GR Yaris Knock Correct Learn Value

[MarcoB]

Starting a thread on this specific value to seek more information on this.
Particular on improving this number.


What I found so far, without disregarding the fact it is part of a safety feature of the ECU:

-This retarding value for optimum power levels maxes out around 25.

-There seem to be 2 KCLV’s operating regions: low RPM and high RPM (below and above 4000RPM, about the same as the changeover from port to direct injection).

-The low KCLV changes faster than the high KCLV value.

A company in Japan called Vehicle Field
apparently developed a special method that brought both regions to 24.1 in approximately five minutes.

Now that would be very interesting to know!
Since slow driving alters this value too, it seems it does not stick.
However the low-rpm KCLV recovers quickly.

A quick glance in Techstream showed mine at 21.9
I took this reading from a standstill and idling, so this value says nothing about the high RPM KCLV.

ChatGPT found some more info on this matter from that same company.

They describe a modification intended to:
-Make the learned value less likely to fall.
-Make it easier to recover after falling.
-Maintain higher learning values during summer.

But the detailed procedure is behind a paid article. So Chat is not able to come up with more details.
But it would certainly be interesting to know how we could maintain this higher value.
(Besides the obvious as to fill up consistent premium fuel.)

 

[tonic]

This is from the GR Yaris service manual:

 

[MarcoB]

This is from the GR Yaris service manual:

View attachment 42901

Excellent information! Thank you.
I think this ads to the puzzel of why this ECU is not making the same amount of power every time we take her out.

Ofcourse it has all to do with safety.
But when there is (or better yet ‘was’) no knocking present I wander why the timing advance does not stay at the highest number for that specific load?

For the low-RPM region, it quickly returns.
But in those instances when you ask the most power (the high-RPM region) it takes al lot of time to get that maximum power back.
Because it does not retain that highest learned value.

I guess when those guys in Japan understood the way this ECU works they came up with this sort of ‘sequence’ to get the ECU to learn those high values within 5 minutes of doeing something

 

[Boosted8]

@MarcoB do you have the means to collect a complete datalog for full and part throttle across the whole rpm range? The risk is you take action on a singular data point that is unessisary.

 

[MarcoB]

From CerN in a different thread:

If you do a 5th gear pull from 2k RPM, the value very quickly adjusts.

Keep in mind, the value is split in two. There is one knock correction learned value for sub 4k RPM, and then once you cross over 4k, it switches to the other value.

Fuel quality seems to matter a lot. I filled a brand new tank of 98 octane any my values were sitting on high 21s and low 22s. Then a fresh tank from a different station brought me up to the 24 range.

According to some japanese youtube video I watched a while back (I think it was this one: https://www.youtube.com/watch?v=3wdLpvu-p58), they tested a car with low knock correction learned values, and they were losing well over 50hp when it was below/around 18, which can be the case even on stock cars if the car is not happy with fuel/spark plugs/injection etc.

I don't think its effect on power is linear. Being anywhere in the 20-24 range doesn't seem to change power a huge amount, but somewhere close to 18-19 the car loses a lot of power.

 

[MarcoB]

@MarcoB do you have the means to collect a complete datalog for full and part throttle across the whole rpm range? The risk is you take action on a singular data point that is unessisary.

I am aware of the possibility -as suggested above and found in other sources- of 2 different values.
Both found by trail and error from different people (published on fora and yt who tried this and logged all the info).

So indeed I will have look into this.
(Staying away from the low-rpm value increase ‘trick’ by heavy load under low rpm).

Unfortunately the only real source of information is the part from the service manual.
And that clearly gives us an formula:

Ignition Timing = Most Retarded Timing Value + Knock Correct Learn Value + Knock F/B Value + Each Compensation Amount

Actual ignition timing also depends on other stuff such as engine speed, load, temperatures and -ofcourse-knock feedback.