Cayenne V6, NVH to locate origin of engine knock, click, misfire

[Rfmotors1]

Hi all current and future Pico users, I have one case involving NVH which I think is good to share.

Porsche Cayenne 3.6 V6, old but still nice car with complaint “engine vibration after morning start”. Its good to mention cold start means 28 or more degrees Celsius as this is in southeast Asia tropical climate.
There was no DTC and no warning light, but the vibration was easy to feel simply by seating in the vehicle and keep it idling after cold start. There was also noticeable clicking or knocking noise, probably come from cylinder head area, most likely from the valvetrain.

Engine knocking idle, cold.mp4

(1.51 MiB) Downloaded 142 times

By monitoring the misfire count via diagnostic tester, the cylinder 4 was having misfire counts which were slowly disappearing as the engine warmed up, together with the clicking noise which also almost disappeared. However, the idle was not perfectly smooth even at engine operation temperature. On higher RPM and under the load, there was nothing noticeable at either temperature.
So where to start? have confirmed knocking noise from cylinder head area and cylinder 4 misfire counter. Given there should be not any noise, next logical step would be to remove valve cover and visually inspect, if we do not have Picoscope.
So, lets confirm what cylinder is really misfiring, where the noise comes from and verify the in-cylinder compression pressure and valve timing and here we are.
First is to confirm what cylinder is really misfiring by simply analyse the crankshaft speed from hall sensor signal with sync by cylinder 1 ignition coil. All is easy to reach, backpin on engine ECU and ignition coil Cyl 1.
Here is the result, firing order is 1-5-3-6-2-4 so the cylinder 4 fires just before cylinder 1 fires.

Engine misfire, crank speed.mp4

(2.62 MiB) Downloaded 129 times

1 Misfires at idle, cold engine, crank speed.psdata

(27.38 MiB) Downloaded 101 times

The diagnostic tester was correctly reporting the misfires caused by cylinder 4 have been confirmed.
Next is to record the in-cylinder pressure waveform on affected cylinder 4 and any other good one, in our case cylinder 3 which does not fire close to misfiring cyl 4, avoiding double misfire.
Below is the result, with help of reference waveform function we can overlay both for easy comparison, focusing on valve timing or leaking signs.

2 Compression idle Cyl 3 vs cyl 4, dual view.psdata

(1.06 MiB) Downloaded 58 times

The cylinder 4 has higher pressure rise around second TDC 180 degrees, this is likely due to both, EX/IN valves closed or nearly closed at the same time. Could be due to leaking hydraulic tappet or worn-out cam lobe. The exhaust valve opening and Intake valve closing time is not significantly different from good known cyl 3.
Next step is to bring the Picoscope NVH which should be good to accurately locate the origin of the clicking noise. This is done by calculating the speed of sound in cylinder head material and then once we know the time difference, (delay) between two accelerometers, we are able to calculate location of the click/knock origin in reference to accelerometers. (I am using this technique to locate squeaking or knocking on body/suspension of luxury cars as Bentley where even tiny common click has to be located and where other techniques failed to locate it)
By placing the accelerometers outside of point of interest, here on front and back side of cylinder head. There are bolts where the magnets were easy to be attached. Then make artificial knock by screwdriver beyond front accelerometer. The accelerometer which is closer to my knocking has picked up the vibration first, the other accelerometer located at the end of the cylinder head picked the same signal with delay. Using the trigger, we have the knock recording where is easy to measure the time difference with Picoscope rulers.

3 knock by screwdriver, determine speed of sound.psdata

(1.02 MiB) Downloaded 59 times

Next is to measure the distance between the two accelerometers with measure tape and write both to simple calculator I have created in excel (NVH.rar excel file is at the end). This will give us the speed of sound in meters per second.
Once we know the distance between the accelerometers and speed of the sound in the material, we just need start the engine and capture the click or knock, then measure the time difference between the two accelerometers signals and identify what accelerometer captured the “knock” signal first.

4 Engine Idling, knock source identification.psdata

(11.39 MiB) Downloaded 60 times

Distance between accelerometers 51 cm
Time delay after screwdriver knock 190 microseconds
Time delay when engine idle 20 microseconds
Accelerometer B captured the idle “knock” first
Result from the excel calculator is
22.8 cm from accelerometer B
28.2 cm from accelerometer A

Using the measure tape we have marked the knock location related to our accelerometers, this is perfectly aligned with suspected cylinder 4.

Also looking on the highest knock amplitude, measuring the angle related to cylinder 4 giving us value 588 to 598 degrees and placing the rulers to idle waveform, it corresponds with angle where the intake valve is closing.

I think it is good demonstration of the NVH powerful tool combined with pressure transducer, Picoscope functions as rulers, math channels and Excel. Of course, it is possible to calculate the “knock” source with pen and paper, but why not to utilise excel when we already have laptop on the bench?

Please write your comments, I would like to learn other techniques, especially Picoscope NVH.
(excel file here)

Harshness source location calculator NVH.rar

(15.69 KiB) Downloaded 80 times

Thank you for your time reading this case study.

Regards,
Roman

[Steve Smith]

Hi Roman, thank you again so much for taking the time to share this case study.

I have just skimmed over as I have to hit the road now but I will share this with the Pico Software/Hardware engineers as an example of PicoScope in use at the "coal face"

I will read this fully tonight rest assured

Take care.......Steve

[PicoBarney]

Just popped over to say "Nice work", Roman, after Steve shared this with us within Pico. It's a great bit of diagnostics. Thanks for sharing.

[Steve Smith]

Finally read your case study fully Roman and found it both riveting and inspirational THANK YOU

As Barney mentioned, I have shared this with the Pico teams and with VM’s too as it speaks volumes of the ground breaking potential of diagnosis with PicoScope

I found the use of “calculations” to determine the propagation of sound through the cylinder head material awesome as not all cylinder heads will be the same (Rather than just using a standard value for Aluminium)

Here you first qualified the propagation time of your cylinder head material which has returned a highly accurate pin-point location of the offending noise

The bigger picture here is the inspiration of the next generation who will read this case study and wish to know more and no doubt replicate the techniques you describe as nobody wants to struggle with "hit-n-hope" diagnosis. Hats off to you and thank you again

Take care.......Steve

[intermotive]

Nice work Roman! I am impressed! Do you mind if I translate it to polish and show it at Polish Pico group?

[Rfmotors1]

Yes please, that would be very nice of you.
I could write it in Czech but I don't have any Czech connections even I am originally from there. Sending my regards to all Pico users in Poland
Roman​

[HarleyMike]

Thanks for the write up!
I understand the whole process except the last part.
How is the 588* determined?
How does it apply to cylinder #4

I understand how #4 was pointed out by the measurements.
Just not how to relate the highest amplitude to degrees on #4

Thank Mike

[Rfmotors1]

Hi Mike,
Thank you for the question, you are right as I did not properly explain how I got to the numbers.
Engine firing order is 1-5-3-6-2-4 and I had the COP pick up probe on Cyl 1, then just by first look on the NVH sensors the highest amplitude was visible and I have marked it by rulers to range 468 to 478 degrees in relation to cylinder 1.
Later when it was clear the noise comes from cylinder 4 area, I have virtually moved the 0 and 720 degree phase rulers 120 degree left where it would be if I have set the phase rulers to cylinder 4 ignition COP pick up points.
This would mean every number set by measuring rulers in the original waveform would be for 120 degrees higher.(as the phase rullers move left but the meassurement rullers stay)

So as my highest amplitude in relation to Cyl 1 was 468 to 478, then in relation to Cyl 4 it would be plus 120 on each ruler, therefore 588 and 598 degrees.
Then I have opened any pressure waveform and placed the rulers to position 588 and 598 degrees and this is the point where the intake valve is just about to start closing.

(In the picture annotation, I have wrongly typed 599 instead of 598)
I hope it helps to understand my logic I have applied to suggest the knock comes at cyl 4 intake valve closing time.
Regards,
Roman