Sale ends todayGet 30% off any course (excluding packages)
Ends in --- --- ---
Quick question on MegaLog ViewerHD: Feedback and Fine Learning Knock Correction (4byte)
I went ahead and added together; summed for total Knock Correction. Is this correct? similar to Total Trim (what Andre does in his practical reflash video lessons).
Also, should I be worried about These knock values at a load of 0.67 and ~3400RPM through 3500RPM
I asked this on Facebook and they said to shut off (increase the low thresholds for knock to 1.00 load and ~ 2500RPM because somehow bellow 1.00 load is irrelevant and should be ignored?
(picture of my log)
Im not a pro but my take would be:
- FBKC: This value is the ECU reducing ignition timing in response to the knock sensor (which is an audio sensor stuck to the block) in an attempt to stop the engine from knocking. If the knock doesn't go away it will continue to reduce timing and you will see increasingly more negative values as it tries to remove more timing to stop the knock.
- FKLC: These are learned values in specific ranges of RPM and load. The ecu stores these ignition timing reductions to apply them pre-emptively in those areas in the hope that it will stop knock from occurring to begin with so that the FBKC system doesn't have to react to it after it already happened
Theres a lot more detailed info on it here:
https://cobbtuning.atlassian.net/wiki/spaces/PRS/pages/338723039/Subaru+Knock+Monitoring
To answer your original questions though of...
Should you sum FBKC and FKLC?
IMO you could if you want and you are aware of what each component is, but you could also just look at them independently. May also be misleading to combine them as if you look at them combined on a histogram you cant tell what portion is from FKLC which is just a learned value applied even though it might not actually be knocking, where as FBKC is actual knock events.
Side note: Id probably also take a good look at knock related stuff on a timeline so you can see what else was going on when the knock happened regardless of whether you sum FBKC and FKLC. The histogram gives you a good idea of the problem area in terms of load & rpm, but it may obscure some stuff as well since the histogram bins get filled by averages or min/max values.
Should you be worried or do something about the knock values?
It depends. Check out the cobb article but long story short... getting FBKC values means the sensor is picking up noise. It might be due to a noisy engine with loose tolerances, or it could be rough shifting or some other source of noise that isn't a concern in which case perhaps increase the threshold might make sense.
If its not those things though and the knock is due to some real issue then you would probably be better off figuring out what the cause is and addressing that cause rather than just adjusting a threshold so the signal goes away.
The COBB article helps a bunch.
I view them separately, but yes the total knock based correction is the sum of the two values, plus the correction applied based on DAM * DA.
Also keep in mind FLK and DAM are both "learned values", but they only drop when knock is detected. For example if the vehicle has no FLK, then drops for the first time in a given RPM/load area, that is from an actual knock event, whether or not there was also FBK at the time.
The same goes for DAM. Every time it drops, knock occurred at that moment, whether FBK or FLK shows it or not.
And disabling knock detection at low load or RPM isn't something I would suggest doing as a rule. There may be some instances where it makes sense, but I'd leave those values stock unless you really have a good reason to alter them.