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

[mavisky]

On the contrary. Timing is pulled immediately once knock is detected.

I thought that was the case for the 2g ecu, but not the 1g ecu. Perhaps I've been wrong.

 

[dsm-onster]

I was under that impression for a 1g too. I know for a 2g that the knock detector is ignored under 2500rpms or so. I've logged with dsmlink and saw no knock retard. Then went to my scanmaster and saw 1-2 counts. randomly. Then went back to dsmlink and still saw nothing for several pulls. Not that this is a guarantee that the sensor was just not detecting knock when running the link logger. But, I was confused until I researched and found guys were saying this. . .

TMO Datalogger 101 . . . Scroll down 2/3s of the way.

 

[wishihadatalon]

No just a fast 4g63.... I wish I could see more people stretch the stock shortblocks for as long as some people can. But they just blow them up because of their horrible tuning abilities. I bought that motor @ 60K and when I sold it last month I had 110K on it. Still runs like a champ in another locals car...

I never had a problem with phantom knock, I get 0 knock when I have my tune perfect.. If I did not run out of fuel pump(walboro 255 not the high pressure version)and injector. The car would have done 132+mph. But thats in the past. I have a 2.4L + t67 now. I'm just saying Not everybodys problems are a knock sensor. I mean This is tuners for crying out loud not NABR. I can guarantee your pretty much in a .001% catagory that runs Haltek here.. And you may get like 5-10 guys here than run AEM and thats it too... So the mitsu knock sensor will save 90% of this boards motors if you follow me...

13.5:1 just seems really lean for any boosted motor. Not knocking it if it worked, but there is no real harm in running it a little on the rich side, and not much power to be gained by leaning it out a whole bunch. Everything I have ever been told is that if you run it a little rich then you will always have enough fuel for all of the air in the cylinder, thus taking advantage of potential power.

 

[mavisky]

I was under that impression for a 1g too. I know for a 2g that the knock detector is ignored under 2500rpms or so. I've logged with dsmlink and saw no knock retard. Then went to my scanmaster and saw 1-2 counts. randomly. Then went back to dsmlink and still saw nothing for several pulls. Not that this is a guarantee that the sensor was just not detecting knock when running the link logger. But, I was confused until I researched and found guys were saying this. . .

TMO Datalogger 101 . . . Scroll down 2/3s of the way.

Thanks for finding that info Matt.

 

[Black95TSIawd]

I was under that impression for a 1g too. I know for a 2g that the knock detector is ignored under 2500rpms or so. I've logged with dsmlink and saw no knock retard. Then went to my scanmaster and saw 1-2 counts. randomly. Then went back to dsmlink and still saw nothing for several pulls. Not that this is a guarantee that the sensor was just not detecting knock when running the link logger. But, I was confused until I researched and found guys were saying this. . .

TMO Datalogger 101 . . . Scroll down 2/3s of the way.

There was a large thread on the link forums in regards to this same situation. One of the creators of Dsmlink even contacted Todd and discussed this with him. It seems that over the years, information became misinformation became believed information. If you can comprehend what i just typed.

 

[architechnik]

I would have to say with a 1g gsx at 110k miles on it I'll pull 15-20 phantom knock while shifting above 4.5k rpms into 2.5k read off a pocketlogger. I'm still running stock and it's all just vibrations but it kills on the advance drop-off. It wouldn't be so bad if it took into account that the vehicle wasn't in gear!

Does anyone have a link that would show the variations between timing pull for the different year ECU's? I've modified my stock EPROM for the current settings, but if it can be done we should be able to compare those specific knock ratings for retard correct?

There is another thread headed up right now about someone creating an adjustable setter to tune properly. I don't know what the advantages of having the double sensor setup they are suggesting, as it doesn't seem to matter if it's coming off 2-3 or 1-4.

I would have to second that, "warmheartedly" on running 13.5 seeming real lean for a pull on hard boost.

 

[pneumo]

That mis information came about when the section in the ecu that handles 'octane' was mis interpreted as a section that handles 'knock', since parts of the code for knock and octane do similar things.

Thanks to the good work of the members of the DSM-ECU list the differences between knock and octane are easier to see. It turns out knock will retard timing startng with one count. It takes about 2.83 counts of knock to pull one degree of timing. So one count of knock pulls .35 degree timing, or about 1/3.

Octane is the long term trim applied to ignition timing. If knock is below 3, then octane is increased. If knock is between 3-7, octane is held steady. If knock goes over 7, octane is decreased, which will reduce timing all the time, even when there isn't any knock. That's why the 'Octane Reset' feature is popular in eprom chips.

 

[turboglenn]

That's some good info in the post above, I'm just glad i don't' have to deal with all that "crap" anymore. The 2G ECU was just a PITA when you can only get pseudo-91 octane around here. And i don't' even want to know what a barrel of 118 runs now days. I use to get it at 3.00 a gallon and would buy 55 gallon drums and run off it all summer, but th guy who use to sell it to me at that price retired from the petroleum world and now I'd probably be paying 5.25 or more a gallon, so I'm going to just add meth injection to my wanna-be-91 and see how far i can go on it.

I really would like to switch to E85 but my car is hard as hell to get started in the cold (15* and below) as it is now, and those are normal temps around here for half the year Mainly because you only get one chance to tune the first start, after it's fired up once and there's some heat transfer you have to wait till it cools all the way down and come back to it, i guess i haven't had the time or drive to be out there in the cold that often, i usually start it and come back inside for 20-30 minutes when it's that cold. Today we hit 30* woohoooo I'm going motorcycle riding here in a bit... Now can you see how cold it normally is here, when I'm nuts enough to ride a motorcycle at 30* because it's a warm spell for us

 

[diambo4life]

Thanks for finding that info Matt.

That is actually a long-standing DSM myth. Knock count is knock retard just expressed differently. If you have knock counts, you have knock retard. The whole idea of timing advance only getting pulled over 7 counts etc is rubbish. The author of that article didn't exactly explain that concept very well.

I personally have my knock sensor mounted very loosely on my block. It's nowhere near tight and I am comfortable with it's sensitivity.

Also, on my old setup, my knock sensor was only active from 5,500rpm+ ; Basically, when the engine was under full boost.

 

[turboglenn]

haha great pic, thanks hack!

I just hope i don't do that on the dyno here in about 12 hours

I'm really hoping for some good numbers, and i think this thread is biting me in the ass because i'm about to trust a high PSI tune to alcohol and pump gas, and now i'm worried the MSD unit isn't going to detect something tomorow. But then again, i don't know why i am worrying, it's the same motor with new cams and just an alky system added, it hsould have less heat from less restriction in the intake with the new cams and teh alky should help stave off detonation ( I am hoping for nough to make 425 street driven pump gas horsies. Of course tonight on the way to borrow a working camcorder to tape teh runs i was breaking th etires loose so hard in 3rd that teh entire front end was changing lanes on it's own on the interstate. Even 4th gear was torque steering so bad i could barely keep it on the road without some decent concentration. That's a pretty big boost for my car, i've been driving it at ~360WHP for so many years that it feels like nothing to me, so i know i have to be near 400 right now, i just hoep this time my tires don't slip on the rollers at teh dyno, last time we actually put traction compound on my tires and had 2 guys sitting on the strut towers to keep from spinning, and the cams have unlocked a TON more power

The only thing that has me worried is that i've tuned enough 4g63's on C16 and better to know exactly what that motor wants with that gas at high boost, i have no clue what it wants on pump gas + alky at those same boost levels muhahahhah...Should be a fun day~!

 

[dnhieu]

I use dsmlink and tuned for 0 knock on all my runs. This was on my old stock 6 bolt bottom and .63 a/r 35R setup.. Full weight 2g doing 128-129mph @ 28-29 psi and 23 degrees advance with 13.5 AFRs 0 knock 400+ track passes hundreds of steet races and daily driven since it was put together a couple years ago. I've even had the wastegate line come undone 2-3 times logging boost to 42psi and 12-15 counts of knock on one gear pulls. I just sold this motor right before christmas with this exact setup to a local guy and its already up and running fast.

I really don't see the problem in tuning for 0 knock?

No just a fast 4g63.... I wish I could see more people stretch the stock shortblocks for as long as some people can. But they just blow them up because of their horrible tuning abilities. I bought that motor @ 60K and when I sold it last month I had 110K on it. Still runs like a champ in another locals car...

I never had a problem with phantom knock, I get 0 knock when I have my tune perfect.. If I did not run out of fuel pump(walboro 255 not the high pressure version)and injector. The car would have done 132+mph. But thats in the past. I have a 2.4L + t67 now. I'm just saying Not everybodys problems are a knock sensor. I mean This is tuners for crying out loud not NABR. I can guarantee your pretty much in a .001% catagory that runs Haltek here.. And you may get like 5-10 guys here than run AEM and thats it too... So the mitsu knock sensor will save 90% of this boards motors if you follow me...

That is actually a long-standing DSM myth. Knock count is knock retard just expressed differently. If you have knock counts, you have knock retard. The whole idea of timing advance only getting pulled over 7 counts etc is rubbish. The author of that article didn't exactly explain that concept very well.

I personally have my knock sensor mounted very loosely on my block. It's nowhere near tight and I am comfortable with it's sensitivity.

Also, on my old setup, my knock sensor was only active from 5,500rpm+ ; Basically, when the engine was under full boost.

i cannot agree with these people more. i have basically disconnect my knock sensor and tune around engine noise, afrs, plugs and timing. i stick with a basic safe tune with no signs of auditable knock nor signs on my plugs. i have been doing this for over a year now at 23psi, meth, STOCK block, STOCK headstuds, and STOCK headgasket. my car see's 23psi and 8k everyday.

im not telling the everyone on the boards to go out and do this but i have found it takes alot to blow up a well built 4g63.

 

[absolute_DSM]

Even 4th gear was torque steering so bad i could barely keep it on the road without some decent concentration.

Anyone else I would have laughed at, but I actually believe you. Thats some good FWHP! Your dyno results should be very rewarding.

i cannot agree with these people more. i have basically disconnect my knock sensor and tune around engine noise, afrs, plugs and timing. i stick with a basic safe tune with no signs of auditable knock nor signs on my plugs. i have been doing this for over a year now at 23psi, meth, STOCK block, STOCK headstuds, and STOCK headgasket. my car see's 23psi and 8k everyday.

im not telling the everyone on the boards to go out and do this but i have found it takes alot to blow up a well built 4g63.

Thats about what I believe too. Its pretty much all in the tuning.

 

[dnhieu]

, my really good buddy Hutch would break loose in 4th with his old gst. leaving me like WTF

 

[Black_Bullet]

Interesting thread...

I dont worry too much about knock counts, but i just tune mines to were i dont see any significant timing getting pulled back on my logs. If its pulling back timing, whether real or fake knock, your still not getting all your power. Might as well just set it to were its not yanking timing dramatically...

I use a knock buzzer that is trigger by the ecu at 10 counts of knock and above.

The knock buzzer has proven very reliable for me in the past, and i base most of my tuning off of it when doing tuning pulls, because I dont think under 10 counts of knock is enough to harm anything.
The stock knock sensor is so inconsistent; due to phantom knock, gear, temp, etc. that its pointless for me to get a good tune for zero counts based on the logger alone without making my tune way to conservative.

As precaution; (especially for the days i dont have my laptop logging screen in front of me) and my tune on the edge of what my knock sensor considers 10counts. I can let off if theres any "possible" problems at the sound of the buzzer because then it will be over the limit of what i consider safe just in case the knock was real.
So in the mean time i just ignore the random 4-6 count blips on the laptop screen and just wait for the buzzer...

 

[steve]

I use a knock buzzer that is trigger by the ecu at 10 counts of knock and above.

...

(It) Has proven very reliable for me in the past, and i base most of my tuning off of it when doing tuning pulls just because the stock knock sensor is so inconsistent.

So how does using a buzzer that is based off the readings from the stock knock sensor get around the "inconsistent" stock knock sensor?

 

[soldave]

Did he say it was definitely based on the stock knock sensor?

 

[Black_Bullet]

So how does using a buzzer that is based off the readings from the stock knock sensor get around the "inconsistent" stock knock sensor?

No it is based off the stock knock sensor,
i typed out what i ment to say in my post wrong.
I was speaking in regards to me saying that i wasnt worried about a lot of my knock counts ( logger wise )
and that i dial in my tune more so with relying on the knock buzzer.
My knock counts on the logger are inconsistent at times due to phantom knock or etc.
As long as it doesnt go off ( 10 counts +) I dont worry about the tune so much.
It will randomly like to throw out 4-6 counts and sometimes no counts at all.
Therefore i cant get a good tuned based off of logs only since its random.
If i just try to tune with my logger while aiming for zero counts all the time id have to make the tune
way too conservative. So under 10 counts i dont worry about it too much since i consider half of them being possibly fake knock.

Tuning this way, even if it bleeped out some knock on the logger here and there;
the timing still doesnt ever really seem to be affected and i get the full timing i dial in on the chip. If i ever hear the buzzer then it will definitly be pulling timing on my logs...

I edited my post up above.

 

[pneumo]

If you have a chip, then knock can be desensitised with phantom knock code and knock decay rate adjustments.

 

[soldave]

Make that "If you have the ability to emualte or burn a chip". I have an Ostrich emulator but finding the knock thresholds and things like that on an early Evo chip isn't very easy. I don't know if many people know where it lies in the code of an Evo .bin.

 

[pneumo]

There's been some work done with the 952 bin since it's similar to the usdm E931. What's the number on the EVO1 chip?

 

[Black_Bullet]

If you have a chip, then knock can be desensitised with phantom knock code and knock decay rate adjustments.

So far all i know of is the octane value reset and the ignore phantom knock code
which i have both... I ignore knock up to like 3200 rpms.

 

[Joesmoke]

After battling with my KS. I grounded it. Why? I know the car ran rich, I run straight meth injection and a big IC. Actually, I was still running the stock smic and 14b when I pulled it.

The other reason I pulled it was I know my car, and was in for some experimentation. If it blew up so be it.

So, with the logger counting timing in the mid twenties, and boost at about 20 PSI (on the 14b) the car ran like a scared cat!

I ran like that as a daily commuter for 12 months with no problems. Had a failure in the meth pump once, and had some very light auduble detonation for a split second before I backed out of it. No harm done.

The latest incarnation is a 20g and big fmic. I did lose the HG once, but that was with run away boost (30 psi +) and there was no damage done (so no knock) Pistons fine, all fine. the HG was a stock composite and it was obvious that it just gave way.

So, Ill continue to pay attention to the car, and enjoy no KS. I have no room for error if I lose the meth, but thats fine with me.

My 2 cents.

edit- I should say though that once I have a way of tailoring a knock curve (standalone) I would utilize that.

 

[turboglenn]

After battling with my KS. I grounded it. Why? I know the car ran rich, I run straight meth injection and a big IC. Actually, I was still running the stock smic and 14b when I pulled it.

The other reason I pulled it was I know my car, and was in for some experimentation. If it blew up so be it.

So, with the logger counting timing in the mid twenties, and boost at about 20 PSI (on the 14b) the car ran like a scared cat!

I ran like that as a daily commuter for 12 months with no problems. Had a failure in the meth pump once, and had some very light auduble detonation for a split second before I backed out of it. No harm done.

The latest incarnation is a 20g and big fmic. I did lose the HG once, but that was with run away boost (30 psi +) and there was no damage done (so no knock) Pistons fine, all fine. the HG was a stock composite and it was obvious that it just gave way.

So, Ill continue to pay attention to the car, and enjoy no KS. I have no room for error if I lose the meth, but thats fine with me.

My 2 cents.

edit- I should say though that once I have a way of tailoring a knock curve (standalone) I would utilize that.

That's the spirit !!!!!!!!!!!!! AS long as you dont get crazy with timing or lean AFR's and since you seem to know what your motor wants you'll continue to be fine i'm sure. I am confident without knock capabilities. but after getting My J&S i wouldn't give it up for anything...But i wouldn't run a stocker again either!

Being that it's tunable you can dial out the BS phantonm knock, noisy lifters etc... I mainly prefer it as a tuning tool for easy setup of the timing curve, but knowing there's something there should i run out of meth ( i am dependant just like you with a tune tht's built on a lot o fit )

I've been trying to help the owner with spreading into th eimport scene more and he's agreed to assemble of few of the boxes to fit our cars in a test market so to speak and can have them ready on short notice

PM if you're wanting more info on them or interested in one for our cars specifically

 

[Joesmoke]

Ahh, refreshing not to get a "The sky is falling because you have no KS!!" reply.

It comes down to common sense (of the internal combustion engine kind)

The J&S system was high on my list of priorties, at one time. Did a lot of looking at it and really liked what I saw. Mostly its ability to retard timing on a per cylinder basis. Thats its true value.

Then I couldnt decide to go piggyback on piggyback or standalone. Im leaning towards stand alone again, but who knows.

The KS is truly there for 98 percent of the people. I spent a lot of time geeking out to the concept of knock, and a lot of time making sure I wouldnt have a problem. More then the average guy is probably willing to do.

And meth, it does a Talon good

 

[JrCRXHF]

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.