Mitsubishi Lancer Evolution Forum banner

1 - 19 of 19 Posts

·
Registered
Joined
·
303 Posts
Discussion Starter · #1 ·
I've looked around and I can't figure out what "engine knocking" is. I'm sure its simple but if someone could answer this question I would appreciate it!
 

·
Registered
Joined
·
303 Posts
Discussion Starter · #3 ·
Thanks, I looked into it and I understand how knocking happens. I don't think my car "knocks" but I just wanted to know what to listen for. I looked on youtube and from the sound quality it didn't sound "pingy" like wikipedia said. I just want to know what to listen for and how to fix it if the problem ever occurs.
 

·
Registered
Joined
·
1,103 Posts
Knock and ping are old car guy terms describing ignition irregularities. It's a common misconception, but in internal combustion theory (4 stroke), the air fuel mixture does not ignite and explode. If it did, you would have uncontrolled combustion, which is what is referred to as knock.

You want the air fuel mixture to burn completely evenly, imagine a field of grass. The flame front travels across the field, it doesn't all go up and once. That's what produces power by creating pressure that acts on the top of the piston.

Knock is really post ignition, which can be anything ranging from the failure of the ignition system to ignite the mixture and an alternative ignition point spontaneously forming to multiple competing flame fronts hitting each other head on in different directions. This is defined as after the point at which the spark plug was supposed to fire the mixture. When it gets super volatile, it just seems to completely explode at the same time.

Pre-ignition, which is called pinging, occurs when there is extremely high pressures caused by lean conditions, too much boost, etc, at a point before the ignition system tries to fire the mixture.

Both knock and ping are detected by a microphone (the knock sensor) which listens for high pitched noises that are associated with both knocking and pinging. The sensor has a tight threshold, so it is easily fooled by noises that are not associated with post and pre-ignition. It'd really an inexact sensing method, but short of putting $5000 pressure transducers inside the actual cylinder it's the best method for detecting problems.
 

·
Registered
Joined
·
1,103 Posts
Bah internets.
 

·
Registered
Joined
·
5,594 Posts
i thought it was where your engine jumps out of the car and knocks on your front door to tell you "come drive me"

lol...
 

·
Registered
Joined
·
20 Posts
Knock and ping are old car guy terms describing ignition irregularities. It's a common misconception, but in internal combustion theory (4 stroke), the air fuel mixture does not ignite and explode. If it did, you would have uncontrolled combustion, which is what is referred to as knock.

You want the air fuel mixture to burn completely evenly, imagine a field of grass. The flame front travels across the field, it doesn't all go up and once. That's what produces power by creating pressure that acts on the top of the piston.

Knock is really post ignition, which can be anything ranging from the failure of the ignition system to ignite the mixture and an alternative ignition point spontaneously forming to multiple competing flame fronts hitting each other head on in different directions. This is defined as after the point at which the spark plug was supposed to fire the mixture. When it gets super volatile, it just seems to completely explode at the same time.

Pre-ignition, which is called pinging, occurs when there is extremely high pressures caused by lean conditions, too much boost, etc, at a point before the ignition system tries to fire the mixture.

Both knock and ping are detected by a microphone (the knock sensor) which listens for high pitched noises that are associated with both knocking and pinging. The sensor has a tight threshold, so it is easily fooled by noises that are not associated with post and pre-ignition. It'd really an inexact sensing method, but short of putting $5000 pressure transducers inside the actual cylinder it's the best method for detecting problems.

Its nice to see real answers and not some cocky ass remark! I though this was a better forum?

Well said discogodfather!
 

·
Registered
Joined
·
2,657 Posts
it will sound like a "tink" rather than a ping.

like tapping the edge of a coin on a steel table top.

with the aluminium block that the evox has, its pretty easy to hear (unless you have a 12" exhaust/woofer)
 

·
Registered
Joined
·
1,038 Posts
first time I ever heard it on my first turbo car it sounded like "a box full of marbles and gravel being shaken"

I know that's pretty loose in description, but as one of the early responses said, actually detecting it with 100% accuracy requires equipment too expensive to run on the car. Going to a good tuner, they will uses the forces of witchcraft and wizardry to extract more power without pushing too much timing or boost and getting the peak cylinder pressures too high for your fuel accommodations.

how much that's acceptable will usually change per model of car. When I dyno'd my X (100% stock) my first run I had 12 counts (insanely too much). My tuner was datalogging and also guessed it was from the factory tune overfueling and because of the heat soak I had. My second run had 1 count. He said that was totally fine for the X.
 

·
Registered
Joined
·
2,657 Posts
so when looking at datalogs etc. how much knock is acceptable? at what point to you consider adjusting the timing or tune?
depends what software your using,

KnockSUM is what the ECU see's, anything above 3 is bad, but only if its consistently above 3, and above 110load...

KnockRetard (i think Cobb calls it) is just KnockSUM * 0.35, so in this case anything consistently above 1 is bad (same 110load as well)
 

·
Registered
Joined
·
303 Posts
Discussion Starter · #14 ·
Thanks for all the info. I hooked my AP up last night and went on a 20 minute drive, going WOT a couple times and monitored the "knock". On the log the "knock" was always at 0, so it looks like I'm good. I listened for any strange noises too, everything seemed normal.
 

·
Registered
Joined
·
1,038 Posts
yeah I think you're more likely to just hear the Evo rattle than knock anyhow. I mean it is a Mitsu :p
 

·
Registered
Joined
·
1,103 Posts
Evoscan on my MR (bone stock) picks up tons of clutch rattle and calls it knock, up to about 10 counts. Hard to believe because it's happening at 1 psi and 600 rpm, so mechanical noise and other problems can cause heart attacks and false alarms.

Like Tephra said, look for consistent knocksums at higher boosts and loads to have a legitimate heart attack.

This is a video of combustion in a coke bottle, not really a cylinder, this video shows a good flame front:


And this is when it has some uncontrolled problems, similar to what would happen with knock:

 

·
[NEEvo President]
Joined
·
15,153 Posts
depends what software your using,

KnockSUM is what the ECU see's, anything above 3 is bad, but only if its consistently above 3, and above 110load...

KnockRetard (i think Cobb calls it) is just KnockSUM * 0.35, so in this case anything consistently above 1 is bad (same 110load as well)
Didn't one of the last AP updates change from KnockSUM to Knock Count so you don't have to *.35?
 

·
aka anEVOx
Joined
·
356 Posts
how much that's acceptable will usually change per model of car. When I dyno'd my X (100% stock) my first run I had 12 counts (insanely too much). My tuner was datalogging and also guessed it was from the factory tune overfueling and because of the heat soak I had. My second run had 1 count. He said that was totally fine for the X.
Yup, when I did my baseline pull my knock count was maxed out on the stock tune. I think mine had alot to do with the lean spool and timing.
 

·
Registered
Joined
·
14,174 Posts
As Teph said, up to 1.1 on the Cobb AP is probably alright... I like to err on the side of caution and have my tuner adjust my maps so I sit at flat 0.0 with the occasional 0.4

I'd rather give up a few ponies than give up $10k.
 
1 - 19 of 19 Posts
Top