PULSCAR ANALYSIS — P1326 (HYUNDAI AND KIA KNOCK SENSOR DETECTION SYSTEM) IN NHTSA'S PUBLIC FILES PulsCar analysis of NHTSA public data, October 2026. License: CC BY 4.0 Attribution: "PulsCar analysis of NHTSA public data, October 2026." Published alongside https://www.pulscar.io/blog/p1326-code ------------------------------------------------------------------------------ Prepared October 2026. Source data: NHTSA public flat files downloaded on 7 October 2026 (complaints received 1 January 2015 to 4 October 2026 — the latest received date in the files — 782,908 records; manufacturer communications; recalls since 2010; investigations) and documents downloaded one at a time from static.nhtsa.gov. 1. Complaints (1,766, all read in full, two readers) First search: every complaint whose narrative names P1326 by any spelling we searched (a space, hyphen or period after the P, a letter l or I for the 1, the suffix 00 as in P132600, or a bare 1326 after "code", "codes", "DTC", "error", "CEL", "coded" or "fault"), names the system ("KSDS", "knock sensor detection", "engine protection mode"), or — on a Hyundai, Kia or Genesis — mentions the knock sensor in owners' spellings or describes a blinking check-engine light together with limp mode or a held engine speed: 1,457 complaints. Completeness audit: before the second batch was coded, an independent completeness search of the same file for what the first search missed proposed 812 complaints; the 309 whose narratives name the system in words were added (found_by column), and 503 that cite only a recall or campaign number were not (see codebook.txt). Coding: the codebook (codebook.txt) was written, dated and hashed before any complaint was read; an addendum for the second batch, carrying the decision rules written during the first, was hashed before that batch was read. Two independent classification passes coded all 1,766 separately, each from the codebook only, without seeing the other's codes, each seeing only the codebook and its batch of complaints. One category (N not about the system on this vehicle; E engine replaced or rebuilt; D engine damage or failure not yet replaced; S knock sensor or its circuit; O another cause repaired; M software only; C checked and cleared; U no diagnosis or repair stated) and four flags (activation = P1326 or the KSDS warning on this vehicle; ksds_done = the knock-sensor software was installed at some point; covered = the maker or dealer paid the whole repair; denied = a refusal, a share demanded, or the owner paid). Agreement over all 1,766: category 1,610 (91.2%), Cohen's kappa 0.884; activation 0.918; ksds_done 0.926; covered 0.840; denied 0.917. Second batch alone (309): category kappa 0.766. Every disagreement (268) was read in full and settled under written rules R1-R20 (codebook.txt); two consistency sweeps corrected 65 agreed codes that contradicted those rules; the resolution and rule_applied columns record both. Final: not relevant (N) 337; of the 1,429 relevant: D 578, U 502, S 151, E 73, O 54, M 44, C 27. Activation 870 (E 41, D 240, S 132, O 34, M 19, C 20, U 384). ksds_done 388. Covered 62; denied 682 (E: covered 36, denied 25; D: denied 412). 485 relevant complaints print P1326 (any spelling above). Makes among the 1,429: Kia 742, Hyundai 687. Models: Sonata 279, Optima 267, Soul 181, Tucson 158, Sorento 140, Sportage 95, Santa Fe 86, Santa Fe Sport 72, Forte 48, Veloster 37. Median model year 2016 (one complaint with an unknown model year excluded). 380 complaints give a mileage from 1,000 to 400,000; median 93,738.5. By year received: 2018 5, 2019 97, 2020 56, 2021 139, 2022 223, 2023 226, 2024 195, 2025 260, 2026 (to 4 October) 228. Mention counts (by script, not coded; case-insensitive, in the relevant narratives): oil consumption, burning oil or quarts (whole word) 187; previous, second, prior or subsequent owner or bought used 60 (47 among the 682 with a refusal or payment); maintenance records or receipts 60 (46); salvage, branded or rebuilt title of the car 20 (19; the script matched 25 and 23, and five mentions of a salvage engine, a salvage value or salvaging the car were removed by reading). Elantra N and Kona N: 10 relevant complaints, 8 mentioning rain, a car wash, water or a wet car (11516164 and 11612224 do not). Figures count complaints, not vehicles. 2. Owner-reported amounts (167 figures in 141 complaints) Every dollar figure (and "dollars", "k" amounts) in the relevant complaints was extracted by script (303 candidates) and decided one by one under one rule: keep every figure the narrative ties to the diagnosis or repair of the P1326/KSDS event or the engine — a test or inspection fee, the software update, a knock sensor or its wiring, another repair named on the way, or the engine — whether paid, billed or quoted; exclude mileages and other non-price numbers, vehicle prices, trade-in and salvage values, loan balances, towing, rentals, storage and fees, repairs unrelated to the event, restatements of a figure already counted, figures inside a stated total, and totals that mix unrelated items (139 exclusions, each with a reason, in our working files). 112 figures are quotes, 30 described as paid, 20 billed, 5 not stated. Groups: diagnosis, inspection or test 33 (median $230); knock-sensor software update 2; knock sensor or its wiring 9 (median $600); another repair 38 (median $1,201.50); repair not specified or several repairs 10; engine replacement or rebuild 75 (median $8,000; 69 quotes; the lowest, $200, is an extended-warranty deductible). 3. Manufacturer documents (642 distinct files, read) First search: every manufacturer communication whose NHTSA index summary names P1326 or P132600, KSDS, knock sensor detection or engine protection mode (any make), or — for Hyundai, Kia and Genesis — the knock sensor, connecting rods, rod bearings, bearing clearance, or long-block/short-block coverage: 332 NHTSA ids. Completeness audit: 44 more ids (campaign siblings whose summaries lack the keywords, the T3G bulletin, dealer best practices, SC147 side bulletins, piston-noise documents). Recalls: the files of 13 campaigns (15V568, 17V224, 17V226, 17V578, 18V907, 20V746, 20V750, 21V259, 21V301, 21V727, 21V844, 23V877, 25V099). Investigations: EA21-003, PE19-003, PE19-004, RQ17-003, RQ17-004 and DP18-003, read from NHTSA's investigation flat file (no PDF). 649 PDFs were downloaded (one request at a time); 13 were byte-identical copies (md5); the remaining 636 plus the 6 investigation records are 642 distinct files. Near-identical files (5-word shingle Jaccard 0.6 or more) were grouped into 271 clusters; one representative of each was read in full and every other member was compared with it sentence by sentence (a diff file per member) and read for its dates, vehicles and terms. Pages without a text layer were rendered and read as images, and the six image-only recall PDFs were transcribed page by page. The op-code flow diagrams on pages 22-32 of Hyundai 26-EM-002H (Santa Fe 3.3L, TXX8, not a P1326 program) were read as images too; none mentions P1326 or the KSDS. 200 files print P1326 or P132600 in their own text (the code inside Hyundai labor op codes such as 8P1326 is not counted); 93 of the 277 cluster representatives and standalone files do. Relevance as recorded: DIRECT 278, RELATED 360, LISTED 2, NOT RELEVANT 15 (members inherit their representative's class where their own entry gives none). Every verbatim quote in the article was checked by script against the document or complaint text; quotes from image-only pages were checked against our transcriptions. Coverage status in September 2026 is our computation from the terms printed in each document, not a statement by the manufacturer. No document prints an individual car's in-service date. Limits: complaints over-represent failures, refusals and expensive repairs; owners report diagnoses as they understood them; a narrative that cites only a campaign number is not counted; searches of NHTSA's index cannot see documents whose summary names none of the search terms. Nothing here is a failure rate.