Spark Retard - When it actually happens?

[vetteboy86]

I'm working on a tune for my setup. My work lately has involved eliminating spark retard.

I'm working on my timing curve, and at the point where I see the ECM start retarding timing, the knock event must occur before that, right? I believe it must take some time for the ECM to identify and take action. My scanning software records every 100 ms or so.

Any information would be very helpful.

 

[vetteoz]

the knock event must occur before that, right? .

If you are logging you should see the knock counts. On mine the spark retard as programmed happens in the same recorded frame , at the same time the KC appear ; no time delay.

 

[vetteboy86]

If a knock count is what triggers knock retard, I still have to believe that there must be a delay. I'm wondering if a knock count is recorded at the point where ECM retards timing?

Does that make sense? Perhaps I'm flawed in my thinking. It must take some time for the ECM to detect the signal and react to it. I'm trying pin point the exact moment that the spark knock occurs.

 

[Schrade]

If a knock count is what triggers knock retard, I still have to believe that there must be a delay. I'm wondering if a knock count is recorded at the point where ECM retards timing?

Does that make sense? Perhaps I'm flawed in my thinking. It must take some time for the ECM to detect the signal and react to it. I'm trying pin point the exact moment that the spark knock occurs.

I know what you're asking, and oz is right - the delay/response time is almost 0sec timewise. One test for proper KS function is to tap the block with a screwdriver next to the sensor. A properly operating sensor will make the motor stumble right away, without any 'delay'.

 

[vetteoz]

- the delay/response time is almost 0sec timewise.

From one of my datalogs.
Black is knock counts , Red is spark retard.
Notice both start from the same point , no delay.

 

[vetteboy86]

So should I subtract timing from the point at it occurs, or a few records before that?

 

[vetteoz]

What program are you using?
In TunerPro you can bring up the 3D chart of spark advance giving SA v RPM v Load.
You can just pull the advance down in the areas where you are getting KC.
Not on the log above , but I was only getting KC going through 3000 WOT (hi load ) .I just flattened out the SA in that area which is easier than punchng in numbers in all the relevent cells.

 

[vetteboy86]

I have not had the car on a dyno yet. I'm trying to get what I can done. I have a wide band installed in the car. My AF ratio is decent, but I want to correct the knock retard before I start hell bent on tuning AF ratio.

So my issue is that I am getting spark knock down low, probably where the car makes max torque, right around 2400-2800 RPM's.

So is it just simple that I start a few hundred before, and a few hundred after, grab that section and subtract. What I imagine is starting at 2200, take 1 or 2 degrees, than at 2400 take 3 or 4 degrees, than at 2600 give 2 degrees back, than at 2800 give 2 degrees back. So it looks like an even trough around 2600.

Now this may not be what I do, but in my mind it is somewhat what I think should happen.

I'm using tuner pro for modifying the bin, and datamaster to scan. The software I am using is the free version of tuner pro. I have been meaning to donate, or to buy the other one. So perhaps it could be that I can just grab what I need.