Knock sensor detuning - OBDII

[V6'er]

I'm going to post this on the V6 forum because no other forum uses OBDII motors (that I can think of, and NO I'm not counting a V8).

I'd like some well-founded opinions on how much, percentage wise, you feel that the "over-sensitive" 1MZ knock sensors could be reduced by, and not do any engine damage.

People complain about it all the time, but who's fixing it? I already know how, it's too easy.

[Oteck]

atm i believe my swap is suffering from the .4v coming out of the 3mz sensors instead of the .8v the 1mz suppose to have. probably think it's dead but it hasn't given me a code.
Another person has swapped out using the 1mz one and worked out for him. I'm now thinking of doing it.

Are you thinking of jamming a resistor in there?

[V6'er]

The output impedance is surprisingly low. It takes about a 4K resistor as a load to drop about 20%.

But a better way to do it is a pot on each sensor. A 10K pot would barely load the sensor, and it's very easy to calculate the percentage of reduction using a DVM.

[jdchmiel]

they used to use a rubber gasket in the subaru world betweeen sensor and block to try and dampen the sound a bit.

I dunno excatly what you are trying to accomplish thoughm so ignore my comments if they dont fit at all. You can remove the knock sensor from the motor, ground it to the block somehow, and keep it from getting any vibrations by some sort of isolation mount. this obv removes all protection and will let the ecu advance as far as it is capable of.

you might be able to use an adjustable voltage clamp to remove any peaks from the signal to the ECU. this again will prob remove the safety net.

You could treat the knock sensor as a microphone of a known imedance ( use a multimeter) and build a passive crossover. you can look at the signal on a scope, see what frequency ranges are peaking a lot when you KNOW you have no knock, and use a notch filter to take jsut them out, or you can do a low or high pass filter or combinations of various filters to basically shape the signal that the ECU gets.

[V6'er]

^ Yes, I could do any and all of those things. (I am so inclined, I'm an EE.)


But, I'm also looking for some opinions on "how much", because ZERO is not an option. We have to remember that this is a 10.5:1 compression ratio motor - detonation can and will damage it with time.

And, I have it on VERY good authority (the guy that used to build the Toyota Celica GTS Celebrity Race cars) just what can be done to a 2ZZ or a 1MZ. This is a discussion for enlightening the rest of you as well as myself.

[d's nuts]

Im so stupid when it comes to stuff like this but what will you accomplish from tampering with the knock sensor?

[V6'er]

Only V6 guys that have driven for a while are going to know that the 1MZ takes a major performance hit with 87 octane, and even 93 octane isn't the final answer. This is because the knock sensors pull timing even with 93 octane on an 80 degree day.

I frankly don't think that this motor should be so sensitive to temperature, and it's the knock sensors that cause the hit to performance. Running 100 octane fuel does make this problem go away I'm told. All I want to do is make it run better, especially on a warm day.

[Mark_Dorman]

Subscribed...

[Jason.MZW20]

Originally Posted by V6'er

Only V6 guys that have driven for a while are going to know that the 1MZ takes a major performance hit with 87 octane, and even 93 octane isn't the final answer. This is because the knock sensors pull timing even with 93 octane on an 80 degree day.

I frankly don't think that this motor should be so sensitive to temperature, and it's the knock sensors that cause the hit to performance. Running 100 octane fuel does make this problem go away I'm told. All I want to do is make it run better, especially on a warm day.

This is a particularly interesting problem because it seems the 1MZ is overly sensitive in its timing retard. Mitch, do you believe this to be a deliberate action by Toyota, much like with the 3VZ, to reduce the performance of the 1MZ vs. the 2JZ? The reason I ask this is because the 1MZ's knock sensors are widely used by other major Toyota engines, including, IIRC, the 2JZ, 2AZ, all UZs, and most other xZ engines. It could also be that, as this is a Camry motor, Toyota has higher expectations for longetivity than with a performance oriented application like the 2JZ, so they aim for the 1MZ to see over 200k miles.

This begs the question: is it the actual knock sensor, or is it the 1MZ ECU's interpretation and action (timing retardation) of the knock sensor signals? Or could there be some external frequency (maybe a vibration creating similar harmonics) the 1MZ block is susceptible to that the other blocks are not, which would cause this apparent problem?

So, your solution focuses mainly on the output of each sensor, right? Would it be more efficient or beneficial to change the voltage the ECU sees at the ECU 'KNK' terminals (much like another 'black box' solution) and even make it user adjustable with around 3 settings?

[V6'er]

Jason,

The guy I talked to said that even the 2ZZ in the Celica GTS was quite susceptible to the issue on pump gas. (He's also very familiar with racing the 1MZ's.) They are in Southern CA, and they run the event in the spring, so how hot can it really be? He said that all cars were tested stock, and there was a great variation in how quick they were on 91 octane pump gas. They went to 100 octane and all cars finished the track in the same time.

So, yes, I'm sure it's the "interpretation" of the knock sensor signal, but that's not easy to change. Changing the input level of the signal is very easy, so let's start there. One 10k pot on each knock sensor leg would work fine, this makes it continuously adjustable. And it was told to me (same guy) that if the engine does NOT pull timing on 87 octane, I've lowered the signal too far.

Newbiewan changed his crossmember out to a '93 well after his SC'd 1MZ was running. Under acceleration, the exhaust would roll down and bang the right track rod, causing the ECU to drastically retard the timing. He moved the exhaust and his performance came back.

My setup has no cat and I'm using the stock Turbo muffler. It sounds good, if you don't mind raspy, and it's quite loud. I think that this extra exhaust vibration causes high output from the knock sensors and kills performance.

[Levi]

I read on the camry forums (Toyotanation.com/Solaraguy.org) that they relocate the 1mz knock sensor to a different location and it read much better. I never looked further into though =/

[jdchmiel]

what is your solution if you already know the answer? this thread doenst make much sense as it seemed like you asked a quesiton, but then stated you know the answer but never told the answer...

[Jason.MZW20]

Originally Posted by jdchmiel

what is your solution if you already know the answer? this thread doenst make much sense as it seemed like you asked a quesiton, but then stated you know the answer but never told the answer...

But he did ...

Originally Posted by V6'er

One 10k pot on each knock sensor leg would work fine, this makes it continuously adjustable. And it was told to me (same guy) that if the engine does NOT pull timing on 87 octane, I've lowered the signal too far.

[mr220v]

There was a guy on the other board with a 20v silvertop who ran for a long time on 87 octane. The silvertop runs 10.5:1 compression. It took 30k+ but finally he damaged his engine enough to warrant replacement.

Now, this was with the stock knock sensor attached. At 10.5:1, with a desensitized knock sensor, I would imagine the damage would be more severe and show up faster?

I'm thinking if you did something like this, you would never want to run 87.


BTW, the beams 3sge ecu is obd2. It's a somewhat different standard than what is found in this country, so usdm obd2 scanners don't work.

[Mark_Dorman]

I don't think anyone was suggesting running 87 in the 1mz. The car pulls timing on 91 octane...the performance hit I think is the main concern.

[mr220v]

^I don't know, I wouldn't put it past some people.

[Jason.MZW20]

Yeah, we definitely don't want the ECU to NOT pull timing on 87 octane, which is exactly what Mitch said. It's a little ridiculous that the 1MZ pulls timing on 91 octane during a mild 80 deg day. It takes 100 octane (almost the octane of 100% E85 ethanol) to cure it. That, to me, is unnecessary.

So, yes, we do want to dial back the knock sensors a little, but not to the point where they no longer retard timing on anything less than 91 octane. This is why Toyota received so much flak when they moved to the J1349 revised SAE rating system. Some suggested Toyota were overrating their engines. It's not the case. They most likely used the absolute best fuel for the engine, because when retested on 87 octane, the 1MZ dropped a massive 20hp (210hp to 190hp).

[V6'er]

^ What he said.


I don't own a Camry, but my money says that the fuel door does NOT say that Premium fuel is required. Which means that the 1MZ is supposed to run 87 Octane fuel all day long. It's the "everyman's" car, right?

On that assumption (verified by "my guy" at the race shop), Toyota expects a performance hit on 87 Octane. It's those of us (and them racers in California), and Toyota Japan, that know better.

No, I have no plans to run 87 Octane to tear up my motor, which as some of you know was brand new when I got it five years ago. It's a virgin 45k miles motor, never been opened up since the factory. I'm surely not in the mood to kill it.

What I'm getting at is that I find it hard to believe that it truly requires more than 93 octane to run right. (Yes, in NC we have 93 Octane, not 91 like so many other places in the US.) I think it's because of the "vibration" caused by the very open exhaust. Add that to an "overly sensitive" pair of knock sensors, and my car takes a hit in warm weather.