Proper procedure for removing spark due to knock

[Protec]

Gents,

Have a 2009 L9H / LS3 truck motor in a 11' JK

Having some slight knock retard issues when lugging the motor either on highway or back roads. Almost always when the truck is in 6th gear turning along and maintaining speed with throttle input without down shift.

Seeing 1-8* at varying conditions 1200-2000 rpm & 0.32 -.56 Grams/cyl


Fueling is good - Showing .96 - 1.0 Lambda at all ranges - Zero KR when "on the throttle" or WOT



So My Question.

I understand that you should look to take the highest knock count for the specific cell and divide by 2 and subtract that amount from the spark table and see how that effects the KR..

My question is do you blend the surrounding cells a bit or if they are not effected OR leave them alone? I didn't know if its ok to have a 4* difference in one cell with full timing value on the adjacent??? Kinda gives you the up and down scenario.

I would think that with normal driving you would be all over the table anyway and never transitioning across or through adjacent cell in a linear fashion negating the need for smoothing, but I want your opinion and expertise.





I have tested this by decreasing spark and eliminating the KR in some areas so I believe most of this is actual knock along with some drive train noise..


Respectfully,


John

[nonnieselman]

yea thats what i do.. subtract half of the KR and smooth the cells.. works good so far

[swingtan]

It's probably more involved than just adjusting the spark timing. I'm going to make a few guesses here, based on your post...

1. KR is predominantly when the engine is under load.
2. RPM is low as you're holding 6th.

And a couple if points before we get into things....

• Fueling is just as important as spark timing.
• Spark timing cells are "blended" so a step of 4' will not cause a sudden step, but a fast rise. This might be an issue, but in the normal cruising cells, there shouldn't be steep rises like this.
• If this is an engine transplant from one vehicle to another, and there are a few modifications to the motor, then the knock sensors are not going to be tuned to the new environment.

So, here's what I'd do.

1. Start with fueling. Get it right:
   Make sure the trims are good in all parts of the tune ( if running CL ). Having the trims working too hard at light throttle will mess up your fueling during PE mode. The LTFT's remain active during PE mode so you need to ensure the fueling is correct through the whole range of operation. Looking at your post, you're only going to 0.96 Lambda which is very lean for PE mode. If you're concerned with fuel economy, then I'd try starting at commanding 0.88 Lambda at low RPM and richening up to 0.85 around peak torque. This will help control knock.
   
   Talking about PE, I don't like the std. GM programing for the enablement of PE. In all the stock tunes I've seen, GM rely solely on throttle position to trigger PE. You need to have 70% or more throttle before PE mode will kick in. But you can labor a motor with much less throttle, putting the engine in a condition where it could really use some more fuel. So I swap the settings around an run a very low throttle setting and trigger on the MAP. Anything from 75kPa and above works well.
2. Move on to spark:
   As mentioned, in the normal cruise cells there shouldn't be steep changes between cells. If you find this occurring, then there's probably something else going on. There are a few ways to fine tune the spark table, here are the common steps...
   
   1. reset the octane scalar: You need to start at a known point. Another trick here is to make the low table the same as the high table, the only issue you have with doing this is that you need to keep them in sync when making changes.
   2. Keep as many external conditions the same between runs. If you don't have access to a dyno, drive the same roads when climatic conditions are the same.
   3. Log everything..... Use the V2 in all it's capabilities. If you are on the road, set up audible alarms for KR so you know when it occurs. If you hear the alarm, take note of what was occurring just prior to the alarm and try to repeat the conditions ( helps identify false knock ). Also use the V2 function keys to flag any points of interest in the log file.
   4. Review the data and make changes. You can create a KR map and use it like when tuning the VVE ( just subtract the KR levels from the spark table ), or if there is not too many different points, you can manually highlight the instance of KR (Scantool) and look for the cells in the spark table (Tunetool) and subtract the KR values.
      When the table is getting close and the KR values are only 2' or 3', I usually remove a little more timing than appears as KR and also remove some timing from the surrounding cells. Then I smooth the local area.
   5. repeat as long as required.
3. Knock Detection:
   This is a tuneable function like the rest of the engine, but it's pretty hard to get right without the use of a dyno and "knock cans" to listen for real knock. From the factory, the sensors might be a little too sensitive, but it's hard to tell if you don't have the right gear. You "could" desensitize then a little and it'll help, or you may just no longer detect KR....

Finally, the "correct" way to tune the spark table would be to put the car on the dyno and log every load cell, matching the dyno calculated power at each cell. You then make changes to the spark table and retest, noting where the dyno calculated power increased or decreased. Your aim is to get the highest dyno power in every cell while running the lowest possible timing. With no RoadRunner for the E38, this will take a bit of time, but is the only way to really know how the spark table should be.

Simon.

[Protec]

Simon,

Very much apreciate the detailed response. I've sent it to evernote to add to my knowledgebase. Again - Thank you!

To start to drill down and get more specific on a few points the .96-1.0L was at cruise, not WOT/PE... My apologies for not making that more clear. I see .85-.90 at peak.... Your point about PE and cruise number (.88) is well taken and something I can work on.

I thank you for the detail on your method of spark tuning and will incorporate that into my calibrating education.

I want to learn to do this on the road, but when I'm comfortable with the software and procedures I will take this to the dyno.

Appreciate your time-


Best,

John-

[IJ.]

Thread mining!

Is there a way using a Knock Retard Data Table to Paste the correction in at 50%?

[joecar]

You mean create a map of knock retard that matches the axes of your HO spark table...?

Yes (even tho it might not be the correct way to do it... you have to first determine what is causing it to knock):
- in tunetool goto spark table, highlight the whole table, do copy-with-labels;
- in scantool goto Map tab, select an unused map, edit its properties, on the Row and Col tabs click Paste with labels,
- then on Data tab select the knock retard pid,
- on the Data, Row, Col tabs checkmark Show Units,
- save map with a filename.


Once you populate the map with data, you might apply a filter (depends) and then copy the map, and in the HO spark table you paste-subtract-with-labels.

[IJ.]

Thanks Joe