mtnfreak

I know a lot of people have dealt with this, but has anyone actually resolved their issue? CEL came on for the 1st time earlier this week - code 52. On warm days, if I drive a couple of miles, shut off the engine and restart, the code is gone. If the temp is under 65 degrees or so, the code remains. While the CEL is on, I have a serious loss of engine power, especially in the mid to high rpms. I doubt the sensor itself has gone bad, so I'm going to look at all the wiring this weekend. Anyone have suggestions on other things I should look at or that I might miss?
89 4Runner, 3.0 V6, 210k miles, most recent engine work was in early March so I doubt there's a connection.
Thanks!

AgRunner06

If your sensor is that old, I would just replace it and the wiring.

mtnfreak

Hoping to avoid replacement for several reasons:
1) Usually takes something to cause electronic parts to go bad - nothing has happened, to my knowledge.
2) I know many people have replaced theirs but not cured the problem.
3) Gaskets alone can top $150 for this job.

mtnfreak

Update: Pulled the throttle body and ECU today. Traced down all of the wires above the intake manifold and couldn't find anything I thought was wrong. One wire went to ground - through the ECU! Might be a problem there. It read 1.6 ohms, but I have no idea what's normal for this circuit. The other wire also went to the ECU and read 0.6 ohms - only slightly above what my extension tester reads on it's own, but, again, I have no idea what it should read.
Cleaned the TB and checked the TPS while I had everything apart. TPS wasn't perfect, but it's probably as close as my clumsy, scarred hands can get it.

midiwall

Just another view of it...

1. I've had 2 sensors go bad in the last month(ish). Yes, they really were bad. I couldn't believe it, they're just piezo microphones. Heat from the block may have been a contributor.
2. On an OBD II vehicle, we get a separate code for each sensor. When I first dove into mine I crosswired the sensors (sensor 1->input 2, sensor 2->input 1) to verify that it was the sensor. The code shifted to the "other" sensor which confirmed that the wiring was good. I eventually tore down the top of the engine, replaced the sensor and the codes went away - for about two weeks, then the other one went.
   
   I have a hack running right now to get around tearing the top off again, where I have both ECU inputs reading from one sensor.
3. A full set of gaskets (two intake manifold gaskets, the plenum and the top water hose) for mine are that expensive as well, but I got "lucky" in that they're pretty reusable. I repaired some tears in the cooling jacket portion of the manifold gaskets with Permatex red and things are fine.

btw, the reason that the engine runs like crap with the CEL on is that when the ECU sees a knock sensor fail it pushes the timing into full retard as a failsafe.

I don't know your engine, but if you have two knock sesnors you may be able to cheat things by Y'ing the ECU inputs to the "good" sensor.

miket223

Reading resistance values on the knock sensor is pointless. Toyota's defines a knock sensor test and hammering on the exhaust manifold and observing if the timing retards. Another variant of this is to measure AC voltage coming out of the sensor. That is AC, not DC voltage, that the knock sensor outputs.

mtnfreak

Interesting - AC, huh? How does that work with being grounded to a DC system? Is there something in the ECU that can make the switch?
I've tried hammering all over the engine bay, but the V6 is so crmmed in that it's hard to get a good 'knock' without damaging anything. I've tried hitting an extension bar placed against the headers and block, but didn't see any change in the timing.

midiwall

The knock sensor is AC 'cause it's basically a piezo microphone. It's a small crystal that reacts well to shock waves through contact versus changes in air pressure like a normal dynamic mic. That's why the test process is to strike the block with a (preferably) steel hammer. Or, tear out the manifold, pull the sensor and strike it directly.

While it's true that automotive systems are largely DC based, the ground reference is just that - a reference. It doesn't matter if you float AC or DC against it, the important thing is to know what to expect when you measure it. And, in this case, the ECU needs to keep it isolated from any DC sensor signals as well.

There are other AC components in the electrical system. The crankshaft and camshaft for example are hall effect transducers which effectively generate AC. You can see it on a scope, you'll have potential above and below the ground reference as the magnet "makes" and "breaks" across the transducer.


Note that if you do decide to measure the voltage, it'll be small... My FSM is out in the garage, but I think you should expect under 100mv.