OpenFlash Tuning Tutorial: Crawford Billet Power Block Mapping changes

[steve99]

Originally Posted by Silver Ignition

So these two numbers are taken from the maps and applied to the FLKC for the final number that we're seeing in the Ignition Advance column? Am I in the area here?


This is so frustrating because I have a really good grasp on how engine performance works but this whole thing is driving me up a wall

Yes the two values you highlighted are taken from maps then you add the FLKC which is a negative value in this case .

then if you IAM was less than 1 you would have to use the full equation.

[Silver Ignition]

Originally Posted by steve99

Yes the two values you highlighted are taken from maps then you add the FLKC which is a negative value in this case .

then if you IAM was less than 1 you would have to use the full equation.

Gotcha so it's like a balancing act between Base Timing B and Knock Correction Max A.


So looking at this weird spike, should I pull another degree or two of timing? Issue is, looking at the chart, the Engine Load is pretty steady in the 1.1-1.25 range across the board but there's that random downward spike on the FLKC...so how do I tune this out?

[steve99]

Originally Posted by Silver Ignition

Gotcha so it's like a balancing act between Base Timing B and Knock Correction Max A.


So looking at this weird spike, should I pull another degree or two of timing? Issue is, looking at the chart, the Engine Load is pretty steady in the 1.1-1.25 range across the board but there's that random downward spike on the FLKC...so how do I tune this out?

All you do is look at the Engine load/RPM point where the knock is occurring.

In this case 1.2 to 1.3 load and RPM 6100-6700 and the ECU is pulling about 2 degrees so remove about half that ie about 1 degree or slightly more from the cells in that area ie load 1.2-1.3 and rpm 6100-6700 area out of the Base_Timing_B table.

[steve99]

Originally Posted by Silver Ignition

Gotcha so it's like a balancing act between Base Timing B and Knock Correction Max A.


So looking at this weird spike, should I pull another degree or two of timing? Issue is, looking at the chart, the Engine Load is pretty steady in the 1.1-1.25 range across the board but there's that random downward spike on the FLKC...so how do I tune this out?

here is a log with a few more timing parameters that makes it easier to see graphicly.

http://datazap.me/u/steve99/98-knock...zoom=1846-2750

[Silver Ignition]

Originally Posted by steve99

All you do is look at the Engine load/RPM point where the knock is occurring.

In this case 1.2 to 1.3 load and RPM 6100-6700 and the ECU is pulling about 2 degrees so remove about half that ie about 1 degree or slightly more from the cells in that area ie load 1.2-1.3 and rpm 6100-6700 area out of the Base_Timing_B table.

I think I'm getting this now...


You're awesome dude thank you so much

[Kodename47]

Originally Posted by Silver Ignition

I think I'm getting this now...


You're awesome dude thank you so much

Here's it in Laymans terms:
Base timing table is just that, your base ignition timing value.
Most people reference the KCA table just as the advance table and is why the IAM is the advance multiplier. So the calculation of IAM x advance table value is then added to the base timing. This gives you the output timing.
FLKC (Fine Learned Knock Control) will then be subtracted from above when necessary. If you log FLKC you will see where the timing is being reduced, and if you look at the total timing, usually you'll see the sudden change, especially as the ignition timing should be relatively smooth. If it's consistent then look to remove timing and in the area.


If you can see the timing as a graph (total timing ideally), then look to make it smooth. It could be a non-smooth table is inducing knock.


Using VGI's tool can make it quick and easy to spot any correllations etc. You could alternately try changing the fuelling a little if timing doesn't seem to make much difference. I would test both methods.

[Silver Ignition]

so what is the ECU seeing/doing when I have changes to FLKC? Is there actual knock events happening or is FLKC adjusted to prevent knock?

same maps/settings as the last log, only this is a 3rd gear only, 2-7.6k pull
http://datazap.me/u/silver-ignition/...=0&data=1-4-12

[Kodename47]

Originally Posted by Silver Ignition

so what is the ECU seeing/doing when I have changes to FLKC? Is there actual knock events happening or is FLKC adjusted to prevent knock?

FLKC is in response to what the ECU though was a knock event. It's not instant, but is stored for when that area of timing is passed again. The ECU will try and add timing back in after time to "test" the area again but if it continues to see that it's a problem area it will keep the timing removed.

[steve99]

Originally Posted by Silver Ignition

so what is the ECU seeing/doing when I have changes to FLKC? Is there actual knock events happening or is FLKC adjusted to prevent knock?

same maps/settings as the last log, only this is a 3rd gear only, 2-7.6k pull
http://datazap.me/u/silver-ignition/...=0&data=1-4-12

As per @Kodename47 post above

also read link below its for WRX ECU but the process is same but IAM calculated different in BRZ/86 but it appears most of it applies to 86/BRZ

http://www.romraider.com/forum/topic1840.html

http://www.romraider.com/forum/topic5371.html



It appears the ECU applies the FLCK correction to the timing and also stores it in a fine learning knock correction table similar to below (this is WRX one. As Kodename47 says if the value in the learning table is negative the ECU will always be trying to "return " the timing if the next time it passes that rpm/load range it see no or less knock. This is why you sometimes see positive FLKC as the ECU "returns" timing previously pulled.

Thats the way I think it works anyway

[Silver Ignition]

Originally Posted by Kodename47

FLKC is in response to what the ECU though was a knock event. It's not instant, but is stored for when that area of timing is passed again. The ECU will try and add timing back in after time to "test" the area again but if it continues to see that it's a problem area it will keep the timing removed.

okay so that's why the "spike" looks the same each time I do it, because it's learned, not an instant thing.

Originally Posted by steve99

In this case 1.2 to 1.3 load and RPM 6100-6700 and the ECU is pulling about 2 degrees so remove about half that ie about 1 degree or slightly more from the cells in that area ie load 1.2-1.3 and rpm 6100-6700 area out of the Base_Timing_B table.

I did exactly that…flashing now…I'll run it tomorrow and post up

[Walla Walla]

Originally Posted by Kodename47

FLKC is in response to what the ECU though was a knock event. It's not instant, but is stored for when that area of timing is passed again. The ECU will try and add timing back in after time to "test" the area again but if it continues to see that it's a problem area it will keep the timing removed.

so, if we want to track down the right timing correction we need to make, we should track those when the fuel map are learned (close loop to be apply open loop) but before FLKC has to much data learned, so it wont mask knock by removing some timing (and probably remove more than is really required, to keep the engine in a safe zone) ?

[Kodename47]

Originally Posted by steve99

This is why you sometimes see positive FLKC as the ECU "returns" timing previously pulled.

I nearly wrote the same, but it's not quite right. Positive FLKC is when IAM is less than 1 and timing is being added above the IAM x Advance calculation to test if the IAM can be increased. The FLKC value will just return to 0 if the advance adjustment is removed.

Originally Posted by Walla Walla

so, if we want to track down the right timing correction we need to make, we should track those when the fuel map are learned (close loop to be apply open loop) but before FLKC has to much data learned, so it wont mask knock by removing some timing (and probably remove more than is really required, to keep the engine in a safe zone) ?

I would certainly wait until the fuelling has settled down, as a change in AFR can make a difference.


The way I have used to easily see what's happening is as below, using VGI's Tool.
- Go to Log Stats screen
- Import Log
- On the set axis button, copy in your base timing table and verify (save for future if you want)
- Set X axis as engine load
- Set Y axis as RPM
- Set Data as FLKC
- Set a filter as FLKC < 0


I would then check the mode value and the count. You can repeat for FBKC too. Obviously if you log pull by pull then you won't need to do that, but if you're logging more than that, it can be useful to see other problem areas.

[Turdinator]

Originally Posted by Kodename47

The way I have used to easily see what's happening is as below, using VGI's Tool.
- Go to Log Stats screen
- Import Log
- On the set axis button, copy in your base timing table and verify (save for future if you want)
- Set X axis as engine load
- Set Y axis as RPM
- Set Data as FLKC
- Set a filter as FLKC < 0


I would then check the mode value and the count. You can repeat for FBKC too. Obviously if you log pull by pull then you won't need to do that, but if you're logging more than that, it can be useful to see other problem areas.

Is this the same VGI's tool that is used for MAF scaling?

[Kodename47]

Originally Posted by Turdinator

Is this the same VGI's tool that is used for MAF scaling?

The very same one....