------------------------------------------------------------- SOURCE FILES -------------------------------------------------------------------------- NHTSA ODI complaint flat files COMPLAINTS_RECEIVED_2015-2019, _2020-2024, _2025-2026 -- https://static.nhtsa.gov/odi/ffdd/cmpl/ (dictionary CMPL.txt) 777,040 distinct vehicle complaints (ODINO, product type V) received 1 January 2015 to 8 September 2026. NHTSA recall flat file FLAT_RCL.zip. NHTSA manufacturer communication files TSBS_RECEIVED_2015-2019, _2020-2024, _2025-2026 -- https://static.nhtsa.gov/odi/ffdd/tsbs/ (dictionary TSBS.txt) Loaded into SQLite and queried with Python 3 and the sqlite3 module. -------------------------------------------------------------------------- CASE SENSITIVITY -- READ THIS BEFORE REUSING THE QUERIES -------------------------------------------------------------------------- NHTSA complaint narratives are ALL CAPITALS for complaints received through 2020 and mixed case from 2021 onward. A case-sensitive regex, or an upper-case SQLite GLOB, silently drops most of the post-2020 file while still returning thousands of rows, so the query looks like it worked. That matters a great deal here: the largest single cluster in this set, the Ford 3.5L cam phaser rattle, is overwhelmingly 2021-2026 filings. Every query below is case-insensitive on purpose: - the SQL stage uses LIKE, which is case-insensitive for ASCII in SQLite; - every Python regular expression carries re.IGNORECASE; - no GLOB, and no upper()/lower() comparison against a cased literal. Word lists are printed in lower case for readability; they match in any case. -------------------------------------------------------------------------- WHICH COMPLAINTS COUNT (the rattle set: 5,789) -------------------------------------------------------------------------- STAGE 1 -- SQL pull, case-insensitive LIKE on the narrative: descr LIKE '%rattl%' OR '%ping%noise%' OR '%pinging%' -> 9,699 candidates. STAGE 2 -- Python, per sentence. A complaint counts when at least ONE sentence contains a rattle word that survives the filters. rattle words (regex, IGNORECASE): rattl* | ping* THE PING RULE: "ping" is only counted when the same sentence also contains an engine or fuel word (engine, knock*, detonat*, octane, fuel, accelerat*, spark, timing, uphill, under load, climb). Without that rule the word matches "shipping", "stopping" and every other -pping verb, and it also matches the parking-sensor sense of ping. A stop-list of common -pping words (shipping, helping, stopping, keeping, sleeping, camping, jumping, whipping, dropping, slipping, gripping, zipping, clipping, tipping, capping, napping, wrapping, snapping, chipping, skipping, equipping) runs on the matched token itself as a second guard. Makes excluded: motorcycle, heavy-truck, bus, trailer and RV makers, and rows where make is UNKNOWN -- the same list our other analyses use (Harley-Davidson, Indian, Triumph, Ducati, Kawasaki, Yamaha, Suzuki, Polaris, Can-Am, KTM, Vespa, Piaggio, Aprilia, Moto Guzzi, Arctic Cat, Royal Enfield, Husqvarna, Zero, Mack, Hino, Kenworth, Peterbilt, Paccar, International, IC Bus, Freightliner, Western Star, Blue Bird, Thomas Built, Navistar, Prevost, Workhorse, Spartan, Autocar, Forest River, Volvo Trucks, UD Trucks, Sterling, Gillig, U-Haul, Cummins, Winnebago, Grand Design, Jayco, Thor Motor Coach, Keystone, Heartland, Coachmen, Newmar, Tiffin, Entegra, Airstream, Lippert, Carson, PJ Trailers, Peach Cargo, Dutchmen, Prime Time, Crossroads, Palomino, Shasta). RESULT: 5,789 complaints. SAMPLE READING (required before publishing any keyword count) Thirty randomly drawn matches were read in full. Twenty-nine described a rattling noise from the vehicle. One did not: a narrative about rough running where the rattle word appeared in a separate clause about engine performance. FALSE-POSITIVE RATE: 1 IN 30, ROUGHLY 3%. This is a lower rate than our other noise analyses because "rattle" is an unusually unambiguous word in this corpus -- the ambiguity lives in "ping", which the rule above handles, and the ping subset is only 96 complaints. Counts under about 30 should still be read as "a handful of owners". -------------------------------------------------------------------------- FIELDS AND RESULTS -------------------------------------------------------------------------- Outcome flags (share answering Y, of 5,789): towed 3.39% (196), crash 1.04% (60), fire 1.05% (61), 43 injuries, 0 deaths. All-complaint baseline (777,040): towed 5.11%, crash 4.37%, fire 1.82%. Mileage: NHTSA MILES field, rows with 0 < miles < 400,000 (3,332 rows, 57.6% of the set). Median 72,000; quartiles 40,000 and 105,000; 23.1% at or under 36,000 miles. All-complaint baseline median (392,935 usable rows): 70,000. WHERE THE NARRATIVE PLACES THE RATTLE (categories overlap; a complaint that places it nowhere matches none, so shares do not sum to 100) Engine or under the hood 3,015 52.08% Interior, dash, trim, seats, windows or doors 1,200 20.73% Suspension or steering 892 15.41% Transmission, converter, clutch or flex plate 639 11.04% Brakes: pads, clips, shields, calipers 412 7.12% Exhaust: heat shield, pipe, muffler, converter 367 6.34% Something loose: jack, spare, tools, hubcap 156 2.69% Turbocharger 75 1.30% PARTS THE NARRATIVE NAMES BY NAME Cam phaser 432 7.46% Timing chain, tensioner or guide 317 5.48% Engine pinging or detonation 96 1.66% Catalytic converter 84 1.45% Lifter, rocker or valvetrain 50 0.86% Heat shield 35 0.60% The heat-shield row is the one to read carefully. It is small because cheap problems that get fixed do not generate federal complaints, NOT because loose heat shields are rare on the road. CIRCUMSTANCES THE NARRATIVE DESCRIBES Names a recall, extension, campaign or bulletin 1,825 31.53% (all-complaint baseline for the same regex: 32.49%) On a cold start or in the first minutes 650 11.23% Over bumps or rough road 293 5.06% Engine was replaced, or needs replacing 266 4.59% At idle 225 3.89% Rises or changes with engine RPM or acceleration 152 2.63% Shop could not find it (strict no-fault-found) 155 2.68% (all-complaint baseline for the same regex: 2.64%) Told the noise is normal 154 2.66% THE CAM PHASER CLUSTER (432 complaints) Ford F-150 317, Ford Expedition 97, Lincoln Navigator 9, and 9 others across six makes. Median mileage 72,000 (261 rows report one); only 8.0% are at or under 36,000 miles. 109 of the 432 narratives name a Ford customer-satisfaction programme or a customer-service programme number. The Ford F-150 alone is 610 complaints -- 10.5% of the entire rattle set -- at a median of 67,562 miles, and 52.0% of its narratives name a cam phaser. Its biggest model years are 2018 (186), 2017 (121) and 2019 (112), which is the 21B10 population. MANUFACTURER BULLETINS (2,467 distinct documents) Rows whose Summary contains rattl*, case-insensitive, collapsed from one row per make/model/year to distinct NHTSA document IDs, after the heavy-truck and engine-maker exclusions. Of the 2,467, by full summary text: names body or interior parts 894 36.2% names the engine, a valve, lifter, phaser or chain 561 22.7% names suspension 250 10.1% names the exhaust 247 10.0% remedy mentions foam, felt, tape, a clip, a fastener, tightening, torque, grease or insulation 259 10.5% mentions software, reprogramming or a calibration 129 5.2% names brakes 89 3.6% Counts by make (top): Chevrolet 249, Honda 173, Jeep 165, Volkswagen 164, Mazda 142, BMW 138, Audi 134, Ford 133, Acura 129, Land Rover 128, Cadillac 83, GMC 79, Toyota 64, Subaru 60. Bulletin counts are counts of documents, not of vehicles or repairs. RECALL CAMPAIGNS (2) Campaigns whose defect, consequence or remedy text contains rattl*, case-insensitive: exactly two in eleven years, covering 182,129 vehicles by the manufacturers' own potentially-affected counts. 10V-179 (GM, 162,129 Hummer H3) -- hood louver clip-tabs fracture, "causing the hood louver rattle against the hood and become loose", and if unnoticed the louver could detach and strike a following vehicle. Remedy: adhesive, free of charge. 11V-371 (Nissan, 20,000 Altima) -- transverse link bolt not torqued to specification, which "could cause the bolt to become loose while the vehicle is being driven, creating a rattling noise and vibration"; if the bolt comes out completely, difficulty controlling direction. A rattle is almost never a federal safety defect. Both of the campaigns that exist describe a rattle as the audible early warning of a part working loose. PRIMARY DOCUMENTS OPENED FOR THE ARTICLE (PDFs fetched from static.nhtsa.gov/odi/tsbs/, HTTP 200 on 18 September 2026) Ford Customer Satisfaction Program 21B10 Supplement #3, 13 September 2021 (10201643) Ford 21B10 owner letter text, 21B10-S2, September 2021 (10201088) Stellantis 09-012-20, 17 October 2022 (10223986) -------------------------------------------------------------------------- LIMITATIONS -- READ BEFORE QUOTING ANY NUMBER -------------------------------------------------------------------------- 1. Complaints are self-reported, unverified and voluntary. They are not failure rates and cannot be converted into one: NHTSA publishes no denominator, so nobody knows how many of each model are on the road. 2. Counts favour high-volume models and models whose problem has been publicized. Nothing here ranks reliability, and no honest reading of complaint counts can. 3. THE SET IS SELECTION-BIASED TOWARDS EXPENSIVE AND UNRESOLVED RATTLES. A $4 exhaust clamp does not produce a federal complaint. The location table is a map of which rattles become disputes, not of which rattles happen. 4. Mileage, towed, crash and fire are the owner's own entries; mileage is blank or unusable in 42.4% of this set. 5. The crash share (1.04%) is far below the all-complaint baseline (4.37%) because the baseline is dominated by airbag and seat-belt complaints filed BECAUSE of a crash. 6. "Mentions software" counts the word, not the outcome. -------------------------------------------------------------------------- FILES -------------------------------------------------------------------------- rattle-where-owners-place-it.csv location category, complaints, share of the 5,789. rattle-parts-owners-name.csv part named, complaints, share of the 5,789. rattle-complaints-by-make.csv make, complaints, median miles, miles reported, percent towed, percent fire, percent naming the engine or under-hood. rattle-top-models.csv same plus percent naming a cam phaser and the top model year, for the 60 most-reported models. rattle-bulletins.csv 2,467 rows: NHTSA document ID, bulletin number, date, make, model years covered, component, flags for body / exhaust / engine / suspension / software / a pad-tape-clip-or-tighten remedy, and the first 200 characters of the summary. No personal data is published: no VIN, no city, no dealer name, no owner free text beyond short quoted snippets in the article itself.