PulsCar analysis of NHTSA public data, September 2026 Crank but no start: what 867 owner complaints and five recall campaigns say 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/ (field 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. Campaigns counted DISTINCT on campaign no. NHTSA investigations flat file FLAT_INV. https://static.nhtsa.gov/odi/ffdd/inv/INV.txt -------------------------------------------------------------------------- WHICH COMPLAINTS COUNT (the crank-no-start set: 867) -------------------------------------------------------------------------- Step 1. FTS5 match on crank OR cranks OR cranking OR cranked OR the phrases "turns over", "turned over", "turn over", "turning over" -> 7,831. Step 2. Keep a complaint only if its narrative contains an explicit crank-BUT-no-start construction. Writing CR for (crank|cranks|cranked|cranking|turns over|turned over|turning over|turn over): POSITIVE, within one clause (no sentence-ending punctuation crossed): CR [<=80 chars] [but|,|and|yet|however|though]? [<=30 chars] (will not|won't|would not|wouldn't|does not|doesn't|did not|didn't| never|not|no|fails to|failed to) (seem to|actually|ever)? (start|fire|catch|turn on|run|ignite) NEGATIVE, checked in the 45 characters BEFORE the match -- if present the match is discarded, because it is the OPPOSITE complaint (a car that will not crank at all): (will not|won't|would not|wouldn't|does not|doesn't|did not|didn't| never|no|not) (even )? CR Result: 868 complaints. Step 3. Drop powersports makes (motorcycles, scooters, ATVs). One complaint dropped (a Husqvarna motorcycle). FINAL SET: 867. SAMPLE READING. Thirty random matches were read. Twenty-eight described a true crank-no-start. Two were no-crank reports whose phrasing defeated the negation filter ("would not crank or start"; "would not hardly turn over and would not start"). False-positive rate 2/30, about 7%. The article states this. WHY THE SET IS SMALL. This is a deliberately narrow query. It requires the owner to have written the crank-but-no-start construction explicitly; it does not try to infer the symptom from a description of a tow or a dead car. A looser query would be larger and dirtier. 867 is a small file next to our other symptom sets, and every percentage on the page should be read with that in mind. NO NORMALIZATION. Raw counts of unverified owner reports with no denominator (NHTSA publishes no vehicles-on-the-road figure). These are NOT failure rates and cannot rank reliability. 284 of the 867 narratives (32.8%) mention a recall, and recall publicity is itself a reason to file, which inflates the makes with the largest campaigns. -------------------------------------------------------------------------- HEADLINE FIGURES -------------------------------------------------------------------------- Fields TOWED_B / CRASH / FIRE = 'Y': crank-no-start set 1.27% / 0.23% / 0.35% all 777,040 5.11% / 4.37% / 1.82% injuries 2, deaths 0. THE TOW FIELD DISAGREES WITH THE NARRATIVES, AND THE ARTICLE SAYS SO. The structured towed field is set in 1.27% of these complaints, which is below the all-complaint baseline and implausible for a vehicle that cannot move. The word "tow" appears in 28.3% of the narratives. The field is optional and often skipped. We report both and tell the reader to trust the text. Mileage (field MILES; drop missing, zero, negative, >= 400,000): 544 usable. Median 85,000. Quartiles 56,000 and 114,326. <= 36,000 miles 12.3%; <= 60,000 28.1%; <= 100,000 65.4%. Narrative subsets (regex, case-insensitive, whole set of 867): died / stalled / shut down WHILE DRIVING before the no-start 265 30.57% stranded, roadside, highway, intersection, parking lot 246 28.37% the word tow 245 28.26% mentions a recall 284 32.76% intermittent / sometimes / comes and goes / random 172 19.84% second, third or fourth visit for the same fault 110 12.69% says there was no warning or no indication 110 12.69% shop could not duplicate / no fault found 12 1.38% hot-only pattern 60 6.92% cold-only pattern 30 3.46% -------------------------------------------------------------------------- WHICH PART OR SYSTEM THE NARRATIVE NAMES -------------------------------------------------------------------------- Independent regexes over the full 867; categories overlap, so shares do not sum to 100%. Published as crank-no-start-what-owners-named.csv. Fuel pump, pump module or fuel module 292 33.68% Relay, fuse or the integrated power module 239 27.57% Wiring, harness or connector 43 4.96% Battery replaced or called good 42 4.84% Spark plugs, coils or ignition (incl. ignition switch) 37 4.27% Crankshaft or camshaft position sensor (or P033x/P034x/ P0016/P0017) 36 4.15% Immobilizer, key, fob, security system, Passlock 36 4.15% Engine control module or software 28 3.23% Timing belt or timing chain 22 2.54% Fuel filter 6 0.69% THIS IS A MAP OF VOCABULARY, NOT OF CAUSES. Owners write down the part a shop named out loud. Nobody files a federal complaint because a gallon of gas fixed it, which is why the cheapest cause on the page is invisible here. -------------------------------------------------------------------------- OWNER-REPORTED DOLLAR FIGURES (45 figures, 41 complaints) -------------------------------------------------------------------------- Published as crank-no-start-owner-reported-costs.csv. A dollar amount counts when ALL of the following hold: - it matches \$\s?([0-9][0-9,]{2,6})(\.\d{1,2})? in the narrative; - its value is >= $80 and <= $6,000; - a 110-character window either side names a fuel pump, pump assembly, pump module, fuel module, relay, fuel pump relay, TIPM or "totally integrated power"; - that window contains none of: owe/owed, per month, monthly, down payment, purchase price, bought the car/vehicle, loan, financ*, worth, trade-in, lease, deductible, insurance, new engine, engine replac*, remanufactured engine, catalytic, airbag, transmission, camshaft, cylinder head, rental, total expense, spent over, already on this car; - duplicate values inside one narrative count once. FIRST PASS AND WHY IT WAS TIGHTENED. A looser query (wider window, no competing-term list, ceiling of $25,000) returned 86 figures. Reading 20 random matches found about six misfits -- a $16,000 engine, a $14,500 remanufactured engine, $12,070 of catalytic converters, an airbag-light bill, and two running totals for a whole ownership. Tightening produced the 45 figures published here; reading 20 of those found two doubtful ones, about 10%. RESULT. Median $1,100; quartiles $600 and $1,400; range $140 to $2,200; 30 of 45 at or above $1,000. Cheapest complete repair described: a genuine fuel pump relay at $18.88 with a $176 total bill including labor. HOW TO READ IT. 45 figures is a small sample, and it is drawn from people angry enough to file with the federal government, which skews toward large and disputed bills. It also mixes relays, modules and pumps together, because the narratives do. The article says all of this in the text. -------------------------------------------------------------------------- RECALLS AND INVESTIGATIONS OPENED IN FULL -------------------------------------------------------------------------- DP14-004, "Totally Integrated Power Module Failure," opened 25 September 2014, closed 24 July 2015 (petition denied). Summary as published in FLAT_INV. 104 of our 867 narratives (12.0%) mention the TIPM by name. Recall campaign records read through https://api.nhtsa.gov/recalls/ campaignNumber on 18 September 2026: 14V-530 (189,206), 15V-115 (339,525), 19V-813 (528,594), 15E-028 (132,536 Bosch fuel delivery modules), 10V-384 (1,128,659). Recall search used: campaigns whose defect text names a fuel pump relay, an integrated power module or a TIPM (4 distinct campaigns after excluding airbag campaigns that use the word "knock"/"relay" incidentally), and campaigns whose consequence text names a vehicle that will not start or fails to start (15 distinct campaigns; the five used above are the ones whose defect is genuinely a no-start mechanism on passenger vehicles). -------------------------------------------------------------------------- PART PRICES RE-CHECKED THIS PASS -------------------------------------------------------------------------- Read 18 September 2026 at the dealer-parts catalogs ToyotaPartsDeal and HondaPartsNow, listed price against the catalog's own MSRP: Toyota 90919-05060 Crank Position Sensor $105.72 (MSRP $149.02) Toyota 90987-02022 Circuit Opening Relay $79.52 (MSRP $112.09) Honda 39794-SDA-A05 Relay Assembly, Power $22.62 (MSRP $32.22) Honda 12290-5A2-A01 Spark Plug (NGK) $16.44 (MSRP $23.22) These are the ONLY prices on the page re-priced in this pass. Every labor range keeps its earlier date: RepairPal, AutoZone, O'Reilly, NAPA, Advance Auto Parts and PartsGeek all returned HTTP 403 on 18 September 2026, and RockAuto served a CAPTCHA, which we did not attempt to work around. -------------------------------------------------------------------------- FILES IN THIS FOLDER -------------------------------------------------------------------------- crank-no-start-by-make.csv make, complaints, miles reported, median miles, share at or under 60,000 miles, share whose narrative says the vehicle was towed, share that stalled or died while driving before the no-start, share describing an intermittent fault. Makes with fewer than 8 complaints are omitted. crank-no-start-top-models.csv the 45 most-reported make/model pairs with the same columns plus the top model year and its count. crank-no-start-what-owners-named.csv the ten part/system categories above. crank-no-start-owner-reported-costs.csv 45 figures with ODINO, make, model, model year and amount. No VIN, no city, no dealer name, no free text. CONTACT: support@pulscar.io