Splitting the output from one knock sensor into two for the ECU

[DougyD19]

I have an Infinity 94 Q45 that I had a knock sensor die on. The engine is old and the parts are brittle and unfortunately the knock sensors are deep in the middle of the V. I ohm tested the lines going to the knock sensor and the one in my left bank is not outputting any ohms.

I would like to know if it is possible to just wire the output leads from the good knock sensor back into the bad sensor's wiring? With this setup, I'll still have a knock sensor....that can pick up knock, but it just won't be able to hear the left bank as good. I'm also curious if the ECU will know the difference or if the signals will change from splitting the signal.

 

[WGJ]

Wiring the knock sensor as you suggest will change the input impedance that the sensor "sees" at the engine control module.
Just assuming that the engine control module provides a suitable load for everything to still work, then depending on how the software filters the knock sensor signal, you may still not "hear" the signal from the bank with the dead sensor.
Software often uses time windows to look at knock signals and these time windows will be relative to the speed of the engine and (probably) the cylinder closest to the sensor.

Bill

 

[Thecardoc]

The two sensors are in fact different, they respond to different frequencies which allows for a more accurate recognition of detonation, and better ability to disregard mechanical sources of noise which might sound like detonation to a single sensor.

Attempting to jumper them together wont work, because the PCM actively tests the system, and watches for the two sensors to respond independently.

 

[DougyD19]

Thanks for the replys. I think I'm going to have to go to plan B. I'm going to buy a second knock sensor and mount it in a different area on the bank where the dead sensor is.

I now have to figure out how to wire the sensor in.
Here is what I have to work with.

http://www.q45.org/FI KS PINS.JPG

The knock sensor is here -

http://www.autozone.com/autozone/ca...romWhere=&searchText=SU2074&_requestid=904449

I've wired a GM knock sensor into an MR2 before, but this is going to be more difficult.

 

[patprimmer]

As has already been said, the OEM ECU expects a certain signal pattern from the knock sensors. This pattern is dependent on their location, so moving them might have undesirable results.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers &

http://eng-tips.com/market.cfm

for site rules

 

[MikeHalloran]

Uh, "the one in my left bank is not outputting any ohms" does not sound like the appropriate diagnostic test. It would be for, say, a TPS, but I thought knock sensors had pulsed or switched, not resistive, outputs.

Did the ECU say the knock sensor isn't working? Why did you break out the ohmmeter in the first place?



Mike Halloran
Pembroke Pines, FL, USA

 

[Thecardoc]

"the one in my left bank is not outputting any ohms"

Mike wrote. "Did the ECU say the knock sensor isn't working? Why did you break out the ohmmeter in the first place?"

Because that is how horrible trouble trees are for technicians when they are faced with diagnosing a circuit for the first time. It's no surprise that non-technicians will make the same error, no matter how much engineering education they have. People take, "Check continuity of wire XXX" as an indication that an ohmmeter should be used at each end of the circuit. Knock sensor diagnostics are often listed in the trouble trees (GM example) as test resistance of the sensor it should be 2000-3000 ohms. After following examples like that enough times, techs get trained to do that as a legitimate step. Then it takes a lot more training to break them of that habit and get them to learn to test dynamically.

 

[thruthefence]

Aren't these things piezo electric devices?

 

[Thecardoc]

"Aren't these things piezo electric devices?"

Yes, but some (most today) require a bias voltage supplied to them in order to turn them on, and then they produce their signal over-top of the bias voltage. Typical strategy is to use a 5v reference, which gets dropped across an internal PCM resistance when the sensor circuit is complete and inactive. That's part of how the PCM determines "continuity" in the circuit, and is the way techs should test it also, not with an ohm-meter.

With a scope the tech would see, 0v, 2.5v, or 5v at the PCM connector, and then he/she would need to repeat that check at the knock sensor. With 2.5v the tech would then have to go to an active test and directly monitor the signal.

BTW "silver bullet", the car in question typically loses connection between the harness connector and the sensor creating a randomly open circuit. The tech would see reference voltage such as 5v, and then that would possibly spike down to whatever signal the sensor is producing during an active test.

 

[patprimmer]

I would have thought the very first test would be to see if the terminals were tight and if not to squeeze them up a bit so they cut through the oxide layer as you push the connector back on. Only after that does not work would you look to test or replace potentially faulty parts

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers &

http://eng-tips.com/market.cfm

for site rules

 

[Thecardoc]

Hi Pat. That would be the first test, if one didn't have to do 3 hours worth of work removing and reinstalling the lower intake manifold to do so.

 

[DougyD19]

I checked the code output on my ECU and it said knock sensor. I then checked the ohms on the pins on the last connector before it dissapeared beneath the intake mani. My right bank said 550 ohms and my left bank said 1. I also checked the voltage. My right bank said 4.8 volts and the left bank said 2.4 volts.