Notices
Engine Management and ECU Remapping This section is to discuss the various aspects of engine management modification for your Subaru.

ESL and its Darn Knock control !

Old 11 May 2017, 08:40 AM
  #1  
The Rig
Scooby Regular
Thread Starter
iTrader: (13)
 
The Rig's Avatar
 
Join Date: Dec 2002
Posts: 7,793
Default ESL and its Darn Knock control !

Since i reset my ECU i keep getting a negative figure of retarding timing oin the knock control map at load 40 at 2600 rpm.

Now, before i reset the ecu it was running fine for many many many mths with the same map and my driving style with no adjustments in the knock control.

i hvae removed a degree or 2 in the 39 and 43 load cells At 2400 and 2800 rpm as there isnt a 2600 rpm row so assuming its interpolating between these 2 at this time ?

Anyhoo, it shouldnt even be adjusting in this area as i have my threshold set to 6 and the advanced timing in the affected rows is at 5 ?????? But then again, the annoying thing is with the knock control table is its not calibrated very well, looking again at the table, it could be detecting knock anywhere from 2600 to 3800 (thats a difference of 1200 rpm! )as thats the scale on the knock control rows, so i dont really know at what RPM the knock occurs ????? I dont hear any knock, my knock lights dont flicker at all either. i dont want to reduce the timing if its not needed but as i dont know the exact rpm i could lower the timing at 3600 and 4000 at loads 39 and 43 ?

The knock control table after a reset can be all 0`s for a week or 2, even longer then poof, this rogue negative gets added at some point when i next check.

Anyhoo, anyone else have ghost adjustments in their knock control ?




Last edited by The Rig; 11 May 2017 at 10:10 AM.
The Rig is offline  
Old 11 May 2017, 11:07 AM
  #2  
ossett2k2
Scooby Regular
iTrader: (9)
 
ossett2k2's Avatar
 
Join Date: Jan 2007
Location: Leeds
Posts: 6,170
Default

Is a strange one,like you say it's not supposed to learn knock correction where the advance is lower than the threshold.
Maybe try rescaling the rpm to include 2600rpm?
ossett2k2 is online now  
Old 11 May 2017, 11:11 AM
  #3  
ossett2k2
Scooby Regular
iTrader: (9)
 
ossett2k2's Avatar
 
Join Date: Jan 2007
Location: Leeds
Posts: 6,170
Default

Ah just noticed,,will that correction be anywhere between 2600-3800rpm tho as the knock correction table will inerpolate between these 2 columns.
ossett2k2 is online now  
Old 11 May 2017, 11:51 AM
  #4  
The Rig
Scooby Regular
Thread Starter
iTrader: (13)
 
The Rig's Avatar
 
Join Date: Dec 2002
Posts: 7,793
Default

Exactly bud, as its recording at 2600 yet the next phase of the knock control table is 3800, it could be picking up knock on my timing table anywhere from 2400 2800 3200 3600 and 4000 as ive noticed its never quite accurate when it interpolates.

Also as its hitting at load point 40 and the next load point is 46 on the knock control table, it could be load columns 39 and 43 thats "causing" the correction .

Just seems i have to adjust alot of rpm cells and 2 load cells as a guess to where the "fault" is occuring, im not keen on dropping timing where its not needed as we all want that power lol . Just annoying more than anything .

I could lower load cells 39 and 43 at 3600 and 4000 rpm cells down to 5 as these are at 6 currently and see if that does anything, but its another ecu reset and fuel trims being learnt which bugs me for a couple days lol.

Just wish the kock control cells were not as wide apart as they currently are
The Rig is offline  
Old 11 May 2017, 12:25 PM
  #5  
ossett2k2
Scooby Regular
iTrader: (9)
 
ossett2k2's Avatar
 
Join Date: Jan 2007
Location: Leeds
Posts: 6,170
Default

After you have reset the ecu do you datalog straight away before the correction happens?
Does a knock event or search step show up on the log,and would this show you more accurately where the correction is happening?
ossett2k2 is online now  
Old 11 May 2017, 01:15 PM
  #6  
The Rig
Scooby Regular
Thread Starter
iTrader: (13)
 
The Rig's Avatar
 
Join Date: Dec 2002
Posts: 7,793
Default

I've logged after a reset and no correction occurs in the knock control map , as I say , I can drive as I normally do with no corrections , check the table , nothing changed then maybe check couple weeks later a correction .
The Rig is offline  
Old 11 May 2017, 06:36 PM
  #7  
The Rig
Scooby Regular
Thread Starter
iTrader: (13)
 
The Rig's Avatar
 
Join Date: Dec 2002
Posts: 7,793
Default

A couple more stats to ponder over lol

The Rig is offline  
Old 12 May 2017, 09:01 AM
  #8  
bludgod
Scooby Regular
iTrader: (3)
 
bludgod's Avatar
 
Join Date: Feb 2007
Location: Belfast
Posts: 1,736
Default

have you check the a/f corrections in that area? Might be a little more fuel in that region will help balance things out if it's just occasional noise. Could it be shift knock from some hard driving?
bludgod is offline  
Old 12 May 2017, 11:54 AM
  #9  
The Rig
Scooby Regular
Thread Starter
iTrader: (13)
 
The Rig's Avatar
 
Join Date: Dec 2002
Posts: 7,793
Default

Fuel corrections are not too bad, from memory i think my fuel table requests 13,12.8, 12.6 and 12.4 around those areas, i may richen up slighlty but as you say, im more convinced its shift noise or gearbox noise from when i hoof the loud pedal down in 5th gear and the boost kicks in .

what im going to do is, rather than remove another degree of advance timing to get the advance under the threshold so its not learning, im going to add a degree of timing to the base map then remove the degree of timing from the advance so it will be at 5 in those cells so shouldnt learn as when it was at 6 yet i still keep my timing the same.

Hopefully tht works

ive tested the knock sensor and it read 560k so looks to be ok so not a duff sensor.

cheers all
The Rig is offline  
Old 12 May 2017, 12:03 PM
  #10  
bludgod
Scooby Regular
iTrader: (3)
 
bludgod's Avatar
 
Join Date: Feb 2007
Location: Belfast
Posts: 1,736
Default

yeh the knock advance table is only part of the knock control system, the advance values are more related to how much timing is added/removed as the IAM increases/decreases (and when IAM re-learning is possible) so you'll still have active knock correction potentially outside those figures. The A/F learning I'm not sure on the logic there, if it's anything like a newage car process then when your target is richer than 14.5 or so the ECU won't make corrections so if it's too far out it will remain so.

keeping the same total timing and playing with the knock advance value may help, but if its hearing knock it will likely still pull timing - it just won't use that area during an IAM relearn (should it ever drop below 16).

Is your total timing smooth in that region? could also be a false report from jumping say 5 degrees in timing from one cell to the next, again based on newage car logic here but they will show a false knock reading if the timing goes up too much from one row/column to the next - check out all your cells in that area it might be you only see this when moving diagonally into/out of a specific cell when the lower RPM load is higher in 5th?
bludgod is offline  
Old 28 May 2017, 01:14 AM
  #11  
The Rig
Scooby Regular
Thread Starter
iTrader: (13)
 
The Rig's Avatar
 
Join Date: Dec 2002
Posts: 7,793
Default

Havent altered anything on the map since i started this topic, went for a blat today and checked the knock control map, below








So i though tfeck sake, what bloody now, so on the drive home from where i went i checked again






Things got a mind of its bloody own lol
The Rig is offline  
Old 28 May 2017, 07:51 PM
  #12  
bludgod
Scooby Regular
iTrader: (3)
 
bludgod's Avatar
 
Join Date: Feb 2007
Location: Belfast
Posts: 1,736
Default

are you sure it's not just shift knock getting picked up? would explain why it's learned and then unlearned as you drive. For all it's pulling you could try a little richer there or pull a degree of timing and if it makes no difference then live with it. Ideally get the headphones on and have a proper listen tho.
bludgod is offline  
Old 02 June 2017, 09:59 AM
  #13  
tjmatt
Scooby Regular
iTrader: (1)
 
Join Date: May 2008
Posts: 410
Default

I know nothing of ESL and its knock sensor filtering, but my heat shields can rattle between 2000-3000rpm. Could it be picking up something like that noise?
tjmatt is offline  
Old 02 June 2017, 02:39 PM
  #14  
The Rig
Scooby Regular
Thread Starter
iTrader: (13)
 
The Rig's Avatar
 
Join Date: Dec 2002
Posts: 7,793
Default

Could well be, i suspect its picking up gearbox noise but as its reducing timing its annoying !

Weird thing is, if i lower advanced timing values to below 6 in "false" alert areas ( but increase base timing to compensate) and set the threshold to 6 so it doesnt learn the "false" alert area anymore which have a value of 4 and 5 now, my IAM does not increase to anything more than 10 or 12 and to get it to move from 8 is a real struggle, you really need to WOT at 5000 rpm plus to get it to move to 10 or 12 when if i have the IAM threshold at 4 it goes to 16 within a couple of accelerations.

what i did do was , set the threshol to 4, get it to change the IAM to 16 then i set the threshold to 6 , but i suspect this may be why im struggling to have the car spool up quickly and it suddenly feels flat as its not using all the advance timing now.

Think im going to reset the threshold to 4, leave it and see what knock it detects and either leave it or............

cheers

Last edited by The Rig; 02 June 2017 at 02:42 PM.
The Rig is offline  
Old 02 June 2017, 03:04 PM
  #15  
bludgod
Scooby Regular
iTrader: (3)
 
bludgod's Avatar
 
Join Date: Feb 2007
Location: Belfast
Posts: 1,736
Default

that's not so weird - what your playing with there is the rough correction values. There are a few sides to the knock control system, feedback knock control, fine learning knock control and rough correction.

I'll try and explain as best I can but my knowledge is based more on the newage cars for this (though the system is comparable):

When you start the car from a reset, the IAM starts at the default value (2/8/10/16 etc. based on your map setting) and starts off in rough correction mode. When the knock advance value in the map is greater than the threshold value (say your advance limit is 6 and your advance map calls for 8) then the timer starts and if no knock is heard then eventually the IAM will increase (based on the IAM increment value and assuming it's not already at maximum). If the IAM is now at maximum then rough correction is finished with (until the IAM drops) and you move into fine/feedback only corrections.
If during the learning period the ecu hears knock or the advance value isn't high enough then the timer can't run and the IAM can't increase. So if your knock advance table is generally too low you'll have a hard time getting the IAM up to full strength. Same goes if you have a persistent knock area it will continually be pulling timing and the ECU will be attempting to compensate.

Once the IAM is max then immediate knock is handled as a feedback knock value (which you can log with knock sum) and persistent knock gets saved into the table as a fine learned correction value (which you can view with the ESL software).

If you find you've an area that's always pulling timing, maybe try lowering the base value to see if that clears it up, you can also try adding a little bit of fuel there too (which is best depends on your particular circumstances). But the best way is to have a listen with a set of cans - if it's knocking you'll hear it knocking, if it's not knock then it could be a rattly heat shield/nut not tightened/broken knock sensor/noisy gearbox/noisy clutch etc. - the ECU can't really tell the difference it's just listening during the allotted time window and if the knock sensor sends a signal then it assumes it's knock.

I would be inclined to get the IAM at 16 - then do a run with less timing/more fuel and plot your log with AFR/total timing to see what it looks like. If you have big jumps in your timing table you can end up with a false knock reading so if your timing plot doesn't look nice and smooth then work the table until your happy with the shape.

If i've talked myself round in circles there now give me a shout on facebook I'll try and go through it a bit better for you
bludgod is offline  
Old 02 June 2017, 03:15 PM
  #16  
The Rig
Scooby Regular
Thread Starter
iTrader: (13)
 
The Rig's Avatar
 
Join Date: Dec 2002
Posts: 7,793
Default

Cheers bud good info as always , thing is, my threshold was 6 , my mid throttle advanced timing cells are now 4 and 5 due to these false knocks yet 3/4 and full throttle have advanced timing of 7, 8 and 9 and these cells were being hit yet the IAM didnt increase from 10 o 12. If i lower the threshold to 4, bing IAM goes to 16 nicely.

As you say, think im going back to a lower threshold of 4 and will try and correct the knock from there rather than hide it as its not pulling hard like it did before since ive set timing advance lower than 6 mid range with the threshold at 6.

cheers bud
The Rig is offline  
Old 02 June 2017, 03:34 PM
  #17  
bludgod
Scooby Regular
iTrader: (3)
 
bludgod's Avatar
 
Join Date: Feb 2007
Location: Belfast
Posts: 1,736
Default

when your at IAM 10 and running through the higher revs do you have any hits on knock-sum? The timer will take longer to run at higher RPMs vs the lower rev/spool up area - if you look at the standard map you'll see the knock advance values are spread out with more "potential timing" if you will in the more knock-prone areas.

As you say, i'd go back to your previous settings and initially try to dial the knock out after confirming how the timing plot looks (nice and smooth/no big steps). Far easier to do that with the IAM at 16 than with it bouncing about and potentially changing your timing target mid-pull. Check your AFR's as well - you can get a clue there if there's problems with the ignition so look for spikes in the readings that you can tally up.
bludgod is offline  
Old 15 April 2019, 04:41 PM
  #18  
Ash Webster
Scooby Regular
 
Join Date: Dec 2013
Location: Lytham St Annes
Posts: 853
Default

Sorry to bring up an old thread....

@ OP did you ever get to the bottom of this?

I'm having similar problems with knock sensor circuit at the moment . I'm also on ESL

Was yours throwing up CEL (code 22) too?

Mine has done it intermittently pretty much forever and its only recently I got around to trying to diagnose it. I got a brand new OEM sensor and the code still comes on every now and then. Sometimes after 1-2 weeks sometimes after a few days. Not even under high load....

I'm hoping its just ****ty knock sensor circuit/harness wiring as i dont map it myself.....

Last edited by Ash Webster; 15 April 2019 at 04:42 PM.
Ash Webster is offline  
Old 16 April 2019, 11:33 AM
  #19  
The Rig
Scooby Regular
Thread Starter
iTrader: (13)
 
The Rig's Avatar
 
Join Date: Dec 2002
Posts: 7,793
Default

Hi,
The only way i found around it was to increase base timing and only have advanced timing a couple of degrees in the affected areas.
I tried to increase the advance threshold to stop it looking at the affected areas but all this did was stop the IAM increasing from 8 to 16, always stayed at 8 ( i guess it wasnt looking at any advance as the threshold was too high.

So, increase base timing and then just have 2 or 3 degrees advance ( that is, if the actual timing isnt at fault, as you say, you didnt map it, i do map my own so can see whats going on etc etc ) If you have the cable. If not, well, you are at the mercy of the mapper
The Rig is offline  
Old 16 April 2019, 11:58 AM
  #20  
bludgod
Scooby Regular
iTrader: (3)
 
bludgod's Avatar
 
Join Date: Feb 2007
Location: Belfast
Posts: 1,736
Default

if its throwing a cel though it's gonna be the sensor itself or the wiring for it. if you've replaced the sensor with a new genuine one then take a good look over your wiring and test the wire continuity to the ECU and the grounds. Anything out of place then start replacing. The ECU's knock adjustment alone won't trigger the cel code - the code is for a faulty sensor not how you've configured the advance table.

also check for things rattling about in the engine bay, the sensor is basically a microphone to pick up noise so the ECU can't make any determination between a hose clamp/engine mount/downpipe vibrating against the block compared to the piston vibrating against the cylinder walls, it hears the noise and the ECU acts accordingly.
bludgod is offline  
Old 16 April 2019, 12:10 PM
  #21  
Ash Webster
Scooby Regular
 
Join Date: Dec 2013
Location: Lytham St Annes
Posts: 853
Default

Originally Posted by The Rig View Post
Hi,
The only way i found around it was to increase base timing and only have advanced timing a couple of degrees in the affected areas.
I tried to increase the advance threshold to stop it looking at the affected areas but all this did was stop the IAM increasing from 8 to 16, always stayed at 8 ( i guess it wasnt looking at any advance as the threshold was too high.

So, increase base timing and then just have 2 or 3 degrees advance ( that is, if the actual timing isnt at fault, as you say, you didnt map it, i do map my own so can see whats going on etc etc ) If you have the cable. If not, well, you are at the mercy of the mapper
thanks for the reply bro - but as you said if it was a mapping issue im at the mercy of my mapper and id hope it was something he would have found during his road mapping....

Originally Posted by bludgod View Post
if its throwing a cel though it's gonna be the sensor itself or the wiring for it. if you've replaced the sensor with a new genuine one then take a good look over your wiring and test the wire continuity to the ECU and the grounds. Anything out of place then start replacing. The ECU's knock adjustment alone won't trigger the cel code - the code is for a faulty sensor not how you've configured the advance table.

also check for things rattling about in the engine bay, the sensor is basically a microphone to pick up noise so the ECU can't make any determination between a hose clamp/engine mount/downpipe vibrating against the block compared to the piston vibrating against the cylinder walls, it hears the noise and the ECU acts accordingly.
i think this is bang on what i want. ive just printed off the manuals for checking continuity for the harness etc (error codse 22) but i dont have a ohm/voltmeter, might have to grab one from halfrauds/toolstation or nip to my local garage and see if he can do a quick check for me. could also be worth getting the car up in the air and wiggling a few things around to see if any bolts/pipes are slightly loose as you say...

If i can get this intermittent CEL fixed it will feel like ive got a new car - ive had this pretty much for ages, and its defninitely NOT the engine knocking but have never got round to narrowing it down (mostly due to laziness on my part, then taking the car apart over a year or two to refurb in my garage!). When i got it back from the mapper recently after the FMIC it was fine and again boom CEL back to actuator pressure and having to reset my ecu every few days...

If its not the wiring or something loose in the engine bay im stumped and i may have to find a way of bypassing it (putting the sensor in foam? lol) until my engine blows up and i can build a CDB from scratch,.,,,
Ash Webster is offline  
Old 16 April 2019, 04:52 PM
  #22  
ossett2k2
Scooby Regular
iTrader: (9)
 
ossett2k2's Avatar
 
Join Date: Jan 2007
Location: Leeds
Posts: 6,170
Default

I would also check and clean the block where the sensor sits plus torque to the correct spec,from memory(so double check) 24nm
ossett2k2 is online now  
Old 16 April 2019, 05:23 PM
  #23  
Ash Webster
Scooby Regular
 
Join Date: Dec 2013
Location: Lytham St Annes
Posts: 853
Default

aye im gonna try that too - also just found the last few posts on this thread very useful

https://www.scoobynet.com/905800-kno...lt-code-2.html

same problem as me - turned out to be faulty knock sensor wiring , the words "more common than you think" from an experienced garage give me some hope its just the wiring..
Ash Webster is offline  

Thread Tools
Search this Thread

Contact Us - Archive - Advertising - Cookie Policy - Privacy Statement - Terms of Service

© 2019 MH Sub I, LLC dba Internet Brands

We are a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for us to earn fees by linking to Amazon.com and affiliated sites.