Research Reports
← All reports

2026-09-23 · audited · confidence medium · verified 2026-09-23 · stable · 27 sources · kia-rio-2004-a5d-na-power-by-fabrication-tier.md

  • automotive
  • kia-rio
  • a5d
  • naturally-aspirated
  • header
  • cams
  • ecu
  • mazda-b-series

Kia Rio 2004 (DC) A5D — Maximum Naturally-Aspirated Power, Grouped by Fabrication Tier

Date: 2026-09-23 | Sources: 27 + three-round cross-model audit | Overall confidence: Medium Supersedes the 2026-09-09 report. Round-3 synthesis: 11-SYNTHESIS-R3.md. Falsifiable predictions P1–P12: 03-PREDICTIONS.md.

Executive Summary

The A5D is a Mazda hybrid (Z5-DE block, B5-DE head) with 9.5:1 compression, hydraulic lifters and a Siemens speed-density ECU. Built as a street engine on 95 RON, its defensible naturally-aspirated ceiling is 118–125 crank hp at 6,800–7,500 rpm, reached only with a flow-benched head, cams, upgraded springs and standalone management. Bolt-ons alone land at 103–109 hp. Two independent deep-research passes agreed on every one of those brackets; neither found grounds to revive the 135–155 hp figures rejected in earlier rounds.

The programme is organised by the fabrication work each step demands, because that is what actually gates a solo builder: F0 spanner-only restoration and measurement, F1 garage fabrication, F2 a welder, F3 a machine shop with the head off, F4 engine management and engine-out work. Within that frame the round-3 audit changed five things:

  1. The Mazda parts bin is a head and valvetrain bin, not an engine ecosystem. Valves are documented to interchange with the Kia B6ZE (27.80 / 24.90 mm heads, 5.50 mm stems). Cams, intake manifold, exhaust flange, flywheel and rotating parts are Kia-specific part numbers with no documented Mazda drop-in. Design the header from the A5D head outward.
  2. The stock ECU is identified and the standalone path is known-feasible. Siemens ECM, 60-2 flywheel trigger with an OEM Hall sensor in the bellhousing, with a published MegaSquirt configuration. Speeduino and rusEFI decode 60-2 natively.
  3. The 6,500 rpm lifter ceiling is a stock operating ceiling, not a hard limit. On the B6/BP analogue, fresh springs push usable RPM to 7,000–7,300; a serious spring package to ~7,500.
  4. Two items leave the power budget. A 10 % underdrive crank pulley recovers tenths of a horsepower, not 3–5. A lightweight flywheel makes zero steady-state power.
  5. Header budget is 4–7 hp, not the vendor's 8, with 1.375–1.5 in OD primaries, larger than the 32 mm previously specified. Both reviewers mis-plugged the primary-length formula; the arithmetic is reproduced here.

Codebase Context

This report is the third revision of the A5D power programme. It replaces kia-rio-2004-a5d-na-horsepower-mods-2026-09-09, whose intake analysis and tier list were corrected in rounds 1–2 and whose remaining tiers are re-cut here. It feeds LOG-018 (Phase 3 pulley claim withdrawn, Phase 2b Mazda-header alternative downgraded) and constrains LOG-007 (header built first, 2.0" system off its collector). The sound-build case for ITBs in LOG-032 is unaffected; this report scores ITBs on power only.

Findings

1. Engine baseline — verified, do not re-research

Item Value Class
Code, displacement A5D, 1493.5 cc, DOHC 16V inline-4, transverse OEM
Bore × stroke 75.5 × 83.4 mm OEM
Factory output 97 hp @ 5500 rpm, 133 Nm @ 4500 rpm OEM
Compression ratio 9.5:1 (mymotorlist A5D) Spec site
Lineage Mazda Z5-DE block + B5-DE head (same source) Spec site
Valvetrain Hydraulic lash adjusters; Kia-specific cams 0K30E12420A / 0K30E12440A (Kia catalogue) OEM
Valves Intake 27.80 mm head / 5.50 stem / 96.40 long; exhaust 24.90 / 5.50 / 95.60 (valve set listing) Vendor, verified
ECU Siemens ECM, OBD-II ISO 9141-2, sequential injection, waste spark, Hall CMP, magnetic CKP (A5D engine-management training document); families 5WY1308F / 5WY1309C / 5WY5716A (ECU listing) OEM / vendor
Trigger 60-2 flywheel wheel + OEM Hall sensor in bellhousing; MegaSquirt example trigger angle 67°, A5D "identical setup wise" to A6D (MS1/Extra Ignition Hardware Manual) Project manual, verified
Air metering Speed-density, MAP with integral IAT, no MAF Operator
Airbox inlet ID 55 mm (measured); throttle mouth OD under 63 mm, bore not yet measured Operator
Timing belt Interference engine; failures under 50,000 km, valve damage on failure Spec site
Cranking compression spec 1275 kPa / 184 psi, ≤ 98 kPa spread Kia service data
Mean piston speed 18.07 m/s @ 6500, 20.85 m/s @ 7500 Arithmetic

Still unknown, and now measurement items rather than research items: stock cam card (duration, lift, exhaust-valve opening), stock injector flow, throttle bore, factory rev limiter. Both reviewers searched; one fabricated values with a false source attribution and those were rejected. The Z5-DE (160 mL/min) and B5-ZE (192 mL/min) injector figures in circulation are Mazda data, not A5D data (Club Protegé Z5/B5 thread).

2. The Mazda parts bin — what actually transfers

Part Status Donor Evidence Note
Valves Documented Kia B6ZE (1995–97 Sephia 1.6) Vendor listing, verified Same aftermarket set covers A5D and B6ZE; also sold by GoDPT
Trigger / standalone Documented — MegaSquirt manual, verified 60-2 + Hall; Speeduino and rusEFI decode it
Cams Plausible only Mazda B5/B6 DOHC Kia part numbers are A5D-specific No fitment record found; regrind A5D cores
Valve springs Plausible Mazda B6/BP hydraulic-lifter spring kits (Kelford catalogue) Same HLA architecture Measure installed height, OD/ID, coil bind before buying
Intake manifold, throttle body Plausible B5/B6 Kia-specific numbers; throttle body listing quotes 65 mm outside diameter, not bore No direct-fit listing
Exhaust manifold / header flange Not documented B5/B6 Kia/Mazda hybrid build needed new intake and exhaust manifolds Strongest negative finding
Injectors Electrically compatible family B5/B6/Z5 top-feed Forum A5D flow unknown; log duty before buying
Flywheel, pistons, rods Plausible with machining Z5/B6/BP Forum Not bolt-in; measure

Consequence. Head and valvetrain work is where the Mazda ecosystem helps (valves, seat practice, hydraulic-lifter cam profiles, spring packages). Everything bolted to the outside of the head is Kia-specific and gets fabricated, not bought.

3. Programme by fabrication tier

All gains are crank horsepower, inference anchored on Mazda B-series data, not A5D dyno measurements. Ranges overlap; they do not add.

F0 — spanner only. Restoration and measurement. ~0 hp direct.

Rank Item Purpose
1 Timing belt Interference engine, documented early failures. Gates everything
2 Compression test vs 184 psi If a cylinder is down, this is the whole project
3 Injector duty at WOT (ELM327 logger) ≥ 85 % = solve fuelling before adding air
4 IAT log, highway vs crawl < 5 °C spread = the cold-air build is worth under 1 hp
5 Cam degreeing (degree wheel + dial indicator, both lobes, record EVO) No public A5D cam card exists. Sets header length and cam choice
6 Throttle bore with calipers; WOT MAP depression log Gates duct sizing and the manifold decision
7 Plugs, coil/leads, fuel pressure, O2 sensor Recovers whatever is missing; do not count it as gain
8 Drop-in filter ~0–1 hp
— Throttle body coolant bypass Done

F1 — home-garage fabrication. 0–3 hp, temperature-dependent.

Rank Item Gain Band Gate
1 63 mm reinforced cold-air duct into the retained airbox; resonator and corrugated inlet removed 0–2 Upper half IAT spread ≥ 5 °C
2 Radius airbox entrance and exit 0–0.5 Upper half —
3 Heat shielding at the intake/header interface 0–1 Hot-soak transitions After F2 header
4 Piggyback ECU wiring 0 direct; 2–4 when used to calibrate Where the stock map is wrong Duty and AFR data

F2 — welder / exhaust fabricator. The first substantial airflow gain.

Rank Item Gain Band Note
1 4-2-1 header, 304 stainless, 1.375–1.5 in OD primaries 4–7 4,000–6,500 Spec in §4. Built before the exhaust; the 2.0" run comes off its collector
2 Port-match the flange without enlarging the head port 0.5–1.5 Mid/upper Head-side template
3 2.0" system + flanged cat 0–1 Upper Diameter settled; sound feature retained
4 Custom single-throttle intake manifold 2–5 4,500–6,500 New item. Only if WOT MAP depression ≥ 3 kPa (prediction P12)

F3 — machine shop, head off. Where the engine becomes a built motor.

Rank Item Gain Band Note
1 Three/five-angle seats, throat and short-side radius 3–6 4,500–7,000 Buy by measured flow, not by menu
2 Cams — regrind of A5D cores to a hydraulic-lifter street profile 4–8 (overlaps with 1) 5,500–7,200 Needs springs and calibration. Piper's B6/BP hydraulic grind (266°/256°, 8.89 mm lift) shows the product class exists (Piper BP285HR); A5D fitment of Mazda grinds is undocumented
3 Valve back-cut and margin work 1–3 Upper Flow-validate
4 Springs and retainers 0 direct Raises ceiling to 7,000–7,300 Prerequisite for cams
5 Bigger valves 0–3 alone Upper Flow bench mandatory
— Compression via pistons, ≤ 10.5:1 on 95 RON ~2 hp per point Whole band Frozen until the supercharger question is decided. Never by skimming

F4 — engine management, engine-out, donor head. Where the ceiling moves.

Rank Item Gain Note
1 Standalone ECU + dyno tune (Speeduino / rusEFI on the 60-2 + Hall trigger; Haltech/Link if budget allows) 2–5 over a well-calibrated stock ECU Prerequisite for cams. Scope the trigger with an oscilloscope before building a harness
2 Rev-ceiling raise (springs, HLA strategy, oiling, limiter) 0 direct; enables 5–10 if head and cams flow 949 Racing's BP guidance: lifters short-lived above the mid-7,000s, stock springs float high-7,000s (949 Racing)
3 ITBs 0–5 for power at this ceiling Sound case is separate (LOG-032). The 185 whp NA BP build that used 45 mm ITBs also had cams, a shaved head, Supertech valvetrain and a race header (miata.net)
4 Donor B6/BP head 5–10 only with flow data Needs new intake and exhaust manifolds; timing and oil/coolant passages
5 Forged pistons / rods Compression-dependent; rods make 0 hp Engine-out
6 Lightweight flywheel 0 hp Response only; at clutch time

Removed from the power programme

  • 10 % underdrive crank pulley "+3–5 hp". Both reviewers independently: total accessory drag at WOT with the A/C clutch disengaged is under ~2 hp, so a 10 % speed cut recovers tenths. Keep as an accessory-load experiment only. Damper-preservation warning in LOG-018 still stands.
  • Lightweight flywheel as a power item. Rotational inertia, not horsepower.
  • "Source a Mazda B5/B6 header." Not documented; see §2.
  • Big-bore throttle body until the head and cams are done.

4. Header specification — from data, with the arithmetic shown

Bell's method (Four-Stroke Performance Tuning): primary length P = 850 × ED / RPM − 3 inches, where ED = 180 + EVO (exhaust valve opening in degrees before bottom dead centre); primary inside diameter ID = 2.1 × √(cc / ((P + 3) × 25)) inches, with cc the cylinder volume (373.4 cc). Reproduced in the audit workspace at _meta/audits/a5d-mods-audit-2026-09-09/measure/header-primary-r3.txt.

EVO (BBDC) Tuned 5,500 rpm Tuned 6,000 rpm Tuned 6,500 rpm Primary ID across that range
46° 31.9 in / 81 cm 29.0 in / 74 cm 26.6 in / 67 cm 1.37–1.49 in (35–38 mm)
55° 33.3 in / 85 cm 30.3 in / 77 cm 27.7 in / 70 cm 1.35–1.46 in (34–37 mm)

The independent area rule A = (CID/N) × RPM / 88,200 gives 1.68 in² at 6,500 rpm, a 1.46 in diameter (header calculator). Three methods converge on ~1.4–1.5 in inside diameter, so 1.5 in OD tube (1.375 in OD if the fabricator argues for velocity). The 32 mm (1.25 in) OD in earlier logs is too small.

Both reviewers got the length wrong. One entered 360 − 55 = 305 as the exhaust duration instead of 180 + 55 = 235, producing 37–41 in primaries, about 10 in too long. The other used a formula with no published basis. Lengths above are the corrected values and remain provisional until the A5D exhaust cam is degreed (F0 item 5, prediction P8).

4-2-1 layout. Primaries and secondaries sum to the tuned length; secondary ID about 1.15–1.2 × primary; pair cylinders 1-4 and 2-3 (firing order 1-3-4-2, confirm on the engine); collector 2.0 in ID into the existing system; cat mount retained.

The fabricator must measure before cutting pipe: exhaust-port cross-section at the head face (not the gasket outline), port centre spacing and stud pattern, available route to the pairing point, clearance to steering rack, A/C compressor and driveshaft, cat flange position.

5. Valvetrain ceiling and cams

Round 2 said hydraulic lifters cap useful RPM at about 6,500. That was too absolute. On the closest documented analogue with the same lifter architecture (Mazda B6/BP), 949 Racing's build guidance puts lifter life problems above the mid-7,000s, fresh stock springs floating in the high-7,000s, and recommends shim-under-bucket lifters for sustained running above 7,000–7,300. Kelford and Piper sell B6/BP cam profiles explicitly for hydraulic lifters.

Working ladder, adopted:

Valvetrain state Usable ceiling
Stock springs, stock HLAs ~6,500 rpm street ceiling
Fresh or upgraded springs, HLA behaviour verified ~7,000–7,300
Serious spring package + matched hydraulic cam + validated lifter control ~7,500

Where the remaining head power lives: seat/throat/short-side work ~3–6 hp; cams ~4–8 hp. These overlap. The cam only pays when the head moves the extra air, which is why flow-bench validation precedes the cam order.

Prediction P10: with fresh stock-spec springs the engine reaches 7,000 rpm without float. If float appears below 6,700, round 2 was right.

6. Calibration path

Option Cost Unlocks Later ITBs Later supercharger Verdict
Stock Siemens ECU reflash Lowest, if a definition exists Fuel/spark tables, limiter Poor Poor Only while hardware stays near stock; no PH reflash service was confirmed for this ECU
Piggyback (Unichip-class, ~PHP 30k locally) Low–moderate Fuel and some ignition intervention Moderate Moderate Transitional; speed-density scaling gets awkward as hardware moves
Speeduino / rusEFI on the 60-2 + Hall trigger Moderate Full fuel, spark, trigger, tables, limiter Good Good Best value. Feasibility is documented via the MegaSquirt example
Haltech / Link universal Highest Professional standalone, knock control Excellent Excellent Best capability, worst economics for a 1.5 L

Calibration is a multiplier, not a last step: bolt-ons on an untuned stock map deliver less than half their potential (prediction P7). Injector headroom is unknown; the WOT duty log decides whether Mazda-family top-feed injectors are needed.

7. Compression, ITBs, pulley, flywheel — the attacks

  • Compression. Both reviewers independently: ~10.5:1 is the conservative street target on 95 RON in a 35 °C climate; ~11.0 possible with excellent calibration and knock control; above 11.5 is uncomfortable on pump fuel. Roughly 2 hp per point, but the gain depends on whether timing was knock-limited. Take it through pistons, never a head skim: a piston set can be swapped back if a supercharger arrives, a skimmed head cannot. Still frozen pending that decision.
  • ITBs for power. 0–5 hp over a well-designed single-throttle manifold while the ceiling is ~6,500 rpm, and they force Alpha-N tuning. Spend order for power: head → cams → ECU → compression → ITBs. The sound argument stands on its own in LOG-032.
  • Underdrive pulley. Not credible as a power number. See §3 removals.
  • Flywheel. Zero steady-state power. Response only.

8. Build versus donor swap

No defensible Philippine-peso breakpoint could be sourced. One reviewer offered "PHP 65,000" with no listing behind it; that number is rejected. What is documented: the LTO treats an engine change as a modification requiring inspection (LTO FDM Vol. 2); engine substitution must be registered within the statutory period (RA 10883); PNP-HPG clearance for change of engine requires the donor engine's invoice or CSR, an affidavit, the source vehicle's OR/CR and a physical inspection (PNP-HPG Clearance Form 1).

Mechanically the crossover comes earlier than bare-engine prices suggest, because an A5D build keeps mounts, gearbox interface, axles, wiring, cooling and fuel system, while a B6-ZE or BP-ZE swap replaces all of them plus the paperwork. The measurement that settles it: price a complete running donor package locally, not a long block.

9. Measurements that gate money

Recommendation Measure How Result that flips it
Build the 63 mm duct Highway-vs-crawl IAT ELM327 log, same route < 5 °C spread: worth under 1 hp, money to the header
Buy any airflow part Injector duty at WOT ELM327 at 6,000+ rpm ≥ 85 %: stop, fuelling first
Header length Stock exhaust cam EVO Degree wheel + dial indicator Outside 45–58° BBDC: recompute §4
Custom manifold WOT MAP depression Logger at 6,000+ rpm < 2 kPa: leave the manifold alone
Throttle body Bore Calipers at the blade ≥ 58 mm: no gain from a bigger TB
Cams Actual cam card As above, both lobes Materially different from B5 surrogate: redesign header and cam choice
Springs Float onset Fresh springs, sweep to 7,000 Float below 6,700: round-2 ceiling stands
Head work Flow through lift curve Same bench, before/after Little mid/high-lift gain: stop cutting
Underdrive pulley Accessory-load delta Back-to-back, same accessory state < 1 hp: delete (P9)
Standalone ECU AFR, duty, timing on stock ECU after F2/F3 Wideband + logger Stock ECU cannot hold lambda/timing: standalone moves forward
Donor engine Complete-package price Quote a running package with mounts, gearbox, wiring Below cumulative A5D machine/fabrication spend: economics flip

10. Full-build estimate

Build state Crank hp Peak RPM Condition
F0–F2 103–109 5,800–6,300 Healthy engine, cold air, tuned 4-2-1, calibrated fuel and spark
F0–F3 110–118 6,300–7,000 Flowed head, cams, springs, calibration
F0–F4 118–125 6,800–7,500 Serious head/cam/ECU package, raised usable RPM

Two reviewers, independently: 103–109 / 105–108; 110–118 / 114–118; 118–125 / 122–125. The external anchor is Mazda's own B5-ZE, 125 PS at 7,000 rpm from 1,498 cc (Mazda B engine). Caveat: the B5-ZE is 78 × 78.4 mm; the A5D carries the B5-DE head on a 75.5 mm Z5 bore, so valve shrouding and port-to-bore ratio differ. The analogue supports the order of the ceiling, not a claim that the A5D copies B5-ZE output. A 185 whp NA BP 1.8 exists (miata.net), but it is a different displacement and a race package.

Risks & Caveats

  • Every gain figure is inference anchored on Mazda B-series data, not an A5D dyno sheet. The predictions file records thresholds that make each one wrong.
  • One reviewer's return was rejected on facts. It attributed a cam card, injector flow, piston dimensions, valve sizes and a trigger pattern to sources that do not contain them. Its directional agreements are retained only where the other reviewer reached the same conclusion independently.
  • Header lengths are provisional until the exhaust cam is degreed. Diameters are not.
  • No Philippine pricing could be confirmed for cams, springs, standalone ECUs or a donor package. Local quotes required.
  • Hydraulic-lifter ceilings are extrapolated from B6/BP evidence, not tested on an A5D.
  • Compression and cams alter standing behaviour; a cold-air duct does not. Never bundle validation of the two kinds.
  • Legality: cat retained on a flange; engine swaps require LTO and HPG paperwork as cited.

Recommendation

F0 health and measurements → F1 IAT and restriction test → F2 header (plus manifold only if MAP depression says so) → calibration → F3 flow-benched head → cams and springs → only then the F4 standalone / ITB / donor-head decision.

  1. Timing belt, compression, injector duty, IAT log, cam degreeing, throttle bore. All before spending on parts.
  2. Build the 63 mm duct only if the IAT spread is ≥ 5 °C.
  3. Fabricate the 4-2-1 header at 1.375–1.5 in OD primaries, lengths from §4 after the cam card exists, then the 2.0" system off its collector.
  4. Calibrate. Piggyback if staying near stock; Speeduino or rusEFI on the 60-2 trigger if F3 is coming.
  5. Head work bought by flow numbers, then cams on regrind A5D cores with springs to match.
  6. Decide the supercharger question before any compression change; if compression, pistons not skim.
  7. ITBs, donor head and rotating-assembly work only after all of the above, and only with the measurements in §9 in hand.

Sources

  1. Kia A5D engine specifications — mymotorlist — https://mymotorlist.com/engines/kia/a5d/ — 9.5:1 CR, Z5-DE block + B5-DE head, hydraulic lifters, belt failure note. Fetched 2026-09-23: carries no cam, valve, piston or injector data.
  2. Evergreen intake/exhaust valve set, 01–02 Kia Rio A5D / 95–97 Sephia B6ZE — Spark Automotive — https://sparkautomotive.com/products/evergreen-sa-9761182754-intake-exhaust-valves-fit-01-02-kia-rio-1-5l-a5d-b5-dohc-16v — 27.80/24.90 mm heads, 5.50 mm stems, 96.40/95.60 mm. Verified 2026-09-23.
  3. Same valve set — GoDPT — https://www.godpt.com/products/intake-exhaust-valves-fit-01-02-kia-rio-1-5l-a5d-b5-dohc-16v — second listing, same application.
  4. MS1/Extra Ignition Hardware Manual — msextra.com — https://www.msextra.com/doc/ms1extra/MS_Extra_Ignition_Hardware_Manual.htm — Kia Rio A6D/A5D example: 60-2 flywheel trigger, OEM Hall sensor, trigger angle 67°, waste spark. Verified 2026-09-23.
  5. Kia Rio A5D engine-management training document — manuals.plus — https://manuals.plus/m/65299127b0d6ddfa67fd96b39cd2c2d4057779cf17bf50d964b5320e294be13e — Siemens ECM, ISO 9141-2, sequential injection, Hall CMP, magnetic CKP.
  6. A5D ECU 5WY5716A / 39110-2X567 listing — B-Parts — https://www.b-parts.com/auto-parts/electrical-and-electronic/engine-control-unit-ecu-kia-rio-i-hatchback-dc-15-16v-5wy5716a-391102x567-2000-2001-2002-2003-2004-2005-2006-6854656 — ECU family.
  7. A5D ECU 5WY1308F / 5WY1309C listing — B-Parts — https://www.b-parts.com/auto-parts/electrical-and-electronic/engine-control-unit-ecu-kia-rio-i-hatchback-dc-15-16v-5wy1308f-5wy1309c-simk42-k33e-18881-2000-2001-2002-2003-2004-2005-2006-211707 — ECU family.
  8. Kia genuine parts catalogue, A5D valve system — partsouq — https://partsouq.com/en/catalog/genuine/unit?c=Kia&cid=761988206&q=KNADC223246345434 — A5D-specific cams 0K30E12420A / 0K30E12440A, hydraulic adjusters.
  9. Kia genuine parts catalogue, A5D intake manifold — partsouq — https://partsouq.com/en/catalog/genuine/unit?c=Kia&cid=1916505584&q=KNADC223246345434 — manifold 283102X065, Kia-specific.
  10. Z5-DE into B5-ZE thread — Club Protegé — https://www.clubprotege.com/forum/forum/2nd-generation-mazda-protege-familia-1995-1998/engine-drivetrain-bha-chassis/48067-z5-de-into-b5-ze — Z5 160 / B5-ZE 192 mL/min injectors, B5-ZE exhaust timing ~55° BBDC (Mazda data, not A5D).
  11. Kia/Mazda hybrid build thread — Club Protegé — https://www.clubprotege.com/forum/forum/forum-news/new-members/43460-noob-here-and-with-a-kia-mazda-based/page2 — B6 head on a Kia block required new intake and exhaust manifolds; B6ZE pistons and BP rods with machining.
  12. Planning your BP engine build — 949 Racing — https://949racing.com/supermiata/tech-info/planning-your-bp-engine-build/ — hydraulic lifter life above the mid-7,000s, spring float, shim-under-bucket recommendation.
  13. Piper BP285HR hydraulic cam, Mazda B6 — Piper Cams — https://www.pipercams.co.uk/product/mazmx16bp285hr — 266°/256°, 8.89 mm lift, hydraulic-lifter profile.
  14. Piper BP285HR, Mazda BP 1.8 — Piper Cams — https://www.pipercams.co.uk/product/mazmx18bp285hr — same profile class for the 1.8.
  15. Kelford Mazda B6/BP cam and spring catalogue — manuals.plus — https://manuals.plus/m/19ee65ede920ad8251601737f33db11a92faad5f3a8bac70c96d04ba034b480d — hydraulic-lifter cam profiles and spring package.
  16. Maruha Motors racing cams, Miata B6/BP — https://www.maruhamotors.co.jp/miata/parts/racingcam.php — further B-series cam options.
  17. 185 whp naturally-aspirated BP build — miata.net — https://forum.miata.net/vb/showthread.php?t=593947 — shaved BP4W head, 264° cams, Supertech valvetrain, 45 mm ITBs, race header, 7,800 rpm.
  18. Mazda B engine — Wikipedia — https://en.wikipedia.org/wiki/Mazda_B_engine — B5-ZE 125 PS @ 7,000 rpm, 78 × 78.4 mm, JDM only. Verified 2026-09-23.
  19. Ford Laser (B5-ZE) catalogue — carsensor.net — https://www.carsensor.net/catalog/ford/laser_hatchback/F002/M001G004/ — JDM Laser B5-ZE output.
  20. Header primary tube length calculator — calculator.academy — https://calculator.academy/header-primary-tube-length-calculator/ — P = 850·ED/RPM − 3 and A = (CID/N)·RPM/88,200.
  21. rusEFI — https://rusefi.com/ — Hall/VR support, OEM trigger patterns, sequential injection.
  22. Kia Rio DC throttle body listing — Autodoc — https://www.autodoc.parts/car-parts/throttle-body-12439/kia/rio/rio-saloon-dc/55966-1-5 — 65 mm nominal outside diameter; not bore.
  23. Kia Rio 2001–2005 service/repair manual listing — eManualOnline — https://www.emanualonline.com/cars/kia/rio/kia-rio-2001-2005-service-repair-workshop-manual-emo-175327.html — service-data availability.
  24. Kia Rio 2001 workshop manual (Spanish) — Scribd — https://www.scribd.com/document/460374028/KIA-Manual-de-taller-Kia-Rio-2001-pdf — valve-seat and compression service data.
  25. LTO Field Data Manual Vol. 2, 2nd ed. — https://www.lto.gov.ph/wp-content/uploads/2023/10/FDM-vol.-2-2nd-Edition.pdf — engine change treated as a modification requiring inspection.
  26. Republic Act 10883 — https://ldr.senate.gov.ph/legislative%2Bissuances/Republic%20Act%20No.%2010883 — engine substitution must be registered.
  27. PNP-HPG Clearance Form 1, change of engine — https://hpg.pnp.gov.ph/wp-content/uploads/2024/07/Clearance-Form-1.pdf — donor invoice/CSR, affidavit, OR/CR, inspection.

Methodology

  • Mode: cross-model audit round 3 (deep research). Brief: _meta/audits/a5d-mods-audit-2026-09-09/10-BRIEF-R3-DEEP-RESEARCH.md. Two reviewers (GPT deep research, Gemini Deep Research), asked independently, returns stored verbatim.
  • Local verification 2026-09-23: mymotorlist A5D page fetched (cam/valve/piston/injector data absent — one reviewer's [OEM] attribution rejected); valve-set listing fetched (dimensions confirmed); MegaSquirt manual located on msextra.com (60-2 + Hall confirmed); B5-ZE output confirmed on Wikipedia. Header arithmetic reproduced in measure/header-primary-r3.txt; both reviewers' primary lengths corrected.
  • Groq was not consulted this round: its numbers were rejected in two prior audits and web research was the purpose.
  • Flags: no A5D dyno data exists anywhere consulted; all gains are inference. Torquecars.com excluded at operator request. Kia Forums content remains paywalled; nothing new from it this round. Insufficient data: stock cam card, injector flow, throttle bore, rev limiter, PH pricing for cams/springs/ECU/donor package.