Does anyone know how to electrically check a knock sensor? I am getting a knock at moderate acceleration through the 1500-2000rpm area, which seems to be all the time in my 97 C36. There is, I believe, only a single sensor on this engine, but connecting a timing light and whacking the block with a hammer is not really an option, as I do not have access to the high tension wiring while the engine is running.

__________________
'96 C280 (gone)
'97 C36
'05 C230k

You need oscilloscope to test it

Attached Thumbnails

 

That's just to see it's output, but I'm wondering if there are any static tests I can do that will indicate the integrity of the sensor or lack thereof. Perhaps I should just test the leads and buy a new sensor.

__________________
'96 C280 (gone)
'97 C36
'05 C230k

After a bit of research, I can answer my own post somewhat.

The knock sensor common to all the M104s and several other engines is a broadband piezoelectric device. It is not tuned to a specific frequency of vibration, but as a broadband device, it responds to a range of vibrational frequencies so Mercedes can use the same one in many applications. The sensor itself is made by Siemens of France.

The knock sensor sends a voltage it generates from its piezoelectric properties to the engine computer, and the ECU interprets the voltage and retards timing appropriately. The ECU does not send a voltage to the sensor, so therefore the ECU has no knowledge of whether or not the sensor is even connected, (slight amendment: a completely disconnected sensor will set a code in short order) let alone when it goes bad as the no-knock condition results in zero voltage sent. There is no "live zero" to indicate failure if absent. No code can or will be set for a non-functional sensor. A code is set when the sensor asks for maximum retard on repeated occasions.

All the dealers I stopped at on my way up north the other day to try to find one claimed they had never sold one. I had one ordered and picked it up today ($74 CDN). I changed it in the back of the dealer's parking lot in less than 5min. Problem solved. My C36's engine is quite noticably quieter mechanically, and doesn't knock under load any more, although I suspect I have taken a slight hit on fuel mileage from what I used over a few hours after installing the sensor, but that is to be expected. Possibly the problem is not restricted to my AMG, just more noticable due to the tuning of the engine. I think I'll put a new one on my wife's C280 as well.

__________________
'96 C280 (gone)
'97 C36
'05 C230k

The leads from the sensor on the M104 motor go to pins 41 and 42 on the connector bus near the ECU in case anyone is interested. They tested out fine, which is why I just went ahead and ordered a sensor.

__________________
'96 C280 (gone)
'97 C36
'05 C230k

I wish there were more posts like this. thanks

Mr Benzfan,

This is very interesting. I have had my 98 C230 for 1 year and it has always had a knocking problem. I called the dealer and they said to use regular. No kidding, they said this was recommended by "headquarters" for the M111 engine. For the last year I have tried every different octane and every different brand and none of it made any difference. I changed out the cat and that did remove a secondary "voluntary" rattling but the primary rattle is still there. I cannot swear it is knocking but it fits all the criteria of knocking, sounds like it and is load related. On the other hand I have a hard time believing that a piezo sensor could go bad. Should I just change out the sensor and cross my fingers?

Thanks,

Mike

__________________
1998 C230 330,000 miles (currently dead of second failed EIS, yours will fail too, turning you into the dealer's personal human cash machine)
1988 F150 144,000 miles (leaks all the colors of the rainbow)
Previous stars: 1981 Brava 210,000 miles, 1978 128 150,000 miles, 1977 B200 Van 175,000 miles, 1972 Vega (great, if rusty, car), 1972 Celica, 1986.5 Supra

I'm afraid I can't see the logic in using regular. In doing so, I would expect that you would induce more knock, not less, causing the knock sensor to retard timing even more. This should still make the engine knock as much as before, but perform worse. A knock sometimes indicates low flow in the EGR system, but my C36 doesn't have EGR. I have heard of MAF sensors becoming less accurate with time, caused by a slight buildup on the heated element. This will cause it to under-report the amount of air flow and cause a slight lean condition under certain circumstances, much as an intake leak downstream of the MAF would do, but I have no idea how to test the MAF sensor if it is not throwing up codes other than putting a new one in and seeing what happens.

__________________
'96 C280 (gone)
'97 C36
'05 C230k