Cutting knock sensor inputs to ECM. 3vz-fe

[vibrosies]

I've read in various posts - usually tuning sites - that 3vz-fe motors are quite solid and can take a lot of abuse without breaking down. Mine has close to 300000km on it and runs really smooth until the following error kicks in.

1 site to mention:
http://www.mr2.com/forums/v6-owners-...toyota-v6.html

Error codes show I have 2 bad knock sensors in my 93 Camry LE V6. Both are reading approx 2 MOhms to gnd - measured @ ECM inputs. If I understand correctly, they should read open till knock is sensed which is said to be very rare on 3vz-fe motors anyway according to the tuning posts.

The CEL comes on after approx 5-10 min of driving and at that point I feel a significant loss in power/ fuel economy. This is the ECM going into safe-mode.

Could I just cut the knock sensor wires to the ECM to fool it? The ECM would see this as an open, therefore, an ideal condition preventing it from going into safe mode. I'm not running turbos etc, so I assume im not subjecting this motor to too much abuse. Would this work?

I would change the sensors but that seems like a lot of work ($$$), not worth the value of the car at this point. Just need this one to last me another year without spending too much other than on DYI stuff and basic maintenance.

Thanks for your input.

[vibrosies]

Hey, there's gotta be someone out there who must know the answer to this?

If I disconnect knock sensor inputs, would this fool the ECM and get rid of check engine light?

Thx!

[93celicaconv]

Is your 93 3VS-FE engine OBD-I or OBD-II compliant? What codes did you get when you checked?

If you want to test your theory, why don't you disconnect the wiring harness from the knock sensors, instead of cutting the wires? Or are the knock sensors too difficult to get at to simply remove the wiring harness?

I think (but I am not certain) that disconnecting your wiring will still trigger a CEL code, because your ECM still gets a signal back from the knock sensor when the engine is turning. I think the lack of this signal will still put your ECM into limp mode, preventing proper spark advance and will run your fuel a little rich, all in the hopes to not generate engine knock until the problem is resolved. I'm just not familiar with the programming of your ECM.

I have a 93 Celica with a single knock sensor, and this engine (5S-FE) is OBD-I compliant. It will set a CEL if the ECM detects an open or short in the knock sensor circuit. That is as much as I can help you.

[vibrosies]

Thx for the feedback 93celicaconv!

The 3VZ-FE is OBD-I compliant. The codes set are 52 and 55.

I guess the only thing I can do is try it and see. I will report back the results.

I would love to access the bloody sensor - just too much stuff to dig out and put back - don't have the knowledge / practice to do all without fearing to create bigger problems for myself.

[93celicaconv]

Well, for what it is worth, this was a question/answer from a forum regarding the same issue with the same engine, only on a 93 Land Cruiser. Based on what this says, for both knock sensors to suddenly go bad at the same time (I think that would be highly unusual), there is a reference to the wiring or the shielding to the wiring becoming damaged. I wonder if that is what has happened in your siutation. See below


Question
1993 land cruiser with codes 52 & 55 dies at times and runs poorly.
No misses and it runs smooth without ck engine light on and codes comming up
Submitted: 310 days and 17 hours ago.
Category: Toyota Value: $15 Status: CLOSED
+Read More
Posted by Skyvisions 310 days and 17 hours ago. Answer


Hello,


You have many options with today's internet so Thank you for choosing JustAnswer.com.



These two codes that you listed above are for the knock sensors. They will have no effect on how the engine performs as far as idle and smoothness. They only control ignition timing based on any engine knock or pre-detonation. Can you give me some more detail on your problem? Your initial statement contradicts itself. On one hand you say it dies at times and runs poorly yet you say there are no misses and it runs smooth can you explain?


RXXXX XXXXXX


P.S. I take great pride in providing quality responses for the people I assist. If you believe that I have been professional and helpful in my assistance please don't forget to give a positive rating for this exchange and click the accept button. If you feel my answer does not address your issues, please just tell me and I will opt out, allowing another technician to address your situation. My Goal is to make this experience one you appreciate.


310 days and 16 hours ago. Reply
When I first start up the vehicle and drive it runs fine. After running for about 5 minutes or so and going from a slow speed and accelerating at times the ck engine light comes on and a noticable loss of power is felt. If I stop and shut off the car and then restart it the ck engine light is off and it will run fine untill the light comes back on and loss of power is felt. Sometimes it runs for quite a while before the ck engine light comes on. It usually happens when accelerating from idle but sometimes happens when just driving at a steady speed.

Accepted Answer
Okay so what we are getting here is a loss of performance probably due to the knock sensors causing the computer to go into a default or retard situation on the ignition timing. Ideally in order to test these knock sensors you need an oscilloscope. The signal wire to the computer from the knock sensor is a shielded single wire cable. Normally what we find when you have these codes set is either a bad knock sensor or somewhere in the harness the grounding cable or shielding gets shorted to the inner cable. This will easily be detected on oscilloscope. If you do not have an oscilloscope you'll need to disconnect and test the integrity of the input wire from the computer to the knock sensors. Once the wiring has been ruled out as okay and not shorting out then you need to replace the knock sensors or test them with an oscilloscope. Based on your description that this happens on acceleration makes perfect sense. Check the overall condition of the wiring harness and the wires going to the knock sensors to see if there is any rodent damage or any signs of chafing or rubbing against anything that would be grounding. Let me know what you find. I will check back with you tomorrow. Do you have any way to test as I have described or do you understand what I'm referring to above?



Read more: http://www.justanswer.com/questions/...#ixzz0lO8x25Uz

[73sport]

FWIW: 2 - 3 mega ohms is effectively open. I'd bet your issue is something else. if your fingers are touching the leads or if theres any voltage present (feedback or bleeding down capacitor) those will affect the reading.

Had these measured in the 100's or 1,000's now you have something the ECU may care about.

As stated above, find them and disconnect them. Besides, is sounds like your measuring the input into the ECU too. You need to isolate the sensors and disconnect them. Forgive me if I misunderstood something and maybe you did do this. . .

Question, do you hear the engine knock or ping? Have you checked the ignition timing? Changed gas station quality?

THings that lead to detonation; fouled and carboned plugs, running hot, excessive ignition timing, running lean, bad gas. Are any of these other things true?

[93celicaconv]

Just be aware the ECU doesn't measure resistance levels involving knock sensors. Knock sensors a piezo-electric devices in that they generate AC voltage when they pick up shock waves. I don't know if these knock sensors are full range sensors (in other words, they generate AC voltage no matter what frequency shock they pick up) or if they are resonance knock sensors (tuned to the typical shock wave frequency pre-ignition would tend to generate, which is somewhere in the 5,000 - 9,000 Hz frequency). Since the AC voltage being sent to the ECM is extremely susceptible to interference, the wiring leading from the knock sensors is shielded from other wiring to prevent interfernce and the ECM's interpretation of the signal.

It is so unlikely that both knock sensors would fail at exactly the same time. I'm thinking that the shielding on the wring for these sensors is damaged (exposed and contacting grounded sources, or exposed and contacting other wiring), and that is the source of this issue.

[LUVLEXUS101]

I connected both of my ecu inputs for knock sensors to one of the sensors and the code went away, car worked fine although I only got about 12 mpg.

[vibrosies]

Very informative responses - thx!

One detail that I should have mention initially was that the ECM didn't report faults on both sensors at the same time.

About 3 years ago, only the right bank sensor produced an error.

At that time, I had cut out the bad sensor input at the ECM and spliced in (shared) the 1 good sensor for both inputs @ the ECM. I was ecstatic that this method worked - ofcourse many thanks to TN forum for the guidance.
The ~ 2M Ohm reading was measured with cold engine - maybe engine temp plays are part on the reading, not sure though. Will have to re-measure to confirm.

Recently, I had an overheating issue (bad rad) which may have compromised the shielding (as you mentined) - makes a lot of sense.

Gotta check that out - as much as possible - given tight conditions in the path where these sensors are housed.

[vibrosies]

It didn't work. Opened up both knock sensor inputs @ the ECM, but the CEL eventually came on just like before (after approx 10 min - when rpm raised to approx 2000).
This goes to prove that the ECM is likely checking for open circuit too.

Got no choice now but to find out how to get into the knock sensor locations to inspect them and its wirings (sheilding etc.)

Buy the way, I measured the sensor impedance to gnd the got the following.

Right bank: ~ 600 Ohms to gnd.
Left bank: ~ over 3 M Ohms to gnd.

Now, time to look for well detailed instructions (hopefully with lots of pics) on how to get to these sensors.

[93celicaconv]

Which knock sensor failed 3 years ago, the left or right bank? I don't know that a reistance check is a way to check the condition of a knock sensor (since the ECM detects the AC voltage that knock sensors create), but the resistance difference between two supposedly identical knock sensore is substaintial.

Could it be that you had a knock sensor go bad 3 years ago, and now you have a wiring problem? Could be. I don't think you will like how much work will be involved to get access to those knock sensors - I think it will be a lot.

[73sport]

This excerpt is from the 94-95 manual:

6. INSPECT KNOCK SENSORS

Using an ohmmeter, check that there is no continuity between the terminal and body. If there is continuity, replace the sensor.

In short: if you can find the wires to the K/S, unplug them and measure bewteeen the input leads and chassis ground and it reads less than 1m I'd venture to guess you have an issue.

Still, I go back to this. Does it actually run hot, ping, or detonate?

[NDR008]

That only checks if the sensor is shorted.
The sensor can still become deaf, i.e. is not producing the right waveform - I have compared to identical knock sensors (one good one bad), and tapped them while connected to a scope, the bad one produced waveforms but not as peaky as the good one.
Also, that test won't indicate if it is the cable's that's gone.

For what its worth, for the first time I am stuck on this problem, after having solved many myself on 4E-FTE's... now the victim is a 4A-FE for me.

I am going to pass temporary new shielded wiring to test the wiring.

[73sport]

^^ True statement it will not tell you its good. It just confirms the ECM is not getting a signal which causes bad input (a short).

In all I've read still no answer as to why it runs so poorly making little power. Best I recall in this thread the k/s is suspect, but nothing indicates its actaully bad.

I come back to this:

Originally Posted by 73sport

-
Question, do you hear the engine knock or ping? Have you checked the ignition timing? Changed gas station quality?

Things that lead to detonation; fouled and carboned plugs, running hot, excessive ignition timing, running lean, bad gas. Are any of these other things true?
-

The wiring is still suspect. He measured "open" but that doens't mean they are good or bad. It could still be the wires, the sensor(s), or something else.

[93celicaconv]

Without a knock sensor input to the ECM, the ECM goes into an open loop (failsafe status) regarding engine timing. Timing is retarded signficantly to ensure no chance of ever reaching a predetonation situation in an unmonitored engine. With such retarded timing, fuel efficiency is sacrificed to ensure the engine is not damaged. The CEL being on, with the codes saying it knows a knock sensor (or in this case, both), is telling us the ECM knows the knock sensor(s) are bad, or the wiring between the knock sensors and the ECM are bad. Doesn't matter which, the ECM is now just trying to prevent engine damage.