Hi Everyone,

Just a quick question. I have an intermittent Check Engine light which is signalling a code 52. I have done the research and determined that this is generally considered to be one of the following items:

1) Bad Knock Sensor
2) Bad Knock Sensor Wiring
3) Failed ECM

My question is twofold. First, has anyone simply replaced the end of the wiring harness with the $30 part from Toyota and successfully resolved their code? If I am not mistaken you can get to the connector which is located between the back two cylinders on the drivers side.

Second, is it possible that a code 52 simply signifies that there is pre-detention happening in the engine? Or would that be noted by a different code? As some background the engine seems to have normal power and returns around 17-18 mpg on the highway (with 33x10.50 tires). However, I have been struggling with a low idle which happens randomly (400-500 rpm) and the engine does seem to have a slight misfire (just a small amount of vibration at idle). For a reference the spark plugs and wires were just replaced and I had to bypass the air reed valve in the EGR system because it failed.

I know this is a common problem, so any advice on the subject matter would be great. Thanks for the patience from the older members on here.

3VZE, 186k miles.

Code 52 means the ECU is not getting a signal from the knock sensor. It does not mean the engine is knocking - in fact, in the absence of a signal from the sensor, the ECU will retard timing to avoid possible knocking and subsequent engine damage. Usually this results in poor performance and mileage. Sounds like your issue is intermittent enough that you haven't seen that yet.

The general consensus is that the actual knock sensor fails much less often than the wiring. There are two components to the wiring - the first from the pigtail connector back to the ECU, which is fairly easy to replace, and the second is the actual pigtail to the sensor, which requires removing the plenum to access.

Some people have also claimed that aging capacitors on the ECU can cause this failure, and that replacing those accomplished the fix for code 52.

Go to the 2nd gen (86-95) forum on yotatech.com and search for code 52. There's a huge amount of discussion there about it.

Just a bit more theory on the knock sensor. The knock sensor is not there to tell you that your engine is knocking, at least not fundamentally. It is there to optimize ignition timing so as to advance timing just to the point below significant knocking, because that optimizes engine efficiency and extracts the most power from the fuel. The ECU runs an algorithm whereby it continuously dithers timing to keep the engine just on the threshold of low-level pinging to find the point of maximum allowable ignition advance. To do that it needs continuous input from the knock sensor. If the sensor signal goes away, the ECU is now flying blind and has to retard timing significantly to save the engine from possible destructive knocking. Code 52 is that "I'm flying blind here" call for help from the ECU.

In old days of vacuum and centrifugal timing advance we used to adjust distributor timing so the engine might ping slightly at low speed acceleration, and then assume that was the best we could do. In the early 80's knock sensors and engine computers took over that role to adaptively adjust timing, within limits, on the fly. I think GM had the first patent on this in the late 60's, but the first production application that I could find was by Saab in the late 70's.

Thanks for taking the time to respond @RonJR I took a look at YotaTech and found this huge thread (see below) that has some more information. So the actual consensus is this is safe to drive on and is just an electrical problem which causes the engine to retard the timing and therefore prevent pre-detonation. Just like you described. I think I'll give replacing the pigtail a shot in the hops it's an easy fix. Though I do not think I am that lucky!

Code 52 Tech Thread - YotaTech Forums

So per the thread I linked in my previous post I decided to look at the connector which attaches to the knock sensor's pigtail and then runs all the way to the ECM. Hoping that a short somewhere on this side would save me from having to replace the knock sensor/pigtail. I found something pretty interesting.

It seems that when the head gasket job was done (by the PO) something was done to the knock sensors wiring. As you can see in the photos the connector which should be plugged to the end of the knock sensors pigtail has been disconnected and the ground wire has been cut. There is now a bluewire leading out of where the pigtail should be.

Someone correct me if I am wrong, but shouldn't there be a ground wire running to the sensor as well? Is it possible they ran the ground wire out elsewhere or simply grounded it near the sensor? Looking for any advice I can get since this is going to make replacing the knock sensor/pigtail more difficult.

Attached Images

   

I "believe" the knock sensor wiring is as follows:

- A shielded cable between the ECU and the knock sensor connection.
- Another shielded cable (often called the "pigtail") between the connector and the knock sensor itself.
- The shields should connect together at the connector, so I think the cut wire you have is wrong.
- The shield should not be connected at the knock sensor.
- The sensor signal ground is through the block.

It's a quite noisy environment, so if the shielding is not properly connected the ECU won't get a useable signal from the KS and it will throw a code 52.

As an electrical engineer with a lot of experience in grounding and shielding and making low level measurements in the presence of noise, I will say that the way Toyota did it is not the best way, but it appears to me that they "cut and tried" until they got something that worked. The proper way would have been a fully differential floating and shielded sensor and cable system, electrically isolated from the engine block. I'm sure that would have cost a dollar or so more, though, so some junior engineer was assigned to experiment until he/she found a hack job that worked adequately, given some filtering software in the ECU.

@RonJR based on what I see in the FSM and my searches I would agree with you. However, the confusing part to me is why there are two wires in the connector which links to the pigtail. If the sensor is always ground through the block then why the two prong connector? In fact I have found several threads stating that there are two wires run to the knock sensors on the 3vze where as the 22RE only has a single wire and is ground through the block. Could it be that they installed an older version knock sensor and assumed it would ground properly through the block?

I have looked a variety of pigtails for the 3VZE through an assortment of years and all of them have two wires suggesting they are grounded elsewhere and not on the block by the sensors threads.

Finally, you state that the cut wire I have is wrong. Do you mean to say it should not be there or that the way whoever did this did so incorrectly? Basically are you saying that they did not shield the wire from where the pigtail should be to where it should connect with the cable from the ECU?

BTW I completely agree with you about the knock sensor (if only due to its location).

The second pin in the connector is to carry the shield through from the ECU wiring to the shield on the pigtail. Since the pigtail shield is floating at the knock sensor, its only connection point is at the intermediate connector. If that connection is also broken, the pigtail shield is left completely floating, which renders the shield ineffective.

I've attached a colored wiring diagram of the ECU wiring related to the knock sensor which might be helpful. ED1 (top center-left of diagram) is the intermediate connector that carries the signal wire and the shield to the pigtail.

Attached Images

 

So based on what you have explained here it looks like in my situation the wire is not being shielded properly. This lack of shielding is probably leading to the ECU not being able to detect the knock sensor readings due to noise which results in the Code 52 error code. This means I am probably not going to be able to avoid pulling everything off to get to the rest of the wiring...

As I see it I have two options both of which require pulling everything off.

1) Replace knock sensor with new stock pigtail and rewire the existing two pin intermediate connector with a properly shielded cable and run this all the way back to the ECU. Then connect the intermediate connector and the pigtail in the manner it was made stock.

2) Replace knock sensor and utilize properly shielded cable in place of the current blue cable. This would eliminate the electrical noise component and might be the easier solution for me. I would only have one cable between the ECU and the knock sensor.

What is your opinion of the above? Would both work?

Electrically your two solutions should be equivalent, so either one should work. I think the reason Toyota split the two cables at the connector is they expected the pigtail, which is in a fairly harsh environment, to fail more often. With the connector you only have to replace that part (which is still a big job), but you don't have to get into the wiring back through the firewall and dash to the ECU. Also, the pigtail lives in a fairly high temperature environment, so it needs to be able to survive that. The rest of the cable doesn't have to be quite as rugged.

But as long as you're into both areas, yes, a single cable should work as long as it can handle the temperature all the way to the knock sensor.

And, as the wiring diagram shows, the shield should be grounded at the ECU, and at that point only. (point EC).

Gotcha thanks for all of the help RonJR. It's handy to have a EE around here. Just a few more questions.

My initial gut says to replace the cable with high quality shielded cable and then wrap it with some sort of heat protection to ensure it lasts a long time. I assume this would not cause any issues with noise since the original cable would be shielded. I will update this thread when I get a chance to do all of this prior to the holidays so there will be some sort of resolution. Thanks again for all of your help!

That should work fine. If the shield is good, the electrons won't know what you did outside the shield.

Well @RonJR I was thinking about this tonight while putting together the parts list and instructions for this fix and I decided I had a few more questions and some things I should check before just throwing time and money at it.

One of the things I should check is to make sure their is continuity between the ECU and the sensor. I should be able to do this without removing anything. Ideally that blue wire will run to the ECU and I should be able to get a lead between it and the wire to the knock sensor. If there is no continuity I can trace the wire and hopefully find where the issue is.

Anything else you can think of that I can troubleshoot before committing to this project.

You can check continuity from the connector back to the ECU, and should probably do so. Unfortunately, you can't check continuity to the sensor itself without removing the plenum to gain access to it. The knock sensor is a piezo crystal, and has no DC continuity - it behaves much like a capacitor in that respect, so even a good one will measure infinite resistance (open circuit) with your ohm meter.

It's basically a crystal microphone that will put out several hundred millivolts of signal if you whack the block with a hammer. If you have a way of detecting that signal (oscilloscope, amplifier with a set of headphones) you could check for connection that way.

Hey @RonJR I wanted to thank you for answering all of these questions several weeks ago. Last week I was finally able to to pull everything off to successfully remedy the knock sensor code. I ended up getting a new pigtail from Toyota and repairing the cut wire shown in the thread above. I also replaced the knock sensor which looked original while I was at it. Hopefully this thread will prove useful to someone else with similar questions. Honestly the knock sensor replacement on the 3VZE is a bear only because of all the things you have to remove to get to it, but nothing is actually difficult, just time consuming. Anyway, thanks again!

If you get this, How did you fix the wire on the Engine Harness side?
Mine was severed and spliced the shielding together and spliced the signal wire together but still showing an open signal.
Replaced knock sensor and pigtail as well.

Help