Electrical Difficult Verified Guide

Diagnosing Dead ECT, CMP and Wideband O2 Signals After a K-Series Conversion Harness Swap

3-5 hours 5 views kricha87 via K20A.org (adapted)

Introduction

A K-series engine swap that cranks, starts and datalogs can still leave individual sensor channels blank, which makes the car impossible to tune. This guide walks through isolating why the coolant temperature (ECT), cam position (CMP) and wideband air/fuel (O2) values fail to appear in the KPro software after a conversion harness is fitted. It was written around a 12.5:1 build using a K24A1 block with a K20A2 head installed in a 1998 DC2 chassis, running a USDM PRB ECU from an RSX Type-S with Hondata KPro, and a European CR-V K24 engine harness. Those specifics matter because a EUDM harness does not share the same wire colours or ECU-connector pin positions as the USDM ECU it is being mated to, and because a counterfeit or internally damaged ECU can produce the exact same missing-signal symptoms as a wiring fault. The method below separates the three likely culprits in turn: a fake or failed ECU, a harness that routes sensor wires through the wrong path (for example the C101 bulkhead connector), and genuinely dead sensors. The payoff is a clean way to prove whether the swap wiring is actually wrong or whether the ECU itself is at fault before you keep chasing the harness. Adapted from a community writeup by kricha87 on K20A.org (https://www.k20a.org/threads/strange-problem-with-conversion-harness-o2-and-other-sensors.102259/), including corrections from the follow-up discussion.

Reference

F20C specs → K20A specs → K24A specs → R18A4 specs → D17A9 specs → D16W7 specs → K20Z4 specs → N22A2 specs → LEA-MF6 specs → LEA specs → N16A1 specs → P10A2 specs → LEB-H5 specs → L15B7 (5PA) specs → L15B7 Gen 2 specs → R20Z3 specs → R20Z2 specs → R20Z1 specs → R20A9 specs → R18Z9 specs → R18Z4 specs → R18A2 specs → ET2 specs → ET1 specs → EV1 specs → EW1 specs → EW3 specs → EW4 specs → D15B2 specs → C27A4 specs → F22B5 specs → J37A4 specs → J32A3 specs → J32A2 specs → J32A1 specs → G25A4 specs → G25A1 specs → F23A7 specs → B21A1 specs → B20A5 specs → B20A specs → J30AC specs → K20C6 specs → L15BG specs → L15CA specs → D17A6 specs → D17A1 specs → D17A2 specs → K20C2 specs → B20Z2 specs → B20Z1 specs → B20B4 specs → K24A3 specs → K20Z2 specs → D13B1 specs → ZC1 specs → D16A1-Blacktop specs → D16A1-Browntop specs → B18A1 specs → J35Z6 specs → JNC1 specs → C32B specs → C30A specs → D15Z1 specs → D15B8 specs → D15B7 specs → F22C1 specs → F22B2 specs → F23A5 specs → F23A4 specs → F23A1 specs → J30A1 specs → K24A8 specs → K24A4 specs → J30A5 specs → J35Z2 specs → K24W1 specs → J35Y2 specs → J35Y1 specs → L15BE specs → J35Z3 specs → J30A4 specs → B16A3 specs → K20C8 specs → K20C4 specs → K20C1 specs → L15B7 specs → K20Z3 specs → R20A2 specs → R20A1 specs → R20A3 specs → R18A1 specs → R18Z1 specs → K23A1 specs → K24Z7 specs → K24Z3 specs → K24Z2 specs → K24A2 specs → K24A1 specs → K20A3 specs → D16A1 specs → B18B1 specs → B18C5 specs → B17A1 specs → H22A4 specs → B16A2 specs → H23A specs → F22B1 specs → K20Z1 specs → K20A2 specs → D16Z6 specs → B18C1 specs → B18C2 specs → B18C7 specs → EA specs → EB specs → EC specs → ED specs → EE specs → EF specs → EL specs → EM specs → EN specs → EP specs → ER specs → ES specs → ET specs → K20A1 specs → K20A4 specs → K20A6 specs → K20A7 specs → K20A9 specs → K20B specs → K20Z5 specs → K24V5 specs → K24V7 specs → K24W specs → K24W4 specs → K24W7 specs → K24W9 specs → K24Y1 specs → K24Y2 specs → K24Z1 specs → K24Z4 specs → K24Z5 specs → K24Z6 specs → LEA2 specs → PPA Cylinder Head → PPL Cylinder Head → PRB Cylinder Head → R40 Cylinder Head → RNA Cylinder Head → PLE Cylinder Head → PRB-A00 Cylinder Head → PRB Fuel injector → PRB Throttle body → RNA ECU → RNE ECU → PRB ECU → PRB-A10 Intake Manifold → PRB-A00 Intake Manifold → PLE → PRB → RNA → PRB → PRB →

Tools Required

Digital multimeter (voltmeter/ohmmeter)
Terminal release/pick tool for connector re-pinning
Laptop running Hondata KManager (KPro) with datalogging
USDM ECU pinout and conversion harness diagrams (RSX/K20 PRB) plus the Hondata CTR wideband O2 wiring reference
A known-good running K-series car for ECU substitution testing
Known-good spare CMP and ECT sensors for swap testing

Parts Required

Genuine Hondata KPro on a PRB ECU (verify it is not counterfeit)
Correct plug-and-play / conversion harness matched to the USDM PRB ECU
Denso wideband air/fuel (O2) sensor
Replacement ECT (coolant temperature) sensor, if testing proves the original is dead

Safety Warnings

⚠ Do the 5 V reference and sensor voltage checks with the correct diagram for your USDM ECU; probing the wrong C101 pin (such as the unused RSX multiplexor feed) will give misleading readings.
⚠ Mixing a EUDM donor harness with a USDM ECU without re-pinning to the USDM layout can leave sensors miswired even when the plugs physically mate.

Step-by-Step Instructions

1 Rule out a counterfeit or damaged ECU first

Before touching the harness, confirm the ECU is a genuine Hondata unit. Counterfeit KPro boards are common and produce erratic, hard-to-explain sensor behaviour that mimics a wiring fault. Confirm the ECU is a legitimate PRB unit with real Hondata firmware, and keep it in mind as a live suspect throughout: a swap can be wired perfectly and still show blank channels if the ECU's internal reference supply is bad. Do not assume a running, datalogging car proves the ECU is healthy.

2 Reconcile the EUDM harness against the USDM ECU pinout

The engine harness here comes from a European K24 CR-V, while the ECU is a USDM RSX PRB. On a EUDM harness the wire colours and the pin positions at the ECU connectors can differ from the USDM layout the ECU expects. Lay the harness diagrams next to the USDM ECU pinout and move any mismatched conductors so every sensor lands on the pin the PRB ECU actually reads. Treat 'the pinouts should be the same' as an assumption to verify, not a fact.

3 Correct the TPS connector wire order

The throttle position sensor read nothing until its three connector wires were reversed. If TPS is dead, re-pin the connector so the wire order runs C-B-A instead of A-B-C. Confirm TPS then reports correctly in the software before moving on, since a fixed TPS confirms your re-pinning method works on the other sensors too.

4 Trace ECT and CMP routing through the C101 bulkhead connector

On this harness the ECT and CMP wiring passes through the C101 connector rather than running straight to the ECU, which is what a swapper coming from a direct-to-ECU wiring expectation finds confusing. The ECT feed present at C101 is a temperature signal destined for the RSX multiplexor; that pin is not used on a swap car and can be left alone. CMP does not signal through C101, but it does draw its power feed there. Verify the C101 pins carry what they should for the USDM ECU rather than the donor car's multiplexor.

5 Test the ECT sensor circuit

Measure the ECT circuit at the sensor. The signal conductor is red/white and lands at ECU pin B8. A reading of only about 0.01 V means the sensor is seeing effectively no reference feed or is dead. Note that this harness carries two red/white wires at the C101 male connector, so confirm you are probing the ECU signal path and not the unused multiplexor feed. A stuck-low ECT will also throw a low-voltage fault, so read the stored codes here as well.

6 Verify the CMP sensor power group

The cam position sensor needs three things: power, ground, and a single signal wire back to the ECU. CMP shares a common power group with several other sensors, so if the CMP power feed at C101 is not hot or is on the wrong pin, the whole group loses power and none of them report. Substitute a known-good CMP sensor to rule the sensor itself out; if a proven-good sensor still shows no cam angle, the fault is upstream in the power group or the signal path, not the sensor.

7 Check the wideband O2 wiring and drop the ECT-dependency theory

Wire the Denso wideband to the Hondata CTR O2 wiring reference, then confirm power and ground on each conductor at the sensor. A common finding is that the A1 pin is not grounded, which alone will keep the wideband from reporting. Do not chase the idea that the O2 will only wake up once ECT reaches a coolant threshold: the O2 heater circuit is purely electrical and independent of coolant temperature. It is VTEC engagement, not the O2 heater, that depends on ECT, so fix the wideband's own power and ground rather than waiting on the ECT reading.

8 Isolate ECU versus harness by testing the ECU in a known-good car

When the sensors and harness all check out but signals are still missing, prove the ECU. Install the suspect KPro into a separate, known-good running K-series car (for example an EP3 with a K20A2) and read the codes there. In this case the same ECU threw P0117 (ECT sensor low voltage) and P0112 (IAT sensor low voltage) in the healthy car, which points the fault at the ECU itself rather than the swap wiring. A likely internal cause is a failed 5 V reference/voltage regulator on the ECU board, which starves the temperature and reference-fed sensors while leaving crank/start function intact.

Pro Tips

💡 A car that starts, runs and datalogs is not proof of a healthy ECU; a bad internal 5 V reference can leave the engine running while specific sensor channels stay blank.
💡 Removing the extra harness relay to test made no difference here, so do not fixate on a single relay as the cause of multiple dead sensors.
💡 When several sensors share a power group, one wrong or unpowered feed pin can take all of them offline at once.
💡 Swap in a known-good sensor before condemning wiring, and swap the ECU into a known-good car before condemning the swap harness.

Was this guide helpful?

Related Guides