rb20 knock sensor problem

[Skym]

One of the symptoms of leaking intake manifold is missing at idle (lean missfire). If engine knocks, ECU can protect engine via running engine rich (knock maps), which could make engine miss at high rpm's. Spraying head to intake manifold joint can make engine note change, which means there's a intake airleak.

If knock sensor is faulty, ECU should engage safety revlimiter at around 5000rpm.

[jayplanedin]

Okay, well my plenum is missing one bolt, however I did spray it with carb cleaner and nothing changed, I will check again tho. I figured it could have been a leak, however it almost seems electrical since it seems to happen suddenly out of nowhere. Does the ecu throw code 34 only when the knock sensor is putting out the wrong voltage(borked sensor)? Or when it detects knocks too? And by 5k rpm limited, do you mean that it will not rev past no matter what? Or is that just the safety point where it begins retarding the timing?

[Skym]

Error code on ECU pops up due to voltage of sensor outside certain voltage specs or can't see the sensor because of faulty or incorrectly wired (engine conversion) wiring. Engine manual should have the voltages.

ECU has a separate knock board mounted above main board inside ECU that converts frequency from knock sensor into knock count. When knock count rises above a certain level (this level can be adjusted when ECU is retuned), ECU switches to knock maps. But from what I understand it only does this above certain rpm (4400rpm) due to the way the knock count level is setup. PowerFC is similar with knock count, as it was based off stock ECU. The knock detection on stock ECU is really accurate, as it picks up knock at the same time as a good aftermarket knock detection device. But knock sensor can get fooled by excessive noise from missing coilpack, something loose on engine, etc and report that engine is knocking when it's not.

5000rpm safety rev limiter is a fuelcut and engine shouldn't rev above 5000rpm. It should engage if knock sensor is faulty.

Sometimes engine can run rich due to using a bigger fuelpump than factory, so ECU needs a retune.

With it being random, possibly electrical, that's why I suspect wiring to fuelpump. Can wire fuelpump so it runs at 100% all of the time. Not ideal due to heating fuel, reducing lifespan of fuelpump, but solves the problem.

Knock sensor, TPS, MAF, 02 sensor are the only sensors I know of that can have an effect on amount of fuel injected.

[jayplanedin]

Very informative post, thank you. So with the 5k rpm rev limiter being a fuel cut and since my engine revs past that obviously the knock sensor also throws a code when having knocks. I think I may look into rewiring the fuel pump. It asnt noticeable anymore but with my old loud fuel pump you could hear something that almost sounded like little splashes coming from the fuel pump. Like tsst.. Tsst-tsst.... And so on. These quick noises were paralleled with the sound of the engine's misfiring. Not sure if that could be the cause of the problem or a result of the problem. How is the fuel pump told to run at different loads? A program in the ecu?

[jayplanedin]

Checked fuel pressure under lean misfire. An even 41. Vacuum was 14 in/hg then after revving shortly the vacuum dropped to 10 in/hg. But still a clean 41 psi.

[jayplanedin]

Thinking of getting injectors rebuilt next.

[Skym]

So fuel pressure, knock sensor are eliminated as causes.

I agree, injectors can leak and cause missing problems. Seen some have this problem.

Change injector loom at same time. Coilpack, injector loom are known to cause missing problems when they go hard as a ironing board, get brittle.

[jayplanedin]

i was doing some reading through the r32 service manual, and found that code 34 is only presented when there is a short circuit or break in the voltage of the knock sensor, does not state anything about actual knocking setting off the code, i plan on rewiring the knock half harness and seeing if that would help. i am aware that the injectors can leak, causing a rich mixture, however is it possible for the solenoid within them to have the wrong dwell and flow less fuel?

how do i go about replacing the injector loom? would i just cut and rewire from the harness back? or is it better(or possible) to somehow source a new injector loom?

[Skym]

Knock doesn't set off a error code. Knock is recorded as knock count.

Don't know with injector. Always thought injectors leaked when they fail. Although on one RB20DET, this person had one cylinder going extremely lean (cylinder one), took out the rings. He was set on saying that it was the ECU causing it. But could have been the injector loom???

From what I understand, from what I was told, the injector drivers on stock RB20DET ECU go sequential? at lower rpm's, then batch? at higher rpm's, so can't see how ECU could cause any problems with one cylinder, as batch affects 3x cylinders at a time.

But if one cylinder is going lean, should notice sparkplug tip getting melted or doesn't have soot buildup on that cylinder when sparkplugs are pulled. Reading sparkplugs can tell you alot about what's going on inside cylinders.

I always thought the injector loom plugged into main loom, just behind engine??? New injector loom, as have new plugs.

[jayplanedin]

Checked for boost leaks today, had a **** load, fixed all of them, still same problem... Still want to rebuild injectors and see if anything changes.