Do you mean the sharktooths are positive? No they go from 0 to -25 so its trying to remove fuel (which matches the overly rich WBO2)

Is your MAT in your UICP or IM?

So get the Load for a given MAP vs RPM (from MAF-based-ROM) then use this value in the SDMAPVE table? in the bottom row?

So for instance if I have 40 load @ 2500rpm/55kpa what do I do with the 40 load? where do i put it?

 

Too many purple colors! I see now, yes it is sharktooth because it is trying to remove fuel and not managing to achieve closed loop, this is not normal, mine does not do this, it is usually within +-10%, often +-3%.

MAT is in my IM.

If you have 40 load with 55kPa at 2500, then your bottom row at 55kPa should be 40. If you don't have a 55kPa bin in your top row, use the nearest one and put the bottom row to 40/55*the top row.

 

ok gotcha

so tuning by free-reving to 1500rpm is ok?

How does the RPM VE affect things? ie If I am using your above method to tune the MAPCALVE map but I am hitting a non-100%RPMVE what do I need to modify?

Cheers
D.

 

Free revving to 1500 is just one of the areas to tune.

To try and simplify the VE setup, I've avoided a 3d table of MAP vs RPM as it is even trickier IMHO. However, when tuning a VE area you still need to think in these terms. By plotting MAP vs RPM vs Load/MAP I could see that most of the variation in Load/MAP (VE if you like) was due to MAP not RPM. For the RPM VE, a dip to 90% at idle and another at redline seems about right. I think the table you're needing to nail is the MAP VE calibration (which as you've found does both MAP calibration and VE). Since I think you have the code right and now the MAP calibration right, the rest is VE. It is quite low at idle and then comes up rapidly as MAP goes up. Getting this right is crucial to driveability and low throttle transients, but it seems to have been easier on my engine than yours, perhaps this is because the values were based on my logs not yours. Perhaps you'll see different numbers being required when you do your own logs. The full throttle/performance stuff is really easy in comparison.

If you are changing a MAP VE based on a RPM that doesn't have 100% VE, then the MAP VE * RPM VE = Load. Example: So if RPM VE is 90%, and you want the overall VE to be say 81% then the MAP VE needs to also be 90% of whatever MAP value is in the top row at that point.

Sounds complex, but it isn't when you get the hang of it. Helps me that I designed it

Before we change anything about the design/setup I would like to get things working for you on this setup because I have faith it will work for you

 

So if RPMVE != 100% then MAPVE = calcedVE / RPMVE?

ie if we have come up with 22 MAPVE but we are at 1000rpm (90% RPMVE), then final MAPVE should be 22/0.90 = 24.44?

 

No worries John - Once I get my head around it I am sure it will be cool

 

Yes to post #140.

 

It looks like your load and MAP are equal on the graph on the last page, both about 25. This will make it too rich. Presumably you have the same in the MAP VE table at about 25kPa? I run mine lower than that as you can see - 25kpa would give about 20 load and small positive or negative ST fuel trim.

If you're still having trouble, might be worth emailing me your actual ROM to check out the routines and tables?

Can you post up a picture of your MAP VE table?

 

In regards to RPM VE and/or map VE...I think this may help to explain it a bit (if I understand how John is coding this).

RPM VE should be adjusted for WOT conditions, mostly at higher RPMS. Your engine's natural VE curve is going to curve downward from 100% around 5000 RPM or so to some number (ie 85-90% at redline, depending on mods). I made a quick VE curve for my car in a thread somewhere.

The map VE is the much trickier and sensitive table that needs adjustment because John is using this to tune throttle plate transients. I think John and I slightly disagree on this, but are making the same point: When the throttle position is changed, the VE changes, and so does the map. So, in a sense, you can look at this as map based VE change, but I look at it as throttle based VE change. Either way, both map and throttle are changing in repsect to each other, so this table enables you to adjust the VE for this map change.

For example, if you're just cruising at 2000RPM and 15% throttle, your map may be -8psi, but if you go 100% throttle, that opening of the throttle actually changes the VE. Your map also changes from that -8psi to 0 psi, but you're still at 2000RPM, so the VE changed a lot, all at the same RPM. So, it is this change in map/throttle that needs to be accounted for in the map ve table.

At least that's how I understand John's table. Correct me if I'm wrong, because I want to completely understand this as well, when I get a chance to play with it.

 

I think you have it correct, and your throttle plate logic is appealing. We could tune the VE as TPS vs RPM. Most SD systems tune it as MAP vs RPM so I just follow this convention. It does have the benefit of not worrying about the ISCV though which would be a whole extra headache. In applying patches we have the complexity of making the patch by adding a third table vs the ease of tuning if we have a MAP calibration and a MAP VE table. As it is, I don't think though it is too difficult just to apply the MAP scaling to your top row of the MAP VE table, and then just type the load into the bottom row that you want that MAP to represent when the RPM VE table has 100%.

I would point out that the real changes in VE are considerably greater by RPM and TPS or MAP than we are having to map for. This is because the existing MAF scaling map has a huge drop at low airflows and a smaller drop at high airflows. The real VE at idle is more like 50%.

 

I agree your method should be fine. Just trying to help in the understanding.

The maf scaling map does have a nice VE curve shape to it, so it definitely helps us out. I'm wondering if the mitsu engineers actually used this table as sort of a VE table as well and not just a L/hz vs Hz table, or maf calibration.

When I did a quick log to check my VE curve, I actually caluclated a VE ~60 at light cruise (2500RPM), with very little throttle. Just after going WOT (2600-2700PM), my data showed about 85%-93% VE. During spoolup, I disregarded most of my data because of my slow logging speed. I don't think I tested idle, but I can if need be. At 7000RPM, my car showed about 87-90% VE.

 

I don't know if this has any relvance to this discussion but it seems it might:

When tuning AEMs a method called Boost Comp is generally used. A MAP(y axis) vs. rpm (x axis) table is used with what amounts to injector pulse widths in the table. For each Rpm value the pulse widths are the same irrespective of the MAP value. What it amounts to is a VE table.

Another table is used to scale the injector pulse widths based on MAP, the theory being that at 14.7 psi the fuel requirements will be double what they are at atmospheric pressure and so on.

A final table is used to modify the injector pulse widths for throttle positions less than 100%, to get proper fueling at part throttle.

I hope I explained this clearly. This method works pretty well, although I have found most times it is still necessary to tweak the injector pulse widths in the vacuum sections of the map.

 

jcsb,

Why are you circumventing the OBD baro test? The code indicates that if a value of 1 bar is hardcoded into baro RAM, the OBD test should pass.

The OBD MAF test for USDM cars is significantly different than yours. I'm looking at alternative ways to circumvent it.

 

wreckleford, lots of similar ideas to what we're using here, I had a little read of the AEM manual.

mrfred, I considered substituting the baro value but I couldn't do it by changing just one instruction around the ADC calls, so I just killed the error status but still allowed the replacement of 1 bar so that the fuel and load variables are correct. It was simple, it works for the UK cars, if you need readiness for OBD you might need to be more sophisticated.

Do you have an address for the US OBD MAF test? Plans to circumvent?

 

Hey John,

ok I fiddled with the numbers and the car feels a lot better - I am not sure if its that worthwile using the WB02 as a reference point to how the car is running.

One thing, I have a lean idle, after the idle settles and the STFT also settles its about +10%.

Now I figured I could add %10 to the load for the appropriate kpa cell, but assuming the MAPVE table is correct wouldn't it be a better idea to add 10% to the idle cell for the RPMVE table?