Why is everyone so afraid of the mitsu knock sensor???

[Street Surgeon]

My trusty Autronic SM2 and I fear no Mitsubishi knock sensor!

 

[Black_Bullet]

A little knock when i tune i am cool with esp if it is just in a little area and does not happen all the time. (say knock count of 3 or so on a EVO)

but why knock scars me is because the Cylinder pressure is much higher with a knock then it will ever be with boost. So if a head gasket is going to fail or a head is going to lift in most cases it is because of knock not to much boost.

Gas motors head is not designed to be part diesel

but who knows that is just me.

True, as to why you see more blown head gaskets with stretched head studs than bent rods.

Most of the people who tune around the knock sensor only do it because thats all they know how to do as for tuning... Like me. I wouldnt tell a newbie to tune like this by any means. The factory knock sensor prevents the majority of us from blowing our motors on the regular. A lot of honda and other car circles whom dont have stock knock sensors tune with the methods mentioned in this thread and have a lot more engine failures even at much lower boost levels.

As far as were i stand, running without some type of knock sensor, id be more worried about some random fuel failure or bad gas causing knock than me tuning it wrong. I have times were im cruising on the highway at about 75-85mph and i will trigger some nasty knock ( my buzzer tells me to let off the gas.) due to my extended closed loop keeping the AFR lean while my turbo is already trying to throw out 10 psi at that rpm/ throttle position.

For more advanced users who arent "afaid" of blowing their motors, or those who run on race gas or e85 all the time, and on cars that arent daily driven and are just set up to perform, then running without any type of knock device is perfectly reasonable as long as you know what your doing.

These motors seem to be very durable and put up with a lot even when there 100k + in miles. But i do believe that any substantial knock after a period of time will quickly fatigue the motor and shorten its lifespand regardless of how well everythings built. I dont have time or money to rebuild my engine every six months, year or even every other year, so this is why daily driven cars shouldnt be pushed so hard.

Now if i only drove this thing 1 to 3 days a week and didnt rely on it to get me to work and school then id aim for a much more aggressive tune like you guys do.
Maybe with enough people running different knock metering devices, and different techniques in the future that will give a revolutionary push to get the majority of us away from the stock mitsu knock sensors and to using more advanced knock monitors such as the J&S unit .

 

[We're on Boost]

I'm thinking about disconnecting the knock sensor from the ecu, but I would still run it to my ECU+ and I would let that do the knock retard for me. It will also add fuel if I want it to. But what would I do with the knock sensor wire that comes out of the ecu? Just run it to ground? or do I need a resistor in there, or what?

Gary

 

[soldave]

If the knock sensor is not detected by the ECU, won't it flash up a CEL and retard a lot of timing from the engine?

 

[FonZ454]

Yes, you'll get a CEL and I believe you lose a lot of timing advance or have some fixed timing. You'd need a way to disable it or bypass it.

 

[turboglenn]

A little knock when i tune i am cool with esp if it is just in a little area and does not happen all the time. (say knock count of 3 or so on a EVO)

but why knock scars me is because the Cylinder pressure is much higher with a knock then it will ever be with boost. So if a head gasket is going to fail or a head is going to lift in most cases it is because of knock not to much boost.

Gas motors head is not designed to be part diesel

but who knows that is just me.

I feel about the same way and my car IS my daily transportation,. although i do have an old 89 full size bronco for winter use, i don't like to be caught dead sporting it around town in the summer, worn out tires, busted windshield etc... So i rely on the eclipse full heartedly

As stated before, I have not had a knock sensor in probably 9 years ( i do again now for about 5 months or so) and i've only popped one motor due to a lean condition on a borrowed UEGO.. I ran normal timing numbers of 16 - 18* through out the curve and have never worried about a "bad tank of gas" i know i get em, but i just figure as long as i'm daily driving it, feeling what my motor is doing everyday i'm cool with no sensor. Then again, i've got a lot of years of running motors hard and sometimes blowing them up under my belt, so i kinda knock what to feel and listen and look for...

My trusty Autronic SM2 and I fear no Mitsubishi knock sensor!

Glad to see another soldier against the knock warlords!

 

[We're on Boost]

If the knock sensor is not detected by the ECU, won't it flash up a CEL and retard a lot of timing from the engine?

Yes, you'll get a CEL and I believe you lose a lot of timing advance or have some fixed timing. You'd need a way to disable it or bypass it.

That's what I'm thinking. You might need to have the stock ecu connected to something that looks electrically like a knock sensor so it stays fat dumb and happy thinking everything is ok.
Another knock sensor maybe?

Gary

 

[soldave]

That's what I'm thinking. You might need to have the stock ecu connected to something that looks electrically like a knock sensor so it stays fat dumb and happy thinking everything is ok.
Another knock sensor maybe?

Gary

Another knock sensor will work. Get one that's zip-tied to the firewall and run it as normal to the ECU. Have the one that's actually in the engine running to your ECU+ (whatever that is.. am not sure so can't comment). Thought of that before.

Am hoping the meth kit I've ordered gets rid of the knock I've currently got from 5400rpm+

 

[Black_Bullet]

Am hoping the meth kit I've ordered gets rid of the knock I've currently got from 5400rpm+

If you chose the right meth setup and right nozzles to run i think you will have some fun

 

[spyder97gs]

I threw a CEL today when romping on it and got it checked and it came out to be the knock sensor. I just ran a tank of 87 octane through it to see how it would perform versus premium. They reset the light and asked if I was having any fuel issues. Sound like the low octane caused the light?

 

[pneumo]

Thanks to this thread I found a new tool to deal with knock.
At the beginning of this thread I was thinking it would be nice to adjust knock sensitivity so knock doesn't pull so much timing at light load, but can still fully protect the engine at high load. I've played with some eprom code before, so I thought, how about writing a new section that subtracts a certain ammount of knock before it's used to remove timing? Something like 'load rawknock to Aa, sub X, etc...' I looked into the code and found the stock knock code already does this! Go figure, the mitsu engineers already thought of it.
;----------------------------------
; Compute a = (rawKnock-4)/8
;----------------------------------
ldaa knockSensor ; a = rawKnock
suba #$04 ; rawKnock -4 ( at m-f043)
bcc L1726 ; Branch if rawKnock-4 >= 0
clra ; Use min of 0
L1726 lsra ; divide by 2
lsra ; divide by 2
lsra ; a = (rawKnock-4)/8

Pretty cool, all I have to do is change the number '#$04' to anything higher and the knock will be less sensitive, especially at lower knock counts. My car just got a little faster!

 

[turboglenn]

Thanks to this thread I found a new tool to deal with knock.
At the beginning of this thread I was thinking it would be nice to adjust knock sensitivity so knock doesn't pull so much timing at light load, but can still fully protect the engine at high load. I've played with some eprom code before, so I thought, how about writing a new section that subtracts a certain ammount of knock before it's used to remove timing? Something like 'load rawknock to Aa, sub X, etc...' I looked into the code and found the stock knock code already does this! Go figure, the mitsu engineers already thought of it.
;----------------------------------
; Compute a = (rawKnock-4)/8
;----------------------------------
ldaa knockSensor ; a = rawKnock
suba #$04 ; rawKnock -4 ( at m-f043)
bcc L1726 ; Branch if rawKnock-4 >= 0
clra ; Use min of 0
L1726 lsra ; divide by 2
lsra ; divide by 2
lsra ; a = (rawKnock-4)/8

Pretty cool, all I have to do is change the number '#$04' to anything higher and the knock will be less sensitive, especially at lower knock counts. My car just got a little faster!

That's cool as hell, i only wish i could get a grasp on uning via eprom flashing and writing code!! I've always been amazed by guys that can decipher what all that jibberish means

 

[soldave]

Interesting. Any ideas what to look for in the hex code for that as I'd like to increase the knock threshold a little too.

 

[pneumo]

Interesting. Any ideas what to look for in the hex code for that as I'd like to increase the knock threshold a little too.

I had a peek at the evo I code and it's very different than the dsm. I'd start by finding the address where rawknock is stored. If you have a logger that can view knock, start by talking to the guys who code the logger and find the addresses they're logging. I got that idea from Ceddy. In the dsm e931 chip rawknock is at D1. If you're lucky it'll be the same. The piece of code goes like this: D7 D1 96 D1 80 04 24 01 4F 44 44 44. Maybe you can search for a similar piece in your code?

 

[soldave]

What is interesting is that the Pocketlogger made for DSMs will report knock and all the other values (just about) on my Evo 1. So it looks like the knock signal that goes to the diagnostic port at least is the same.

Hmmm... had a look at the Evo code and it doesn't seem to be in the same place. There doesn't seem to be a similar piece of code in there either May have to talk to the Pocketlogger people.