Checking Knock Sensors

[Nebone]

I'm looking to solve my knock sensor issue.

16712 P0328 Knock Sensor (KS) 1 -G61: High Input (sporadic)

According to Bentley, this sensor is located on Cyl 3 &4. To replace it I'd have to pull the Intake Manifold and all the lines connected to it.

I have a new sensor in hand and was thinking to compare the resistance between the both to verify that the one installed is actually bad (I hope!)

What does High Input really mean?

My guess is that the resistance went down in the wiring somehow and that aplifies the engine noise thinking its some type of knok.

[dougyfresh]

Do you really need to pull the intake mani? I think you can get to it with a long socket extension and just push some of the vacuum lines away. If my memory serves me correct, the socket is a 13mm and the torque value is 20Nm.


Isn't 'High Input' regarding you're getting some high knock voltages? Thus the sensor is acting very 'sensitive'.. Did you log your knock voltages?

[Nebone]

I did a run today with my brother doing the start and stop and he absolutely sucks at it; like its hard to do.

I got logs of Knock Sensor Voltage (Block 26) and Cylinder Retard (Block 20). I unfortunately did not include load or RPM because I simply forgot. More to come later.

Third Gear from 3,000 RPM to 7,000 RPM and braking

Idle and slow 800 to 4000 rpm rev Knock Voltage

You can see the knock voltages are lower for Cyl 3 and 4 for which I'm getting the code for; interminnent code.

[Nebone]

Bump

[Ewok_Fetus]

LOL! Resistance of wires doesn't just go down over time... A wire is basically a short, can't go lower in resistance than that unless you use a different material or superconductor... But if you are getting a high input to the sensor then I would say ECU... The ECU is what controls the voltage going into sensor... Or you could have a short from the supply line to ground, this will lower the voltage to the sensor giving it a lower output and also make your ECU give a high output... My clue to you is to find a pinout for your ECU and check for shorts in your wiring...

http://wiki.ross-tech.com/index.php/16712/P0328/000808

[Jaxspeeddemon]

you dont need to pull the inake manifold....it took me 40 minutes to do mine on my 99 a4 aeb. its just in a tuff spot and that bolt is at an angle so be careful and not to strip it, i think its 7ft pounds of TQ.

[dougyfresh]

Does the bentley really say the code is for cyl 3 and 4?

Start with checking the torque of the bolt that holds the sensor down. Then do as was just mentioned: check the continuity from the connector to the ECU as well as the resistance drop across that line.

[Nebone]

Resistance in wires can change by a simple short in insulation cause by hot coolant lines right near those wires and even wire oxidation.

I found what looks like a typo in the Bentley manual. The manual says that the G61 sensor is the sensor for cyl 3 and 4 and its what I have been following since the start of this episode. Look at #19



I traced the Knock Sensor cables today and the G61 which has the green plug is actually for the cyl 1 and 2. Whats this? I'm not pulling timing on cyl 1 and 2 like I am on cyl 3 and 4. I'm confused now.

Why are the voltages different and there is 4 of them where there is only 2 Knock Sensors? Looks to me like cyl 3 and 4 voltages stay very close and cyl 1 and 2 jump all over the place...but why is 3 & 4 pulling timing like that when 1 & 2 isn't?

[Ewok_Fetus]

Resistance in wires can change by a simple short in insulation cause by hot coolant lines right near those wires and even wire oxidation.

Insulation does not short... Insulation is what keeps wires from shorting. The insulation can melt and cause a short but that still doesn't change the resistance of the wire. It will just change the path of electricity. Plus, the wires used under the hood are meant for high temps which would prevent the insulation from melting... Yes, wire oxidation can cause an INCREASE in resistance but not a decrease like you were trying to say... But other than that, resistance cannot change in a wire like you are thinking...

And there are 4 different voltages because we have what is called cylinder selective knock control. The ECU knows which cylinder is firing at that moment and since the knock sensor is between the two, it calculates the voltage for that cylinder at that moment...

If you have the sensor and the sensor is supected bad, just change it... If it doesn't change THEN go onto other fixes... Most of the time it will be a failing sensor, then there are always those other problems...

[Nebone]

Now it just comes down to which sensor it is thats going bad. Front or back?

Here is a log I have from RPMs and Cylinder retard. Block 3 and 20 Log

THe main cylinder retard happens on cyl 3 and 4 wheres the code indicates G61 which is cyl 1 and 2.

[Nebone]

I replaced the G61 (Front Knock Sensor) and did a ECU reset. Right after the reset it was peachy but the next day, after lots of driving, the same timing pull returns. I'm still waiting to see if the code pops back up.

Did a log last night and you can clearly see there is a strong timing pull on all 4 cylinders.



Knock Voltages at idle have not changed.

Next thing I'm doing iss changing out the cam chain tensioner. It has been rattling at cold starts for the last 60k miles and now its been chattering at low RPMs and idle is getting louder. That can definitely trigger knock sensors and/ or retard timing.

[mike-2ptzero]

Have you done a compression test lately?

[Ewok_Fetus]

Actually, to tell you the truth, those look okay for a chipped car.... When it gets above 6 is when you start to worry and then above 8 and 10 is uh oh... I will try to get mine done today and post it... But I don't think it will look like that since I am dumping fuel like a Boeing 747 on an emergency landing! Another thing I was going to recommend was doing a quick Seafoam through the combustion chambers... You might have some carbon build up causing an increase of compression which will cause that...

[Nebone]

Have you done a compression test lately?

I'll do it this weekend with Jake and probably replace the chain tensioner and chain. Valve cover was never off and its ~129k on the motor.

Another thing, during light load like cruising or slight revving, the timing retard jumps to 6+ but not at full load.

[dougyfresh]

Have you looked at group 026? I think that's the actual knock voltage if I am reading this correctly: http://www.ross-tech.com/vag-com/m_blocks/

Do a pull with groups 010 and 020 to see what your actual timing advance is while also looking at timing correction. Just curious.

I can bring my compression tester with me to Jake's this weekend if you need it (but he probably already owns one).

Do you have the compressor for the rear chain tensioner? Makes life 20x easier, albeit I don't own one.

[Nebone]

I have everything besides a new chain, but that would require removing the exhaust cam whic is on tension by the timing belt which I don't wanna mess with now. If the chain looks beat, i'll close it up and wait for a chain then do it all at once. (I have 1 tool in the tensioner and one extra)

Post #3 above has my knock sensor voltages. Its same now with new sensor installed. I'm waiting to see someone elses knock voltage readings at idle so I can compare.

I'll rent a compression tester from AutoZone tonight and perform the check. I'll replace the plugs too, since I get this uneven idle tone (sorta has skips like slight misfires) since I installed BKR6Es gapped a .028"

Wish my exhaust wasn't so loud so I could actually hear the engine.