CarTalk.com Best of Deals Car Reviews Repair Shops Cars A-Z Radio Show

Knock Sensor to replace or not

So my 99 lexus rx 300 with 200K+ is throwing the codes for both knock sensors. This most likely means that the wiring harness is bad. It’s a lot of work to get to the harness (and sensors) so I’m considering putting in two new sensors (and the coolant bypass hose ofcourse) while I’m changing the harness to avoid this job down the line. The Airtech/Wells sensors are about $100 each from Rockauto. The questions should I just pay for the parts. My only concern would be if there was a problem with the new sensors which I don’t find and I have to do it all over again. Anyone ever install a flakey “new” part only to have to replace it.

I can’t tell if you are going to do the work or if you are asking if you should buy the parts and pay someone to install them. If you furnish parts the mechanic will not warrantee them and if they are bad you will pay labor to remove and replace. Yes, new parts can be bad but not common.

I’ve had a few new parts in the last 2 years that flaked out within a year or didn’t work out of the box. I’d stay away from the budget lines they have. If it’s a lot of labor to install, try to get OEM.

Some examples, I bought a Standard T-Series TPS sensor, failed just after 1 year, original lasted 20+ years.
Airtex Advantech EGR valve, failed within 1 year, original lasted 20+ years.
Thermostat housing from Advance Auto, don’t remember brand, leaked after 4 months, original lasted 12 years.
These were all easy to change. The worst was a cheap fuel pump a mechanic installed, that was complete junk from the start. I replaced that myself with the OEM brand. That’s a lot of labor to change out. Not worth risking an aftermarket part on.

Thanks for the notes guys.

I’ll be doing the install myself. I’m staying away from the budget line of parts. The OEM part is close to $200 a pop from the dealer. The OEM-like Standard Motors (not T line) or Airtex part is about $100, which may just be a reboxed Denso. I believe it’s Denso that makes most of the sensors on this car (at least the others I’ve changed were).

My options are.

  1. Just swap the harness for <$100 in parts and a risk and extra half day of work. (assuming the harness is at fault)
  2. Buy better aftermarket parts and spend about $300 in parts and full day work with much smaller risk of an extra day work.
  3. Spend about $500 in parts and fulld day work and get the guaranteed part (which might just be the same as part #2)
  4. Pay a Mechanic over $1200-1500 to do guaranteed work.

Note: I have an extra car, So I can park this one while I work on it. I could pull the part and compare, reorder if necessary. That’s my trick with old cars. You always need two. One to fix and one to drive while you fix the other. haha. My wife thinks I’m crazy, she’s probably right.

I’m not too good at diagnosing to be honest. There’s a couple guys on youtube who have great vids on diagnostics with the scan tools, schrodinger’s box and scanner danner. Most “pro” mechanics aren’t good at diagnostics either, from my experience. They often just throw parts at it which I can do myself. I’ve got 2 cars also, one 22 years old and other 12 years old. I’d hate to rely on just one. Maybe someone else here will give better advice. You can usually check knock sensors with a multi meter.

On what do you base this assumption?
Is it not more likely that transient pulses are coming from the piezoelectric crystals (the knock sensors) due to pinging or some other cause of mechanical inputs to the sensors?

What’s the overall engine condition? I’ve seen a worn rod bearing be sensed by a knock sensor causing the ECU to “correct” as if the engine were preigniting.

No disrespect intended, but I think more diagnosis is in order before blaming the knock sensor wiring harness. The fault codes may be trying to tell you something important.

The engine is fully service up 200k+

Both Knock sensors went bad simultaneously and the tech who originally looked at it claimed he was able to get a mirror or scope under the intake and observed some signs of rodent damage on the harness. I was skeptical because I’ve fixed symptomatic CEL codes by repairing something different than the stated code. I contacted a different lexus specific tech to advise if the knock codes were in fact not just symptomatic of something else. He confirmed that the Knock sensor codes are one of the rare straight forward codes. He also cited the dealers tech manual for me.

Blockquote
Is it not more likely that transient pulses are coming from the piezoelectric crystals (the knock sensors) due to pinging or some other cause of mechanical inputs to the sensors?

I’m not sure what you mean, If the engine was pinging the knock sensor would do it’s job and retard the timing. The code is 0325 which is saying the sensor is faulty, not that it’s just detected knock.

The Commons Common Causes of P0325 are
Failed knock sensor
Fault in the knock sensor circuit
Failed PCM

I would love to dream up some other reason as to why my problem is happening, and I usually do, but this seems pretty straight forward. Without removing the knock sensor and testing what more diagnosis would you recommend?

The odds against that happening are very very high. If critters might be nibbling on the harness I’d suspect that first. Not sure I’d spend $200 on 2 sensors based on that.

It’s a good bet there’s an electrical problem common to the knock sensors, but there’s other possibilities that would explain knock sensor codes. A tank of bad gas, spark plugs who’s electrodes have too wide of a gap, problem w/the advance/retard mechanism, variable valve timing problem, etc. It may be that even when the computer retards the ignition the maximum allow the engine is still pinging, and the knock sensors are working, and just reporting the pinging they are hearing.

We had a post here from someone a few weeks ago with a knock sensor problem. On their vehicle it was something like an 8 hour job (billing time) to replace the sensor. I guess it was really buried deep in the engine, the designers must have decided it was such an unlikely part to fail, no worries if it costs the owner 8 hours of labor billing to replace. .

Only to a point. If the sensor detects a chronic misfire it’ll trip a fault code.
However the rodent damage in this case just may be the culprit.

I find rodent damage just as often as failed knock sensors. Bad gas, worn spark plug and mechanical problems generally won’t result in a loss of signal from the knock sensors. This is a knock sensor problem, not a detonation problem.

Rodents like closed-in cubbies like that.
I would first get into that area and investigate before buying any parts.
You can always just put it back together and run it while you wait for any parts or supplies.
The harness might be easily repairable.

It’s certainly not a quick disassemble/re-assemble job. I think it bills at about 5 hours. A new harness is inexpensive at ~$30. I’ll probably just buy the harness and bypass hose to have on hand. Disassemble and test the components, then decide whether to do the sensors. If the harness is certainly damaged I may just give the existing knock sensors another chance. I’ll keep you guys posted, if your curious to know what it was.

I think you are making the right decision. Everything points to a wiring issue, like @Nevada_545 says. So you go in, and make your decision when the parts are laid out in front of you. Rodent damage should be pretty easy to see. If the cables are good, then maybe the sensors are bad, but it does seem unlikely that both would go at the same moment.

Owning 2 older vehicles makes lots of sense to me. It costs no more, and it gives you the time to do work like this. I own 4 running, completely functional Honda Elite scooters (85 to 90), and I ride one of them daily, but all 4 get exercise. Why not?

I checked knock sensors once on my car. I don’t remember the details but you just use a multi meter and tap the block and look for the voltage change. There’s a bunch of tutorials on the internet. That should tell you if they’re functioning.