do you have LTFT affecting WOT disabled? if not, then you should

 

i see. i'm reading that thread now. just ignore me, i'm thinking out loud now.

 

Hi mo.

Here's a great example of how spool-up "measured" AFR can cause headaches. These two WOT log fragments were taken on the same drive, less than 15 minutes apart.

In each graph, check out the LC-1 reading at 4000rpm / 230 load (highlighted).


The first was WOT in 3rd at 3500rpm, so the "interesting" measurement occurred right at the top of my (low) spool-up peak...





The second was WOT in 3rd at 2500rpm, so I'd been at peak boost for a bit before 4000rpm / 230 load...




I mean, seriously... choose between any/all of:

1. MIVEC-related blow-by of unburnt mixture causing wbo2 to read "more oxygen" -> "lean".
2. Wideband lag.
3. Load being calculated differently during spool-up (when my MAP_Calc lags MAF_Calc).


It's most likely a combination of all of the above.


Do I for a minute think my "real" fuelling goes from 13.0:1 -> 12.0:1 -> 11.6:1 in 1/10 second increments? No.

But is the spool-up mixture actually leaner than my steady-load mixture? Yes - it is really leaner. I do have to be careful.


My full-noise AFR target is "low 11s". My STEADY-STATE mixture targets (ie. real-world AFR measurements) look like this...



It looks so simple, but that first log graph bears no resemblance to the target AFR data, does it?!


Hope this is all good food for thought!

Rich

 

yea, i've been reading your thread on the subject as well. i concur with the 3 ideas above but i really dont like the possible map/maf calc delay. how are you calculating your map/maf loads individually. i assume it's a formula using maf voltage?

i have mivec load and fuel load (which i would guess is calculated load using both map and maf).

i'm okay with hitting a max of 12.5 AFR on my wideband during spool up but i've built a test map to try when i get back from vacation:


lastly, i may try to modify the calibration fuel maps to richen the area in the 3000-4000 rpm, 220-260 load area to make the AFRs more 1:1, do i just increase the number in those cells and bring them closer to or above 100?

 

Note that my Ralliart ROM has NO "Accel Enrichment", whereas your X ROM does.

I'd love to explain how that works, but I'm not really familiar with it (seeing as it's all disabled on my car).


Here it is for your 58010005 ROM:



Mine is all zeroes... but it doesn't work even if I populate it.


I really should go digging again, and see if I can reanimate AE on Ralliarts!

Rich

 

Anymore work on this subject rich? It was an interesting discussion that triggered my curiosity

 

Yes!

Regarding the lean mixture as boost peaks... Solved!

We needed to increase Load Ramp Rate #1 from its factory 3.53% to something nearer 7.5-8.0%. That stops calculated load lagging behind where it should be, as load rises fast during spool.

[edit]See this thread for lots of details. Load Ramp Rate #1 is now defined on goldenevo.com for most ROMs.

I'm currently using 7.84% with great success. My spool-up mixtures (and timing, and SST pressures) are no longer behind the curve at top of spool.

Rich

 

Is this applicable to GSR's as well?

 

Use RAX Fast Logging to track MAPCalcs, MAFCalcs, ChosenCalc. If ChosenCalc lags behind the "lower of MAP/MAF" when load is increasing rapidly, then yes, increasing Load Ramp Rate #1 could well resolve it.

Rich

 

I'm actually more curious about why the Ralliart is lacking Accel Enrichment....

That seems backwards to me. With a smaller turbo, I'd expect the Ralliart to need it *more* due to an even faster boost response.

This is the table that injects fuel for quick stabs of the throttle right? Why would the Ralliart lack it? Was it there only for sporty throttle? I figured it helped tune out tip-in lean conditions.

I thought I understood it properly but after finding that the Ralliart has it disabled I'm confused again.

 

thanks, I'll check it out. I'm guessing the spool on my EF3 is slow enough that the load calcs shouldn't really be lagging enough to matter.

 

tephra has posted up info on tuning the Accel Enrichment quite recently... and found it's disabled on X as well! There is a critical multiplier table (for ECT, if I remember correctly), that is stuffed full of zeroes, thereby making the whole mechanism do nothing whatsoever.

Rich