Knock sensors - Too quiet on decel/light throttle, throwing codes

[marin]

LS2 in LS1 car (Holden Adventra), P59 PCM. Standard LS1 knock sensors in standard LS2 position on sides of block. Runs fine except for this issue.

Keeps throwing the 0327 and 0332 on decel or light throttle when travelling over 90km/h. Have put scope on KS's when not driving and reved it, and found that the sensors go extremely quiet before picking up again at idle.

My well experienced tuner has tried everything he knows, bit hasn't come across this before. The knock sensors are horrible to change out due to the AWD system, so trying to avoid that (have to almost pull the engine).

Any ideas (or more info i need to provide)? I don't have access to logs.

Thanks.

[Alvin]

Some early Gen 3 cars are an absolute B about getting the knock sensors working. I assume the harness is in good shape? We always replace the jumper harness when setup OEM. I assume you have some sort of conversion harness.

The knock sensors are the right part number for the PCM? it sounds like it by your post. IMHO and contrary to what other people say.. GEN 3 and GEN 4 knock sensors aren't compatible or swapable. I've never seen GEN 4 knock sensors work with a Gen 3 PCM for example.. at least reliably.

Are the knock sensors torqued to spec? Are they reasonably new?

We've had a few be fixed by swapping in another PCM.

[SiriusC1024]

It's a circuit problem. Those sensors are 1 wire. I'd check that all the grounds are installed with good continuity per factory configuration. People tend to miss them during a swap.

[marin]

Some early Gen 3 cars are an absolute B about getting the knock sensors working. I assume the harness is in good shape? We always replace the jumper harness when setup OEM. I assume you have some sort of conversion harness.

The knock sensors are the right part number for the PCM? it sounds like it by your post. IMHO and contrary to what other people say.. GEN 3 and GEN 4 knock sensors aren't compatible or swapable. I've never seen GEN 4 knock sensors work with a Gen 3 PCM for example.. at least reliably.

Are the knock sensors torqued to spec? Are they reasonably new?

We've had a few be fixed by swapping in another PCM.

The knock sensors were brand new in GM bags, BUT I bought them off a bloke who said he didn't use them, so I have no idea about their origin, there is every possibility they are knock offs. They DO ohm out perfectly at 101K ohms each, even through the VCM extension loom (I KNOW the loom is legit, I bought it over the counter at VCM, I'm in Melbourne, Australia). On a scope, they register knock when you tap the block. We probed the PCM connector with the scope when revving it in the shed, and it threw the codes when they went quiet on dropping back of revs. I have brand new Delphi KSs sitting here from a reputable store (but again, who knows these days) if it comes to that, but they are horrible to get to and change.

I was very careful to torque them to spec. I can't remember what that was, but i looked it up at the time, 15Nm I think it was.

I moved the PCM from the strut tower into the cabin in the footwell as we thought maybe heat on the PCM, but it didn't help the issue. I purchased another PCM and we swapped it in, no change to the issue.

[marin]

It's a circuit problem. Those sensors are 1 wire. I'd check that all the grounds are installed with good continuity per factory configuration. People tend to miss them during a swap.

That's a great call, and yes, we did find yesterday that I missed the earth on the back of the block, but we bolted that back on with no change. I also ran a secondary, very large compared to factory, earth wire from the front head position to the battery negative many weeks ago. Unfortunately, no change for either of those earths being connected.

[marin]

The only other thing I have read that I was wondering about was whether the knock sensors/PCM takes a reference point when it meets certain criteria at start up..... maybe that reference is too loud? Causing the quiet to be out of tolerance, throwing the code? Dunno, I'm not at all experienced with this stuff, just been reading too many opinions on the Internet lol

[SiriusC1024]

The code is for a low voltage reference. Have the DMM set to VDC instead of ohms. For example, battery ground to ground post will be under 50mV for a good ground. A weak ground will show low ohms, but measuring voltage with everything energized will give the real answer. Check continuity with the engine running. Rev the engine while testing to replicate the failure conditions.

My far-out idea is that the alternator is causing excess ripple. When I was talking about grounds I was specifically thinking of the one that goes to the alternator bracket. Use the scope to check power/ground of various points, such as across the battery, alternator power and ground, and the ECU, and monitor for disturbances.

[marin]

The code is for a low voltage reference. Have the DMM set to VDC instead of ohms. For example, battery ground to ground post will be under 50mV for a good ground. A weak ground will show low ohms, but measuring voltage with everything energized will give the real answer. Check continuity with the engine running. Rev the engine while testing to replicate the failure conditions.

My far-out idea is that the alternator is causing excess ripple. When I was talking about grounds I was specifically thinking of the one that goes to the alternator bracket. Use the scope to check power/ground of various points, such as across the battery, alternator power and ground, and the ECU, and monitor for disturbances.

Thanks, I will try all this next. It is a brand new factory GM alternator, but who knows until I test more.