I spent a good part of the day yesterday messing with the knock tables after exchanging a few emails with ace123 about ignition tuning. I haven'found much info about these tables, so I wasn't really sure whatt they did. Since I wasn't sure how the changes were going to impact my tune, I started out making very small changes just on the top load lines so I could easily verify the impact on my datalogs. To my surprise, reducing the values by a small amount caused an immediate drop in the knock retard values, which in turn caused an increase in timing advance. Maybe less surprising is that it caused some knock counts. I thought it would reduce the number of knock counts, but it seems to have only reduced the amount of ignition retard.

After this first test, I realized that I needed to manually retard my ignition timing quite a bit, so that I could adjust the knock sensitivity without causing more knock. I think I retarded it about 8 degrees across the whole RPM range. After doing this, I reduced the knock values more, and smoothed the curves out. Then I went for another drive, and found that knock retard was now at 0 everywhere and my IGN values were showing up to be exactly what they are in my timing map, which has never been the case before. Up until now, my IGN values were always 10+ degrees less in some spots than what were in my timing map. I knew some of this was due to the knock retard values, which were as high as 5-6 degrees. Since my IGN values were being retarded by more than just the 5-6 degrees of knock retard, I now think that there might be some other type of "dynameic" knock retard that is applied. Maybe there is a long term and short term knock retard, and only the long term knock retard is showing up in the datalogs???

If you look at the circled area of the knock table below, you'll see that there is a "spike" in sensitivity at 2800 RPMs. If you look at the circled area of the IGN curve below that, you'll see that there is a matching "dip" in timing advance. This dip doesn't seem to be caused by the knock retard, since it is pretty constant at that RPM range (bottom line).

 

Oh, I forgot the whole purpose of this thread was to ask some questions...

I'm not sure what to do next. What should my goal be for the knock tables, timing maps, and knock retard? I know that as long as my actual IGN values are the same, power output will be the same, but is there a benefit to tuning out the knock retard? Is there a benefit to keeping the knock sensitivity high, and just advance the hell out of my timing map to compensate for the timing that's being pulled out by knock retard?

My current thought is to get my ignition maps where I want them, tnen slowly increase the knock sensitivity back up until it starts pulling timing again, then reduce them slightly again. It seems like this would give me adequate protection against bad gas. Thoughts?

 

I'm going through the same thing right now Greg. I've got my AFR's where I want them but I'm still getting a lot of knock retard on the low cam - but no knock counts on either the high or the low. My tuner had reduced the knock sensitivity on the high cam pretty significantly because of some issues with my valves being rather loud.

So I increased my knock sensitivity, and finally got some knock counts on Cylinder 3 on the high cam - I then reduced the ignition by one degree on that cylinder and increased the knock sensitivity by 5% again on the high cam.

I plan to continue playing around with this on the high cam and WOT, once I get that figured out and I get my knock sensitivity back up near the stock levels - I'll be happy. Then I'll start on the low cam/partial throttle and try reducing the knock sensitivity - I'm not getting any knock counts - but I am getting knock retard there - which doesn't make a whole lot of sense but I figure I can reduce the sensitivity if I'm not getting knock counts without damaging the motor.

 

Wasn't Ace123.

As we saw Honda uses this knock table to self tune the ignition values. Each motor is different and won't get personal attention so this approach made the most sense.

I would say the best approach would be to set the knock values appropriately to only reduce timing when there is an actual issue.

 

The help file goes over what the knock sensitivity tables do. If you tune the ecu properly, the ecu by itself should rarely reduce ignition timing on its own (only when knock is actually occuring. This doesn't include ghost knock, which is a difficult thing to figure out)

 

A good thing to always remember is this

knock count vs knock retard vs knock threshold. Use these three properly together and you will have results

 

I think the missing part is how one table controls all three values.

Even on the stock tune it retards the timing!

 

Ok I've done some reading - the ECU is retarding timing based on an anticipated increase in load at low rpm's. Apparently this is completely normal (per the Cobb Tuning guys on the Evo's).

All my knock retarding is happening below 3500RPM and is always when I start to roll into the throttle - hold constant at 15% throttle - no knock retard. The second I start to increase to 20% or more - I get 5-7deg of knock retard. Doesn't matter what gear I'm in. This corresponds to an intake pressure change of 440mBar to 800mBar

 

Interesting. I never saw that much..

 

Below is my car with the stock tune (red) and my brother's totally stock '06 with the stock tune (blue). The top lines are the IGN value at WOT, and the lower lines are the knock retard at WOT. I'm surprised to see how friggin' high my brother's knock retard is compared to mine. My best guess is that mine is lower since I have 1 step colder plugs and a "better" collant mix (80% water).