No power until 5500rpm?

[jdwk]

Oh that's interesting. So it runs in a limp mode if it detects the knock sensor circuit is open. Well I guess the next thing to try is someone else's knock sensor.

 

[stevew145]

I have 2 old motors with knock sensors in my parents garage. Friday when I go out there I will swap and see what happens.

 

[f1yen]

Or instead, if you are confident..... instead of dealing with the CEL when unplugging the knock sensor, plug in one of the spare ones, that is not hooked to the engine, thatway no CEL and no code, but I would not run it that way longer than needed for diagnostics.

 

[jdwk]

Well if more than one knock sensors are detecting knock then he probably is pre-detonating for some reason. Could be running lean, could be carbon build up, could be running hot.

The strange thing is, is that the pcm is advancing timing again at 5500. The higher the rpm, the more dangerous knock is.

When I was tuning with my FIC, anything above 12.0 AFR would start pinging in the upper rpm range. It sounded like gravel hitting the firewall. I didn't ever stay on it to find out, but it didn't feel like the PCM was pulling timing.

I've always wondered how a knock sensor works, I need to look this up.

 

[stevew145]

Basically, a knock sensor is a microphone tuned to listen for certain frequencies. I basically live at sea level, and the air here is very cool and dense, especially in spring. It never did it before, only since I put this engine in, using the knock sensor on the new engine. I am going to bet its either a broken wire(doubtful because it would throw a code for open circuit like it does when its unplugged.), or its a bad sensor.

 

[jdwk]

That's pretty much what I figured. We are at 1300 or so here in Phoenix and there is very little humidity or low temps to worry about, which might be why I was able to get away with some audible pinging. It might have been pulling timing, I had a crappy scan tool at the time and didn't keep an eye on the timing.

 

[f1yen]

Measure the resistance of your knock sensor! i know there is a spec.. I can go look it up....

 

[jdwk]

Wow, just looked up the service manual.

MONITORING CONDITIONS
— Engine load is above 50%
— Engine coolant temperature is above 60 C {140 F}.
— Engine speed is within 1,500—5,000 rpm

So that explains the timing coming back at 5k rpms and why I didn't feel timing being pulled. That seems really dumb to stop monitoring above 5k since there is more stress on the engine then, but maybe it isn't accurate or fast enough at that point.

Here is the resistance btw.

CHECK KNOCK SENSOR RESISTANCE
• Measure resistance between knock sensor
terminals (part-side).
• Is resistance approx. 560 kilohms?

 

[stevew145]

I'd say it doesn't listen after 5000 rpm because there is just too much other noise. Thanks for the resistance figure, I am going to check it now!

 

[stevew145]

When I check the sensor there is no continuity, but it doesn't throw a code for open circuit unless I unplug it. tomorrow I will replace it and go from there.

 

[pbarton14]

im very curious too see how this pans out.

 

[f1yen]

I checked mine today, and it has that oily/gummy stuff oozing out of it. I think i read somewhere that it is an indication of the sensor failing....?
let me know how you got it off and on.... I am stumped. I just cant seem to get a tool in there to grab it...

lmao and looking at the manual: 1)unplug sensor 2)remove sensor >>replace in reverse order ha

 

[pbarton14]

More than likely you'll have to take the IM off

 

[pbarton14]

I tried this on my own msp and it gave me all of my power back too!

 

[destineal]

what did u try? taking the knock sensor out?

 

[stevew145]

The remove the knock sensor;
get under the car, remove oil filter and remove sensor with a wrench. On the top you will see the connector clipped on near the intake manifold/fuel rail.
I checked the knock sensors resistance again today and it turns out it has 530kohms resistance. The 2 others I have are 565 and 559.
When I use the other knock sensors the same thing happens. I took one knock sensor, connected it the the harness and stuck the actual sensor in a random clip on the side of the engine bay. Power is still there, and the engine light does still come on from time to time, but not nearly as often (maybe once per 60 miles). If I mount the sensor near the ignition coils on the top of the engine, the car runs the same as if it is plugged in properly. I think there is just too much vibration in my engine and the knock sensor is going nuts. Maybe this has something to do with stiffer engine mounts, maybe there is a problem with my ecu, maybe there is a problem with the wiring for my ecu. I am thinking about playing with some resistors inline with the knock sensor to see if that changes anything.

 

[pbarton14]

Yes destineal I only unplugged the connector

 

[jdwk]

You would think if vibration was the issue, then there would be a lot more people with this problem, as almost everyone one here has some sort of upgraded mounts. I am thinking you are actually knocking. Unfortunately standalone knock detectors are pretty expensive. Maybe you can find someone to borrow one from.

 

[stevew145]

Even when the sensor was not screwed into the engine, but just placed on it and strapped down it ran the same as if it was screwed in. There is no way it was detecting knock from above the head. I really don't know what to think anymore. The engine runs as smooth as silk, no detonation(I am familiar with the noise) I am running 1 stage colder plugs, the iat is always low, I am at stock boost, and i have tuned my afr even richer now, It never goes past 11.5 at full throttle

 

[jdwk]

That kind of points to the PCM then since even the other sensors trigger it as well. It's rare for PCM's to fail but it does happen.