Knock Sensors

[bobbysmith]

I have a 02 silverado 4.8 with 299,000 miles. I had a knock sensor code for a while and the intake gaskets were starting to leak so I ordered the Kit from A/c Delco with new harness and sensors. I finally got time off work to do the install then cleared the codes. Ran fine for a day or so then knock sensor codes returned. Truck runs great other than that doesn't use much oil for the miles that are on it. When doing a tune up last weekend I decided to check and see if maybe i didn't plug the new harness in or something. It was good but I noticed that the wire on the 2 connectors are opposite each other. The connector can only plug in 1 way. Is it made this way or is it built wrong from the factory? Would this cause the code to return? I thought about trying to pull the connector apart but didn't want to ruin it and have to pull the intake again. Like an idiot I threw the old harness away thinking everything was good then.

[rjgti]

did you torque ks to the in-lb setting

[rjgti]

another thought, if you ran with bad ks for a while then ign timing was retarded and it could be carbon build up in cyl.

[bobbysmith]

I torqued them to what it called for in the manual.

[NSFW]

hmm.

I just finished changing the knock sensors changed in my 00 Silverado this-afternoon and I found this post kinda interesting.
After getting it all buttoned up (TQ on the sensor is 15ft/lb - 180in/lb btw) Started and ran beautifully. Took it out for a rip and it threw a different code then it did before, Before it was #2 sensor showing an open circuit. This time it was the #1 sensor doing the same thing. I cleared it.. and kept on running.

Your in luck I HAVE the old wire so I could tell you which (and tell myself on the receptical plug) goes to which sensor. I'm sure the sensor itself grounds to the block thru the threads, no magic there. I put a dab of di-electric grease on the terminals before I clicked the knock sensors cables in place and siliconed the &^%$ out of the bung-plug as corrosion is what did in the back sensor. Didn't call for it, but I work with electrical regularily and the stuff seems to work.

Now I am not heart-broken that going thru the effort of changing the sensors and its a wiring problem, as the truck has 450 thousand KM on it and they were due regardless. What I am curious about is why the effing MIL light won't go out, especially when the engine works well, dosen't consume oil, and generally isn't a griefer PAST changing the sensors.

AT the very bottom of the MIL snapshot it tells me possible cause is the PCM has failed. If the PCM failed.. would nuthing work? I wonder why they even put that in there.

Sensors, connections.. Pin-out and harness running to the PCM.. Thats all I can think of allong that chain, nothing in addition. I did notice allot of carbon build-up on the intake plenum in all areas not directly being sprayed by the fuel injectors. I ruled that out as thanks to the exhaust re-circ and not timing related.

I will dig that harness out and check it again. With pictures.

[NSFW]

Allrighty. Sorted a few things out as far as connectivity is concerned but so far still scratching my head about just exactly whats up with the truck and more importantly how to bloody fix it.

I'm leaning towards Rjgti's sense with the carbon buildup inside the combustion chambers. My knock sensors have been throwing an OL code on the number 2 bank for about a month which isn't a tonne of time in OL mode. However I have no idea how long it takes carbon to build up, so thats not a very good ellimination. When I had the intake off, it was a mess of carbon inside the intake. The stuff was just flacking off. I'm aware that is due to the EGR circuit and isn't the same as carbon buildup on the inside of the combustion chamber. Still. 450 thousand KM, bound to be some buildup.

I checked the wiring, because I was sure I had an open line to the front knock sensor. Not so. I was relieved. Nothing like doing all that work.. only to rip it all up and do it again.

I got my meter running.. so I set immidiately and checked the resistance on the REAR sensor. With the clip facing to the left.. the lead on the top is the rear sensor. The lead on the bottom is the front sensor.



I have the positive lead stuffed in the connector, and the ground is held against the cyl-head to ground. 100.0 K/Ohm (also a connectivity check also, something wrong with the wiring would show NOTHING on the meter.



The bottom lead is the front sensor. Showing 100.6 K/Ohm.

I grabbed the old sensors that were in the block. I ran to see to compare if there is any difference between the installed ones over the un-installed ones. I learned something here. NOW is NOT the time to check and compare the sensors. Check the sensors when you can handle em. Note to self.. go SLOWER. In my defense, it was cold.. in the driveway...and hardore rules apply.

This first pic is the REAR sensor. I found the resistance to be HIGHER then the ones installed in the truck. I thought this odd because the sensors I checked are a physically shortest electron path, where as the check done on the truck is a much longer current path. I expected the resistances to be lower on the sensors, not higher..



and again the FRONT sensor. Again a higher resistance then whats installed in the truck.



And this last shot is the stock (whats left) of the harness that was in the truck. The short tail is the rear sensor, and the long tail is the front sensor. I used the meter to run continuity test to determine which terminal on the plug is attached to which sensor.

[muddkatt]

I have to ask, what is that galvanized tee and shut off valve set up for?

[NSFW]

It's actually a brass tee. It's a bypass loop installed between the heater hoses to a coffee-pot style coolant heater. Yea... LOL.... product of a case of beer and what started off as a joke. It works, best use of an old coffee pot. It actually has 'delay brew'. I aint kidding. Works like a hot damn.

I found another tidbit concerning the P0332 and P0237 code pair. PCM failure?? How about this. PCM GROUND pulling lots of resistance due to age. Checking the ground the PCM uses for the detection of the knock sensor signal?

You bet. Its located on the rear left side of the block. Or at-least I think based on this photo.



Don't help much cause I got no reference position. Not even any arrows. It's a dedicated ground, that runs back to the PCM as opposed to the regular ground strap. Mine is right here after looking thru my old photos. It's in a absolutely GLORIOUS spot. DID I check it..before I put the manifold back in? NO.. lol.. derp. I had no idea it was even there.



Closeup



Yeah I bet that's the problem. Right there.

[NSFW]

Update.

I checked the PCM ground and its connected. No corrosion or lack of contact present. Passed a wiggle test. Rang out the PCM pinout to the block and checked the PCM plug end in case the pin got leaned over, or got over compressed from a meter needle jammed into it. Everything looks as it should, meter varified.

Last night before I turned in I un-plugged the positive from the battery and let it sit un-plugged all night.

The new sensors were showing 100kOhm and the old sensors were showing more resistance to that value. Remembering that resistance and voltage are inversly perportonate to one another. When the resistance goes up.. the voltage goes down. Below threshold maybe because went from the Rvalue of the old sensors to new sensors without changing the computers stored value. The computer was telling me that the 'voltage was below threshold' on both circuits. That meant either an open ground sure. But if those circumstances are not present.. then what? Exactly.

The PITA about this is I have no way of looking at the raw data from the PCM. So I have no idea if the PCM is saying it has a open, a short, or ground fault.

Checklist.
#1: Sensors Torq 180in/lbs. Threads dry check
#2: Tug test connectors to knock sensors check
#3: Reistance check thru pigtail, thru knock sensor to block. 100kOhm check
#4: Connectivity back thru pigtail to PCM interconnect. check
#5: Ground from PCM to block. check
#6: Resistance from PCM interconnect thru knock sensor to block check

I severed the connection between the onboard computers and the battery, to allow the captured memory values to drop off. I think the last time I had the positive off the battery was last summer. When I replaced the battery. Since then the truck has never been without 'keep alive' power. With no way of knowing if the knock sensor value is a 'keep alive' memory item, or a firmware value. I started to think outloud on it and rule out some impossibilities.

Knock sensors generate thier own voltage, everyone knows this. The PCM interperates the value and compares that to a pre flashed base value. How does the PCM establish this base value? Well it doesn't, thats what the firmware is for. But the specific value is a LEARNED value. The PCM advances the timing till it gets a knock, then retards it from there. The PCM must have a window (threshold) its looking for at first, but as for the EXACT value.. well.. I dont think two knock sensors generate at the EXACT same mV value. My guess is the PCM reads to the thousandths place. (0.000) mV. and experiments with it untill it goes into CL mode. Then it saves the values. Even though they might register at 7.6mV in a knock situation.. to the PCM its 7.584mV and the digital meter is rounding up.

So thats what was running thru my mind with the disconnect of the positive cable from the battery. Clear the memory.

If the PCM is looking for the knock sensor mV value of the old sensors, and the sensors have been changed, then the PCM is going to give the low-threshold error. Because (my guess here not fact.. theory) its (the PCM) looking for the old sensor value, and presented with a new sensor value which won't match up with it's stored bit.

I started the truck this morning, no problem. I let it warm up, and I took it for coffee. I didnt get 1km away from home and it threw a code. HOWEVER.. this time it was a 'fuel trim lean' code bank 1 and bank 2. Which means two things.

One.. I got dirt in the fuel-rail when I seperated the cross over pipe from the rails to clear the PCV valve and syn-flex hose and get it out of the truck.. and I got an injector that now has crap it. Or they all have crap in them. Not ruined.. just never take apart more then you need to. The only way dirt got introduced to the fuel injectors is because I put it there. Derp.

Two.. My intake gascuts are leaking and sucking air. Intake maniold has a leak on one of the hoses. It was maticulously cleaned, and torqued directly to spec when it was installed. BUT.. Where's my vaccume guage.

On the bright side my knock sensors work.. Yea.. It's like celebrating recovering from a back inury by mowing the front lawn, whilst in the process ya run the machine out of oil. Hardly a win.

note* Yes, in the manual they tell you to disconnect the positive battery connector before you start any work...Naaa....

[rjgti]

damn bad luck, did you put in new gaskets. as a side note that plastic intake gets brittle with time and is easy to crack even with the correct torque procedure