Learning View corrections at 1.7+

My learning view typically shows corrections at the 1.7+ column from 2600 rpm and up. My WOT logs look fine and I haven't been able to log any knock in that area. I've tried consistently staying just above 1.7 from 2600 RPM on, than the same RPM at around 2.0, than varying the load, but no luck. Any suggestions?

http://i124.photobucket.com/albums/p38/mrsalty75/Forum/LearningView_SS_5-13-200974009PM.jpg

Assuming you are OS, the first thing I would do is adjust/rescale the load ranges (the columns) so that you can more accurately determine where the knock is (at what load range). Right now you only know it is above 1.70.

In the map, this is under ingnition timing Knock control > Fine corrention columns (load). Adjust the ranges so that they are broader..

Then, assuming you find a more accurate range where the knock is happenening, I would pull some timing and see if that helps.

I think what he is asking is why it isn't making correction entries in the last column above 2600RPMs when it used to. As it sits; he should be catching any learned knock above 1.7 in the last column. Consider it a good thing that your car isn't knocking in that range. It could be as simple as the change in fuel from winter additives to summer gas. If the logs are clean, then there is no knock in that area and therefore no need to make a correction entry in the KL table. It is best if this table is all zeros. Make sense?

I would worry more about the knock that you do have above 1.7g/sec at 5800+RPMs; as it's been pulling almost 3 degrees consistantly/repeatedly enough to cause an entry and therefore requires some adjustment;).

Assuming you are OS, the first thing I would do is adjust/rescale the load ranges (the columns) so that you can more accurately determine where the knock is (at what load range). Right now you only know it is above 1.70.

 

In the map, this is under ingnition timing Knock control > Fine corrention columns (load). Adjust the ranges so that they are broader..

 

Then, assuming you find a more accurate range where the knock is happenening, I would pull some timing and see if that helps.

Sorry, yes I am running OS. I'll play with that and see if I can narrow it down. Earlier today I though to myself it would be nice to be able to adjust that range and here is the answer. Thanks

I think what he is asking is why it isn't making correction entries in the last column above 2600RPMs when it used to. As it sits; he should be catching any learned knock above 1.7 in the last column. Consider it a good thing that your car isn't knocking in that range. It could be as simple as the change in fuel from winter additives to summer gas. If the logs are clean, then there is no knock in that area and therefore no need to make a correction entry in the KL table. It is best if this table is all zeros. Make sense?

I would worry more about the knock that you do have above 1.7g/sec at 5800+RPMs; as it's been pulling almost 3 degrees consistantly/repeatedly enough to cause an entry and therefore requires some adjustment;).

Thanks for the advice. This LV is after about 55 miles. Normally I see like a 1.4-1.7 correction more in the mid range RPM. The 2.8 correction is a new one, but I will for sure look into that one

With the stock tune, the ECU will pull 1.4 degrees every time it uses this table to mitigate knock.[1] It adds the timing back in increments of 0.35. So the 2.8 indicates means you knocked, it pulled 1.4 degrees, you knocked again with 1.4 degrees of retard already in effect, and it pulled 1.4 more.

Might be a good idea to pull ~1.5 degrees from that area of your base timing map and watch it for a while.

You can get more information from Learning View if you change the retard from -1.4 to -1.01, and change the advance value from +0.35 to +0.20. That way, the 100ths digit of each cell will tell you how many times the ECU has pulled timing from that cell since the last time the table was cleared.[2]

[1] Sometimes the ECU mitigates knock using 'feedback knock correction' which does not show up in this table. Sometimes it decreases the IAM instead.

[2] When the ECU changes IAM, it clears this table. For most cars, reflashing also clears the table. I never knock enough to force a change to IAM, so for me this table basically tells me how many times the ECU has used this table to mitigate knock, in each cell, since my last reflash.

Thanks I started reading an earlier version of that. I'll be revisiting that one too

I though this was only possible if IAM was <1?? I did another learning view after about a 6 mile drive and it's all zeros now.

http://i124.photobucket.com/albums/p38/mrsalty75/Forum/LearningView_SS_5-15-200974256PM.jpg

Thanks I started reading an earlier version of that. I'll be revisiting that one too

 

I though this was only possible if IAM was <1??

I assume you mean the positive correction because yeah, that seems weird to me too.

Hopefully Merchgod can chime in here...

From Merchgod's info--the FLKC table values can be positive or negative. Basically, if there are no knock issues, it ups the timing via the FLKC table, or so I understand it.

"Fine learning knock correction (FLKC) allows for positive or negative correction to KC based on knock."

"Fine correction is designed to make "finer" adjustments to timing after the IAM has roughly corrected timing advance where there is no knock. Fine corrections are stored in RAM and are applied to KC all the time (except for certain conditions like idle). "

You can get more information from Learning View if you change the retard from -1.4 to -1.01, and change the advance value from +0.35 to +0.20. That way, the 100ths digit of each cell will tell you how many times the ECU has pulled timing from that cell since the last time the table was cleared.[2]

 

[1] Sometimes the ECU mitigates knock using 'feedback knock correction' which does not show up in this table. Sometimes it decreases the IAM instead.

[2] When the ECU changes IAM, it clears this table. For most cars, reflashing also clears the table. I never knock enough to force a change to IAM, so for me this table basically tells me how many times the ECU has used this table to mitigate knock, in each cell, since my last reflash.

Thanks for the tip.

Thanks I started reading an earlier version of that. I'll be revisiting that one too

 

I though this was only possible if IAM was <1?? I did another learning view after about a 6 mile drive and it's all zeros now.

There should be no positive FLKC possible when the IAM = 1.0. PM me (at the romraider site) the map you had on there when you took this screenshot.

OK, you did what I thought you did. You changed the "fine correction retard value" from -1.40 to 1.01. What this does is add, instead of remove, timing during a knock event when making changes to the FLKC table. Obviously, that is bad. When removing timing from the FLKC, the ECU does not check that value against your (advance map * IAM), so, in your case, where it is adding timing (because of a positive step value), it will allow positive FLKC when IAM = 1.0.

You need to immediately change that value to a negative number.

I had 1.01 for Fine Correction Retard Value when it should have been -1.01

Yep, beat me to it in my long-winded post.