[birdthunder]

First off, I should say I'm using a CDAN4 strategy EECV. Parts of this probably don't apply for other strategies. The below is my understanding of what is going on, it is not necessarily correct - that's why I'm posting it!

Ford 4.6 2v, PI-head motor, SVO supercharger and methanol injection


I have scaled my tune because my injectors are 'too large'. CDAN4 can't handle 40+ lb injectors. To resolve this, my injector parameters, MAF transfer, and SARCHG are all scaled to 50% (maintaining load, etc. while bringing my injectors under the clip limit).

The best document I have to reference for information is the CDAN2 document from ford (CDAN2.txt/pdf). I figure its probably close enough to my strategy to still apply. To calculate filtering coefficients, CDAN2 does something similar to (this only applies under certain conditions, the calculation can change for tip out, idle, etc):

Code: Select all

AIR_NUM_MULT = (11781.34/AIR_MAN_VOL) * FN1052(ACT,ECT)      
AIR_B3_N = FN052N(N)
MAP_AIR_0 = FN052M(N) * FN052BP(BP)
air_k_base := AIR_NUM_MULT / [AIR_B3_N + (MAP_AIR_0/AIR_CHG(old))]
AIR_FK := air_k_base

AIR_FK is used in manifold filtering logic AND anticipation logic to determine final, actual air charge value (and fuel injected).

AIR_NUM_MULT is determined by the 'manifold volume' scalar, AIR_B3_N is FN052N, and MAP_AIR_0 is proportional FN052M.

Note: I have BP in my tune dummied out / locked to 29.88. I use PRLDSW = 0 and the BP term was messing up my PERLOAD numbers.

Code: Select all

PERLOAD = [LOAD * 29.9] /  [FN035(N) * BP]

So, AIR_CHG is going to be 'half' of its physical value. In order to maintain the rest of the equation, I have halved MAP_AIR_0 (it appears you can halve either FN052M or FN052BP to achieve this; with dummied out BP I don't think either affect anything else). By doing this, a real-life air charge should take on the same filtering coefficient as it did before the tune scaling...I think.

I think this is more correct than scaling ONLY manifold volume, as some have done before - if you ONLY change manifold volume, than you don't end up with the same air_k_base as before. If you only change either FN052BP or FN052M...its okay?

Code: Select all

air_k_base := 2*AIR_NUM_MULT / [AIR_B3_N + (MAP_AIR_0/AIR_CHG(old))]
                  = AIR_NUM_MULT / 0.5*[AIR_B3_N + (MAP_AIR_0/AIR_CHG(old))]
                  = AIR_NUM_MULT / [0.5*AIR_B3_N + 0.5*(MAP_AIR_0/AIR_CHG(old))]
                  = AIR_NUM_MULT / [0.5*AIR_B3_N + (0.5*MAP_AIR_0/AIR_CHG(old))]

air_k_base := AIR_NUM_MULT / [AIR_B3_N + (0.5*MAP_AIR_0/AIR_CHG(old))]

Remember that AIR_CHG(old) will be half of its value after scaling. The bottom equation (scaling MAP_AIR_0) cancels the scaling out within the (0.5*MAP_AIR_0/AIR_CHG(old)) term, not requiring any other changes. The top equation (scaling manifold volume only) does not yield the same result (the 'halving' bled over to AIR_B3_N in addition to the MAP_AIR_0/AIR_CHG(old) term, more than we intended), meaning that filtering constants will NOT be the same as before scaling the tune.


My results in doing this have been positive. Between performing this and halving other mass unit values in the tune (transient fuel mass, etc.), tunes seem to act nearly the same before and and after scaling.

Does this seem right to anyone? I'd love to hear feedback / comments / experience. Do other strategies work similarly?


My other question - does anyone have any information on how to actually tune FN052N and FN052M? I can't find any good explanations online of how to empirically gather corrections to either of them. I'm using the stock NPI 2v head values and I think it contributes to fuel error at times (I have 2v PI heads). Could anyone post these functions for a 2v PI head car? I found some spreadsheets from the user mitchb with 2004 2v engine values, but it doesn't list the units (in case he see's this - what are the units used in the file MAP_vs_airchg_19p14k.xls that you posted?)

Really, any information about manifold and VE modelling is welcome, especially for CDAN4 EECs
Thanks guys

[Mitchb]

Map vs Aircharge, FN052N is:

slope of manifold absolute pressure vs air charge and is in units of

inches Hg/(1 lbm/intake stroke).

In your work, remember the following:

Inferred MAP is defined by the linear equation y = mx + b

where m is the slope of map vs air charge, x is air charge and b is the offset.

If AFR is either rich or lean from low to high MAP, then you should increase or decrease FN025N.

If AFR varies as a function of MAP, then FN052M should be made smaller or larger.

I would not start out by changing the offset FN052BP. I would turn off anticipation logic from the start. When you do this and with your scaling, you will need to look at all of the factors that affect transient fueling, including manifold volume, FN1321 - equilibrium intake surface fuel, FN1322AL and DL - time constants for acceleration and decel, FN312S - change in transient fuel puddle mass per change in aircharge and the EISF RPM multiplier FN313 I think. These all interplay with each other and you'll have to invest some time to get a feel of what works. Plot all the tables in excel graphs which will help you better visualize what's happening. 99% of those trying to tune transient fueling are really looking for a simple fix, but this is pretty much involved.

Post back with your progress.

Good luck!

[Mitchb]

These are the stock values for FN052N in excel for a 96 Thunderbird 4.6 NPI.

MAP_vs_airchg_Stock.xls

(31 KiB) Downloaded 126 times