AFR Table PEER REVIEW

[luftvagon]

Hello,
I need a helpful peer review regarding my latest AFR table before it gets put into CERT environment, and subsequently PRODUCTION.

This is for Type4 CV motor, 1970cc (stock 2.0L), 42x36mm intake, and exhaust valves, stock cam using SVDA Distributor.



My previous AFR table was similar, except anywhere you see 15 AFR, 16 AFR, and 17 AFR was set to 13. The fuel map it produced was beautiful with a lot of power, and not a lot of efficiency.
I am addressing economy with this version of the AFR table.

Here is the fuel map with original AFR values:



Please provide me with your opinion, and whether or not I should be doing anything different. Thank you.

[TouringBubble]

I'll just preface this with the fact that I haven't even begun to tune my ACVW on MS yet ... but I have tuned many, MANY Evos and STis, as well as some Nissans, Mazdas, etc. I'm fairly familiar with EFI control, but have no specific ACVW experience.

Cruising at a leaner AFR is fine, so the changes you specifically mention in the low throttle area should be fine. I personally wouldn't run more than about 16.5:1 in my main cruise area, but 17 is probably fine. I would say that you probably don't need to run 15/16/17 in the 4800 RPM+ range at low throttle. You will likely never use those cells, but coming off throttle from a WOT pull you might get some weird lean spikes. Maybe drop those back to 13 to match.

One concern I have is the 13:1 peak in the middle of your 100% row. Is there a specific reason you would not want a flat AFR curve on throttle? On a NA engine, it generally doesn't benefit you to run above 12.8:1 or so as the fuel mix becomes less stable as it leans out and can limit the timing advance you can run before encountering knock. The general approach is to tune in a flat fuel curve that maybe tapers a little toward redline for added safety. So, I'd suggest just running a flat 12.1:1 tapering to maybe 12:1 in the 4800 column and 11.9:1 at 5400 (or even a touch leaner with the same tapering curve). Maybe VW's are a little different, but many people like to target 12.5 at full throttle. Your head temps would tell you if the car likes it.

Also, I'm personally not a fan of sharp transitions between cells, or "block tuning." You generally end up with a much smoother running car if you smooth the maps between transitions. At WOT and cruise you will likely have many cells of the same value, but between them should be a constant, smooth transition for best drivability.

[luftvagon]

I was working under this observation:

You should be aiming to achieve 12-13:1 air/fuel ratios under full throttle, and 15-17:1 under light loads for a naturally aspirated engine. Boosted engine may require a richer mixture under power. Conventional wisdom (not always right) suggests that you tune for ~13.1:1 at peak torque, and 12.5:1 at peak power for a naturally aspirated engine. Boosted engines can run as rich as ~10.5:1 to 11.0:1 under boost. In all cases, you don't want to run lean at high loads - it can be very destructive to your engine.

http://www.megamanual.com/ms2/tune.htm

I have a new AFR table that I have generated, after testing the initial table. In practice I have found that my engine, in lower gear, at low load, does not like to run lean. There is just not enough energy to maintain momentum, which causes bucking. I corrected that, in addition to noticing my 55mph @ 4th gear falls somewhere between 3400-3800 RPM @ 85% load.

[luftvagon]

One concern I have is the 13:1 peak in the middle of your 100% row. Is there a specific reason you would not want a flat AFR curve on throttle?

Max torque: 109 ft/lb @ 3200 RPM

[TouringBubble]

You are tuning for the lean end of the suggested lambda values for gasoline, which I personally don't like to do. When you lean out the fuel, you have to run less timing to keep the burn stable and keep knock at bay. The difference in 13:1 and 12.5:1 (which I'd suggest) would maybe be 1 hp with the same timing value. Chances are you might have to pull a degree of timing to get that 13:1 and that will lose you about 3 hp or more, so a net 2hp loss, estimated.

If your goal is purely fuel efficiency, you still want to tune for the best torque output for a given rev range. Sure, you can spray less fuel and pull timing a little to compensate, but that simply means that you have to give more throttle, and more fuel to make up for the power lost when dialing back the timing.

Every engine is a little different. So, the 13:1 might work for you, but you'd be doing a disservice to not try a little richer mix and another degree of timing just to see how the car likes it.

[luftvagon]

Taking that into account, and with field testing, and some log review:

[TouringBubble]

You could still smooth the table out a lot and probably get smoother transitions between cells.

for instance, you know your cruise is around 85%. With this map, if you go up a slight hill, the load increases to maintain that 55 mph and you end up closer to 90%. Your AFR plummets to that 13:1. You could smooth that transition with, say, 15 @ 85%, 14 @ 90%, 12.8 @ 95% and the current 12.2 @ 100%. that way as you roll on the throttle you have a consistent curve downward, not a quick jump from 15 to 12.5 like in the 2100/2600 areas.

The same is true across the rows to prevent drastic changes between RPM columns. You will likely see that shift from 15 to 13 under light acceleration between 3400 and 3800.

The general idea is to think in curves and not blocks. The 3D map should be as smooth as you can make it.

Also, if you are aiming for 12.2 at peak torque, you want to go a touch richer at higher RPM, not leaner like you have (12.. Also, as VE falls off at higher RPM the load decreases, likely dropping you to the 13.0 cells in the 90% and 95% areas after 4200. You should be able to run 12.5 to 4200 then fall off to maybe 12.3 in the 100% row, but you'd want to bring that 12.3 down to the 95% row at 4800 and 5400, and down to the 90% row at 5400 just to make sure you don't jump lean.

The VE of the engine, and therefore the cells hit in this map, follow the TQ curve. As tq rises, so do the load values ... as it falls, the same is true.

[Piledriver]

You should actually run a test to see what the lean limit of your engine is.

Mine will run at 22:1 >~3K RPM, and it is nothing special, other than the Pulstar iridium spark plugs I wanted to try out.
(They do actually seem to make some difference, and have held up very well)

You are also likely to discover that injection timing matters when running that lean.

My WBO2 setup could//can see single event misfires, verified multiple times, was able to pick up on code bugs and dwell issues on the LS2 coils reliably. I'm not sure about WBO2s other than a JAW 1.03 and the SLC-OEM, they are quite fast.
.
If your VE table is properly calibrated, and injector dead times/flows are ~correct you should be able to set 18 or 19:1 (or 12:1 for that matter) in the target table and the results will match.(assuming you aren't getting a lean misfire)

If the afr results do not ~perfectly match the target, +/- ~.2 AFR the dead time or flow settings need work.
(this is actually how to test the dead time etc)

If you AFR sensor upper limit is 20:1 you will only be able to set a max target of 20:1.

You will need significantly more timing at those very lean settings.

"Step" transitions are really not that noticeable under ~83 kpa in my experience. (again, assuming you have the timing also tweaked and no misfire) but MS does a lot of interpolation of the maps, I used closely spaced table rows/columns at the transitions to get a sharp change.