openecu.org

I've been trying to modify my A4SGE01C ROM to change injector and MAF scaling, but I've been having some issues...

Since I run a Perrin BigMAF, I computed updated values for the MAF curve by multiplying the air flow values by 1.16. Also, I increased my injector value to 800CC (running modified WRX injectors).

My problem is when using an image modified with these two changes, my car runs exceptionally lean. It idles around 16.5:1 - 17.1:1 AFR

I ASSUMED modifying these two areas would be enough to allow the ECU to properly compute its fueling, etc., but based on my experience, it seems like I'm missing something. Any advice?

Log your fuel trims. They're probably maxed out positive. I saw the same behavior when I was experimently with injector scaling on my car. Too large, idle would never go richer than 16-17:1 and fuel trims maxed out. Too small, opposite problem. This is just with injectors and a stock sized (but aftermarket) intake.

But MAF scaling can do the same thing. 16% doesn't sound like nearly enough scaling for BigMAF. I would think more like 35%. If the stock intake is about 3" diameter and the BigMAF is 3.5", that's a 36% increase in cross-sectional area.

Freon wrote:Log your fuel trims. They're probably maxed out positive. I saw the same behavior when I was experimently with injector scaling on my car. Too large, idle would never go richer than 16-17:1 and fuel trims maxed out. Too small, opposite problem. This is just with injectors and a stock sized (but aftermarket) intake.

But MAF scaling can do the same thing. 16% doesn't sound like nearly enough scaling for BigMAF. I would think more like 35%. If the stock intake is about 3" diameter and the BigMAF is 3.5", that's a 36% increase in cross-sectional area.

Thanks for the advice. After reading your input, I re-computed the numbers I've been using for MAF scaling from scratch...

Stock WRX intake: 60.0mm
BigMAF intake: 76.2mm (assuming BigMAF is 3 in.)

Using these numbers, it turns out my scaling value should be 27%.

(76.2/60)^2 = 1.61 or a 61% increase in area...

cboles wrote:(76.2/60)^2 = 1.61 or a 61% increase in area...

Okay, I see where you getting your numbers from... I didn't square my results. Now, the million dollar question...

When I'm attempting to scale the air flow values corresponding to a given MAF voltage, should I multiply by the increase in diameter (1.27) or by the increase in area (1.61)?

Sorry if this is a stupid question, but I've been trying to figure this out for a while now, and I keep getting different information.

rewt wrote:

cboles wrote:(76.2/60)^2 = 1.61 or a 61% increase in area...

Okay, I see where you getting your numbers from... I didn't square my results. Now, the million dollar question...

When I'm attempting to scale the air flow values corresponding to a given MAF voltage, should I multiply by the increase in diameter (1.27) or by the increase in area (1.61)?

Sorry if this is a stupid question, but I've been trying to figure this out for a while now, and I keep getting different information.

MAF works by seeing how much current is needed to keep the wire at a target temperature. For a given cross section, that can scale to quantity of air. It is (largely) pressure/temp compensated, but can be affected by poor airflow conditions. Anyway, figure that the maf is really looking at a small cross section of the flow, and assuming that the flow is similar all over the tube, and the tube equals to xxx times the sampled area to get total mass. As such, you definately need to look at change in area, not diameter/radius.

In reality, there are so many factors at play here:

* flow profile through the pipe
* what part of the profile the MAF occupies
* turbulence / inhomogeneous flow issues
* MAF gain / linearity

The area calculation is just a first stab at the problem. One way to empirically calibrate the MAF would be to use a known, calibrated injector (e.g. the stock one), set the AFR target to be 14.7 everywhere, and then on a dyno move the engine through a range of loads (by varying RPM really) and adjust the MAF calibration for the current MAF voltage such that the fuel trims are zeroed out.

rewt wrote:Stock WRX intake: 60.0mm
BigMAF intake: 76.2mm (assuming BigMAF is 3 in.)

Where did you get those numbers?

Jeff Sponaugle at PDX reports different diameters in an article he wrote a few years ago. A copy is at http://www.pdxtuning.com/technical_big_maf.htm. He said the WRX MAF tube has an inner diameter of 2.60 in. (66.04 mm) and a cross section area of 5.31 sq in (3425.8 sq mm), while the Big MAF tube has an inner diameter of 3.09 in (78.49 mm) and a cross section area of 7.5 sq in (4838.1 sq mm).

So that means the cross section area of the BigMAF is 41% (7.5/5.31=1.41) larger than the stock WRX tube and the additional air flow at a particular MAF voltage should also be roughly 41% higher for the BigMAF. Inversely, for a particular air flow, the BigMAF should result in a MAF voltage that is roughly 29% (5.31/7.5=0.71) lower.

Jeff confused himself about that last issue in his article. However, his data lead him to the correct conclusion:

First, the average reduction in measure flow from real flow is between 20 and 30%, not 41% as calculated.

Second, the additional correction is much less at lower flow then at higher flow. This explains the observations that an offset in injector size is not enough to correct for larger MAF setups.

I think you need to first re-scale the injectors using the stock MAF tube. Then, once that's been nailed down, re-scale the MAF voltage table for the BigMAF. I think it's impossible to try and do both at the same time.

Thanks, Jon.

That's a lot of food for thought, and I guess I need to plan my attack differently. Unfortunately, running a stock MAF tube isn't an option as I got rid of it long ago.

Why would you make large and broad changes to the injector scaling *AND* the MAF table?

Seeing as how (I think) the car should idle OK with a stock ROM + BigMAF + modded injectors, why make huge changes to the injector scaling? I'd consider tweaking the injector scaling slightly from the stock value so that the car idles with good fuel trims, then adjust the MAF table to get you in the ballpark for everything else (cruise and WOT). Use the fuel maps for your fine tuning.

When screwing with the MAF table, be aware of how it affects your ignition timing.

x99percent wrote:Why would you make large and broad changes to the injector scaling *AND* the MAF table?

Seeing as how (I think) the car should idle OK with a stock ROM + BigMAF + modded injectors, why make huge changes to the injector scaling? I'd consider tweaking the injector scaling slightly from the stock value so that the car idles with good fuel trims, then adjust the MAF table to get you in the ballpark for everything else (cruise and WOT). Use the fuel maps for your fine tuning.

When screwing with the MAF table, be aware of how it affects your ignition timing.

Well, I was interested in changing both attributes, so the ECU could use the information to properly compute fueling, etc., and so I could turn off injector scaling on my UTEC.

rewt wrote:Well, I was interested in changing both attributes, so the ECU could use the information to properly compute fueling, etc., and so I could turn off injector scaling on my UTEC.

Like I said, there is no need to make large changes to BOTH. When you do, you end up with this:

rewt wrote:My problem is when using an image modified with these two changes, my car runs exceptionally lean. It idles around 16.5:1 - 17.1:1 AFR

;)

Remember that one of the things the BigMAF does for you is "rescale" things to work better with larger injectors, so that you *don't* need to make large corrections (with respect to fueling ONLY)... but you're trying to "undo" that in the ROM, so that you need to make large corrections! Stop!

What you're trying to do:
BigMAF skews MAF -> rescale MAF +20-30% -> rescale injectors +100% -> add LOTS of fuel because the car idles lean??

My suggestion:
BigMAF skews MAF -> slightly rescale injectors -> tweak MAF table -> fine tune fuel map

I believe cross sectional area is probably the best first guess as well. Diameter definitely does not make sense to me.

In reality you can use injector scaling to compensate as well, but your loads are all going to be off, so it will affect your timing. That's bad!

My procedure is this:

1) Install new injectors with a factory intake. Scale injector size to minimize fuel trims. I would say cruising and light to moderate, constant acceleration are the best areas to check as opposed to idle (idle tends to vary anyway just based on how warmed up the car is, weather, etc.). Check your fuel table against logs and see how accurate it is. It should be darn close even in open loop and really high load with the stock intake.

2) Install aftermarket intake. Scale the whole MAF scaling table at once based on cross sectional area. pi*radius^2

3) Scale MAF scaling point by point to make your fuel table match actual AFR logged. This is time consuming. Once you have a good curve you can probably reuse it in other cars. The MAF scaling should be "attached" to the intake. The curve should remain pretty smooth and still retain the approximate exponential curve shape of the stock system, but really, who knows on some of these intakes how the air will flow at varying velocities.

I really believe in using MAF scaling to make your fuel tables match actual results. This gives you a reasonable belief that the loads and airflows seen by the ECU are actually correct, or at least correlated correctly even if they're off by a fixed %.

If you have to put 8:1 in your fuel map to hit 10.5:1 @ 4.0-6.0g/s @ 5000-7000rpm but your fuel trims look fine in closed loop, your MAF scaling should probably be ramped up faster at high voltage.

I don't expect it to be perfect, or your fuel table to ever exactly match actual results. But you should be able to get close.

Freon wrote:In reality you can use injector scaling to compensate as well, but your loads are all going to be off, so it will affect your timing. That's bad!

This is true for injector scaling in the UTEC (skews the MAF signal), but not for the injector scaling in the ECU.

No, it is true of the stock ECU. If you try to use injector scaling to compensate for a BigMAF for instance, your fuel table might be close but your timing is going to be off.

If you get a large intake, you need to compensate for it with your MAF scaling, NOT injector scaling.

That is, 4.0V on a factory car MAF scaling table is, say, 230g/s. This should be reasonably "true." Now with a BigMAF, if you just scale your injectors, 4.0V still means 230g/s to the ECU for timing purposes, even though it may be more like 290g/s in the actual physical world. Understand?