Mag's OSE 12P Thread

[mag]

thanks VL400 for clearing up the ve learn/blm issue, another piece of the puzzle solved but remembering an earlier post of yours where you said

"VE learn is not the be all and end all, you have to do the hard yards still. It will get it close-ish and then you have to manually tune it."

why manual tune, ve learn & nb o2 sensor should give good ve values for afr 14.7?

also more questions regarding ve values for afr other than 14.7. if afr target is not 14.7 then i thought nb o2 sensor is not much use & therefore logging & tuning using blms would not be valid either. would need wb o2 sensor to measure afr other than 14.7 & setup ve table, or have I got it all screwed up.

[Holden202T]

ve learn should be able to get the 14.7 areas of the VE table almost perfect .... but the reason its not the be all and end all is that it can't handle the other areas and as far as life of the engine they are probably more important.

can you imagine doing a 1 minute burnout at 14.7:1 at WOT .... if the engine did infact make enough power (due to being so lean) to do the burnout then it wouldn't take long for the engine internals to fry....

so yes you can drive around all week with VE Learn and you'll probably find you'll get good results for the most part, but if the non 14.7 areas are not right theres only one way to get them dialled in and thats by manually adjusting them.

you can use NB o2 for these and that will get you in the ballpark, but at the end of the day theres not much that compares to a WB o2 tune for these areas.

eg. my mates VS V8 ... worked over with different injectors, pretty much was starting from scratch with that tune ... was rich everywhere so its safe, as far as torching the motor goes but still needed to get it half drivable.

so i used the NB o2 to dial in the WOT areas, got it to about 820-850mv ... i knew this would be rich still but was as lean as i was happy to go to get it reasonable and still keep it safe. So when i got the WB o2 onto it, turns out it was about 11.8-12.3 in most areas, so not too bad but slightly rich.

and the only reason i knew to get it to 820-850mv was from comparing WB and NB logs from my ecotec .... (820mv is about 12.8 in the ecotec logs)

[mag]

can understand ve learn good for afr 14.7 but thought o2 rich/lean output voltage readings were too unstable to be really useful. anyway probably will get a wb o2 setup & then the risks will be small & expect it will be easier to tune.

[VL400]

Read up how O2 sensors work in conjunction with a cat converter to understand more, but the rich/lean swings are caused by the ECM during closed loop. Its to do with the cat releasing and storing oxygen so it can turn harmful stuff into slightly less harmful stuff. How long the sensor spends on each side of 14.7:1 is what the INT and BLMs are correcting - spend too long on the rich side and the INT drops causing a reduction in fuel, if it reaches a set point and still needs to pull more then the BLM drops a point, resets the INT and it starts again. There is a wealth of info on the net about closed loop operation.

[mag]

VL400, your BLM explanation is best I have read. same principal applies for VE learn but in this case ve table values are changes?

But still a bit puzzled about using nb o2 sensor output as a accurate way to measure richness or leaness. my understanding of o2 sensor was that its output was "unstable" in that it is either >0.8V or <0.2V for rich or lean & afr 14.7 +/- 0.2 gives 0.45V. thus the o2 sensor output cycles between high & low as the ecu tries to hold afr 14.7. if true, then how rich or lean the fuelling is would be hard to tell from a nb 02 sensor.

a few hundred millivolt change covers afr 12 to 14.5 & 14.9 to 17 is not much. is resolution/stability of the 02 sensor output in the rich or lean regions good enough to use to determine afr or are we just looking to make sure we always get an output >0.8V & therefore rich.

[VL400]

VE learn basically replaces the BLM part of closed loop, it still uses the int but instead of adjusting the BLM it adjusts the VE table directly. The BLMs get locked at 128 when then VE learn option is active.

The narrowband sensor should not be used to measure anything accurately other than to tell you if your rich of stoich or lean of stoich, any actual AFR is a guesstimate. Based on logging that holden202 did while using a wideband in conjunction with the factory NB he found that a certain mV from the NB correlated to a certain wideband reading - and as he said "you can use NB o2 for these and that will get you in the ballpark, but at the end of the day theres not much that compares to a WB o2 tune for these areas" I would not be relying on the values you get from the NB.

[mag]

Holden202, below are examples of ve table values. first is from cal that was intially loaded into ecu & second is from cal recovered from ecu after running ve learn for a week or so.
20 30 40 50 60 70 80 90
1200 48.83 61.72 68.75 72.27 75.00 76.56 78.13 79.30

1200 48.82 34.38 52.34 64.06 67.97 68.36 80.08 77.34

from the blm history it appears 20kpa range has not been "learned" but the other cells have been. also the pattern of the big change at 30kPa is repeated for other 30kPa cells up to 1800 rpm.

as you say the 20kPa figures may need to be aligned with the "learned" adjacent cells. just suprised that the learned values deviate so much from the initial value but if data is valid then it shows the value of VE learn.

[VL400]

Can you attach here or email me your cal so i can check the VE learn and other settings that your using? Looking at your VE snippets your getting adjustments at 90kPa. And BLM history? BLMs are disabled when the VE learn option is activated so should be locked at 128.

[mag]

understand that blm disabed (all set at 100) but thought blm history would still show which cells have been visited during ve learn & therefore which might have been changed.

Have attached bin.

[Holden202T]

you wanna know whats changed, compare before and after bins!