Knock Retard hell! Desperate for help

[GHuggins]

There are some rip offs running around I believe.

The data the Greg Banish put out in his first CD cured so many of my problems I had on a Hahn turbo kit LSJ Cobalt I was tuning. I learned a good lesson on injectors with that car and now only insist on ID's or data I have from Greg's CD. Chasing your tail sucks.

[GHuggins]

Read some of the prediction coefficient stuff in more detail. Some of the guys are disabling it all together. Has anyone in here with knock troubles tried to do this?

Do i literally just paste zero's directly into all of the coefficient tables? I'm tempted to try this along with my improved VE table in my "normal" tune and report back with knock results.

Haven't done anything like this for knock, but yes all you have to do is enter in "zero's" in ALL of the prediction coefficient tables... Sssnake hinted at a problem with this. I haven't seen anything wrong with it yet, but don't know for sure?

[ecotec88fiero]

Not to get off topic but it cracks me up the difference between tuning a turbo cobalt, whose owner is balks at $900 IDs, and a CTS V owner who willing spends the $900 for IDs.

[itsdaveonline]

Haven't done anything like this for knock, but yes all you have to do is enter in "zero's" in ALL of the prediction coefficient tables... Sssnake hinted at a problem with this. I haven't seen anything wrong with it yet, but don't know for sure?

I *think* he was suggesting that the various cyl air measurements were not matching due to the predictive coefficients skewing some one of the cyl air measurements and that possibly contributed to the fuel mixture going lean or the PCM perceiving knock. Don't want to put words in his mouth though.

Either way good to know I can just zero out all of the predictive coefficient tables. It's at least worth a shot to see if it improves anything.

[schpenxel]

I zeroed them out on mine as well

I always had more knock when using the MAF for whatever strange reason.. SD only has been smoother, more consistent, and less knock (I'm assuming the consistent AFR and less knock part go hand in hand)

[itsdaveonline]

I zeroed them out on mine as well

I always had more knock when using the MAF for whatever strange reason.. SD only has been smoother, more consistent, and less knock (I'm assuming the consistent AFR and less knock part go hand in hand)

Did you notice any other improvements or changes by taking predictive coefficients out of the picture?

[schpenxel]

I can't say that I noticed any other changes or improvements.

[sssnake]

I *think* he was suggesting that the various cyl air measurements were not matching due to the predictive coefficients skewing some one of the cyl air measurements and that possibly contributed to the fuel mixture going lean or the PCM perceiving knock. Don't want to put words in his mouth though

Exactly what I was saying. To be clear though you can't zero all of the predictive coefficients (at least I dont think you can). You CAN zero the MAP, TPS, and corrected values. This is what I am currently running with much better results than stock. Not completely there yet but getting much closer.

2016-01-06.png

[schpenxel]

I have them all set to 0, though I'm not using the MAF so not sure what it'd do if you're in mixed mode

[GHuggins]

Works fine with "ALL" tables zeroed out with MAF. You know I forgot to do this on the Z28 and had my "mixed mode" set to operate below 2800 rpms, so when I stomped it from 1000 rpms the air fuel went up to 20 right quick and in a hurry and would stay there until 2800 rpms I of course backed off throttle immediately during this, but wanted to see how long the "problem" was there. Once I saw it richen back up at 2800 rpms I knew what I had forgot to do Zeroed out all the tables and she went to the commanded 11 to 1 air fuel at 1000 rpms as designed. Important thing is to have the MAF and VE both dialed in fully.... If you don't have all the tables zeroed out it'll still lean things down, just not as bad - probably why you see improvements, but still no fix? Wound up with a perfect MAF curve and everything working as it should... Of course injection timing and everything goes into "that" equation........

[GHuggins]

Here's some pics from the Z28 tune with a Magnuson Blower on a LS7 motor, cammed, long tubes with no fueling problems other than the baro problem

I couldn't fit all of the prediction coefficient tables on the screen, but they're all (0's)

I've even attached the injection tables to illustrate something on the lines of what you'll wind up with if you use "my" method of dialing in injection timing. It is a little rich on cold starts. I'm going to need to pull about 3 or 4 percent from my open loop tables yet...

In the MAF curve where the red dot is located is where you'll typically see a huge bump if you leave the prediction coefficients in the equation when you put a blower and or cam into an otherwise originally naturally aspirated motor's tune...

Hope this helps you more...

[schpenxel]

Thanks for posting the inj tables, I'll compare them to mine tomorrow for curiosities sake. I got your PM by the way, will respond tomorrow. I found the package.. Will make sure it goes out tomorrow

[GHuggins]

Thanks for posting the inj tables, I'll compare them to mine tomorrow for curiosities sake. I got your PM by the way, will respond tomorrow. I found the package.. Will make sure it goes out tomorrow

OK, I was afraid it got lost in shipping Looking forward to trying that out

[schpenxel]

Here's some pics from the Z28 tune with a Magnuson Blower on a LS7 motor, cammed, long tubes with no fueling problems other than the baro problem

I couldn't fit all of the prediction coefficient tables on the screen, but they're all (0's)

I've even attached the injection tables to illustrate something on the lines of what you'll wind up with if you use "my" method of dialing in injection timing. It is a little rich on cold starts. I'm going to need to pull about 3 or 4 percent from my open loop tables yet...

In the MAF curve where the red dot is located is where you'll typically see a huge bump if you leave the prediction coefficients in the equation when you put a blower and or cam into an otherwise originally naturally aspirated motor's tune...

Hope this helps you more...

What injectors are you running in that one?

[GHuggins]

ID 850s...

[sssnake]

Alright guys the screenshot below is now indicative of the results I am seeing. The air models are agreeing pretty closely (I need to tweak VE some). I have very little burst knock (.4 degrees) and no real knock for a 122 Kpa MAP transient (peaked at about 5.5 lbs of boost - wet this morning). I did work the PE some as well as before it was oscillating into and out of PE. It now uses the CTS V2 values with the hot power enrich enable a little lower, the enrichment ramp in a little faster, and the pe delay rpm halved. This has been the best results I have seen so far for transients. Thanks to the folks in this thread as well as the folks in the other prediction coefficient thread as I would not have given these a second look otherwise.

2016-01-07.png

Before I make another run through the MAF and VE cals I wanted to try and get EOIT adjusted. I wanted to get my below 2500 rpm values taken care of... I setup an excel spreadsheet to graph the cam profile/timing events and do a few of the calcs and the data lookup (pic below). It all seems to work and make sense so far. EXCEPT for determining the injector pulsewidths to use. Idle is easy. However, how do you pick the values for when the car is in motion? I mean a different gear will give you VERY different results. Also, the load on the engine changes things dramatically as well (accelerating, going up a hill, etc.) If there is this much variance in the pulsewidth values then how does this algorithm effectively adjust the EOIT under all of these conditions?

2016-01-08 (3).png

[schpenxel]

Wow that is a really cool sheet. I love the way it graphs the cam profiles.. that makes it very visual.

[sssnake]

You are more than welcome to a copy but it is still pretty rough (not plug and play). It was just easier than flipping back and forth between the cam card and multiple places in the tune. Besides if I can see it I am much less likely to totally screw it up.

[schpenxel]

That would be great if you could send me a copy. I could make it myself but doesn't seem very smart to re-do what you've already done

[sssnake]

for whatever reason this forum is not letting me add an attachment to pms so here it is.

Cam Chart.xls

Like I said not plug and play yet...