No announcement yet.

MS2 - EAE Wall wetting algorithm

  • Filter
  • Time
  • Show
Clear All
new posts

    MS2 - EAE Wall wetting algorithm

    Anyone played with the EAE wall wetting acceleration enrichment on an m20?

    (this is not the delta TPS enrichment, this is the EAE wall wetting one)

    I turned it on the other day and notice an immediate change in throttle response and general crispyness of the engine. I am just using the default correction curves.

    The only thing i have found is that i think it is pulling too much fuel in the lower rpms and when crawling around in carparks the AFR goes quite lean upon light application of throttle. I need to do some experimentation but i believe that means i need to reduce the "adhere to walls coefficient" in the lower rpms.

    The main difference it has made is when you are at top of 1st gear, and you change gears it used to go quite rich on throttle lift, causing it to bog down slightly when changing to 2nd gear fast. This algorithm has really helped with that. Just hammers once you drop the clutch into 2nd.

    I can see it being a very powerful algorithm and when reviewing logs i can see it working a lot to alter the fuel delivery. however that also means that it has the potential to screw everything up with the wrong settings.

    I'd be interested to here from anyone who has played with the 6 coefficient graphs for any tips with the m20.

    The key is knowing there is a delay when you lift (TB closes and there's still some unburnt fuel) as well as a delay in your wideband readings, so it's important to know WHEN and exactly WHY the throttle goes rich when you are shifting. To me it sounds more like wrong AE (accel enrichment) settings. I haven't messed with EAE but I think varg mentioned it's more to do with part throttle and transient situations, not WOT shifting.


      Thanks for the reply. I have attached a photo of the log on a 1st - 2nd gear change.

      I wasn't sure if accelerate enrichment (TPSdot based) was supposed to be used in conjunction with EAE? or are you supposed to turn off AE if you are using EAE?

      the white line along the bottomm is the EAE (not exactly sure what the numbers refer to, they are quite big values?) you can see it pulls the fuel when i Lift. and then it adds it back in when i WOT again into 2nd. the AE is attempting to add it back in as well but it looks like perhaps i need to add a bit more as there is thaht lean spike when i jump back on, then it gooes rich again for a tiny bit then it evens out? should i expect to be able to even this out more? or is this fairly typical?
      Click image for larger version

Name:	AE EAE.jpg
Views:	223
Size:	56.6 KB
ID:	9866872
      to be honest i probably should be focusing on driveability more than logs, but i like numbers and lines:)

      Any comments welcome.


        have you got the steady state tune spot on at all load/rpm sites? i wouldn't try to chase ideal AFR on transient go by throttle response.
        89 E30 325is Lachs Silber - currently M20B31, M20B33 in the works, stroked to the hilt...

        new build thread


          yer its pretty good, this is my AFR table and i am within 5% of (based on EGO correction percentage numbers). Note the Kpa the left, i dont have a turbo (yet) so im only using half the table at the moment. .

          Interested in opinions on target AFR, seems to work pretty good. doesn't like idling much leaner than 14.2. the bunch of 14.7 numbers is to get it to closed loop target on cruise. probably need to smooth it a bit nicer in the 3100rpm band but it seems to work well as is.

          You are probably right about somewhat ignoring AFR on transient. I am just not sure how too tackle the EAE wall wetting. I cant even find the maths on it. I assume its looking at two cells that you are going between (using a prediction algorithm?), and subtracting or adding fuel based on the "fuel sucked off walls" table and the "adhere to walls" table coefficents, And the temperature and RPM correction for both of these are taken into account too. 6 correction maps to look at, But how do you even begin to tune this?

          Click image for larger version  Name:	afrtarge.jpg Views:	0 Size:	65.2 KB ID:	9866899


            I honestly don't think you need EAE at all, or if you do it'd probably be one of the verrrryyyy last things you tune. Your AE settings are probably the cause of any issues between WOT shifts.


              There's a setting in the time based accel enrichment dialog that dictates the relation to engine RPM with more authority than EAE. Full accel below and zero accel above, those settings linearly scale the AE pulse width down to zero as the RPM approaches the zero accel point. This will scale your AE pulsewidth down at high rpm and can cause what is the perfect PW adder when you tip in for a pull at a lower RPM to be too little at high RPM. Without knowing what size injector you're running and how many squirts per cycle you're running it's hard to judge what a reasonable accel PW is but for over 1000%/s, 2.1ms is a fairly small PW adder and this could be the result of scaling due to the zero accel above point. Accel enrichment is a driveability setting, don't get too hung up on having a perfect looking datalog, In events like up or downshifting you'll likely never achieve a nice linear transition with no rich or lean spikes transitioning from target AFR at the RPM you start at to target AFR at the RPM you wind up at, and you wouldn't likely be able to tell just by driving if someone were to alter a 'good enough' tune so that it did just that without your knowledge. Because of this, EAE and AE in general is indeed one of the last things to address, because you can bake AE-like effects into your VE table inadvertently or on purpose and wind up chasing something unrelated to AE quite easily, autotune can cause this sort of thing by altering your VE table in a transient area in response to a spike or dip. If your throttle response feels good in all conditions with no hesitation, I'd call it good. A brief 16:1 AFR lean spike during a quick transient is not a huge deal, and AE, especially EAE, is 100% trial and error and just takes time to perfect.

              Also, recall that if you add too much fuel during the AE event you can cause a brief miss which will read as a lean AFR spike.
              '91 318is, M20B25, T3/T04E 60 trim (18psi), megasquirt, coilovers, Z3 rack, cold AC
              [b u i l d]
              [Car of the month: April 2018]



                Thankyou for your replies. I will definitely have to look at the AEs setting you mention. Something is not right with the time based AE as you have pointed out. injectors are oem m20b23 injectors which are 185cc, 2 squirts per cycle.

                Also im pretty sure the AE setting is also used in reverse as the "deceleration enleanment". If the AE is not dialed in properly then the deceleration enleanment wont work properly either, probbably a contribution to my rich fuel dump after closing the throttle.


                  Well i have been playing more with the E-AE and to be honest it has been reasonably successful. I have kept trying to tell myself that drive ability is more important than AFR values and that has helped dial it in.

                  I haven't touched the AE at all, other than increasing that "Zero accel above" to 6000rpm ( It was set to 5000 before).Ive just been fiddling with the EAE. The EAE maps that make the biggest difference are the "adhere to walls coefficient RPM correction" and "sucked from walls RPM coefficient correction". I dont have enough experience to know but it seems that the other maps are reasonable as they are from a drive ability point of view.

                  I was able to reduce the effect of the EAE at low rpms (crawling around carparks etc) and that made it smoother. I think the EAE is worth pursuing over the traditional EA. I will keep playing and improving.

                  To be honest i need to put a cover over my AFR gauge so i stop looking at it, and just enjoy driving it!