View Single Post
Old 10-17-2010, 05:03 PM   #15
2010 Donating Member
skunk's Avatar
Join Date: May 2008
Location: Massachusetts
Posts: 925

2nd part

Adaptive Control

There is much confusion about how Adaptive Control works especially at WOT. Adaptive control, according to Mike Wesley and Charles Probst, is always functional – at WOT and anything less than WOT. The difference is that in open loop and at less than WOT, this feature monitors O2 sensors and writes correction trim factors to KAM, which are applied to the base table value for future calculations. This is accomplished by varying injector pulse width. The rate at which Adaptive Control writes these correction factors to KAM is slow enough to filter out short duration upsets in the mixture. Short-term closed loop corrections are written to Long-Term Adaptive Strategy KAM if they are repeatedly seen out-of-spec by the O2 sensors . At WOT the Adaptive Control still functions but it does not update (or write) correction factors to KAM for application to the base tables – it defaults to the already stored trim values, if any have been written, for the load and engine temperature experienced in the WOT load range.

So, we can deduce that if you increase or decrease fuel pressure and no target base table A/F ratio values are exceeded (high or low) then no correction trim values will be written to Keep Alive Memory (KAM) to be used as multipliers for the base air/fuel table. Also remember, the corrections are limited to about 25%, so if you change fuel pressures beyond the ability of the EEC to correct, it will not be able to tune out your fuel pressure changes. I have read posts by many people saying that they varied their fuel pressure and didn’t think that the adaptive control function took their change out. Below is a graph of Rick Wagner’s Tweecer log which shows that when the HEGO sensed a rich condition (3.2v occurring at about sample #897) the EEC adjusted Adaptive strategy to try and compensate – this shift is a little harder to see as I had to scale the right hand Y axis to include the larger HEGO signal but the shift is clear if you look closely. In the graphs below, the EEC is adding fuel if the Adaptive Strategy is above 1.0.

Closed Loop

When the EEC reaches a point where it is satisfied with all monitored parameters, it will transition to closed loop operation. In this mode the EEC tries to control air/fuel mixture to the stoichiometric ratio of 14.7:1 and accomplishes this by adjusting injector pulse width with feedback from the O2 sensors to control to the 14.7:1 ratio. This will occur when the EEC table values are calling for the ratios close to stoichiometric 14.7:1 air/fuel ratio. In warm up mode or WOT the O2 sensors are not used because they can only send signals that indicate a ratio above or below 14.7:1 and cannot read a ratio directly. You’ll see in the table that ratios in these operating areas are outside the 14.7:1 range. In graphing the parameters of thousands of data samples during closed loop operation, we have seen the EEC go into closed loop at a temperature less than 170°F since Rick Wagner sent more data from the last revision. When temperature gets to 190°F, the A9L EEC pulls 2° of timing, so a 180°F thermostat is probably the best performance choice. Here are a couple of Tweecer data graphs.

A word of caution to those who install long tube headers. In order for the O2 sensors to operate properly, they must be at, and stay at, a minimum temperature. If you place the HEGO O2 sensors farther away from the head than the stock location, you may have the sensor too far away to keep it hot enough to read correctly due to the cooled exhaust gases. This causes the EEC to enrich mixture and may cause plug fowling. Ceramic heat coating the headers will help maintain exhaust gas temperature. The EEC has a timing table used to time the sampling of the O2 sensors to coincide with the arrival of the latest cylinder exhaust pulse from each bank. Moving the HEGO sensor further away (or closer for that matter) may result in an ill-timed sample, which could also cause improper air/fuel mixtures to occur due to the O2 sensor not sampling when the pulse passes the sensor. One way to compensate for the additional time needed due to adding length in the Long tube headers is to change the timing table with an EEC-Tunerâ or TwEECerâ .

The next two graphs were from data contributed by Dave (sorry didn’t give last name) who goes by TransmarrowBird Eater on the forums. These depict two runs, one in third gear where he slows down and then speeds up and the second is of a 3rd-4th gear run. LAMSE1 is that A/F correction trim factor mentioned previously to get the target air/fuel mix and you can see it react in these graphs as well.

There are significant differences in the design and function of the 94-95 EEC as compared to the older EECs. In 1994 Ford did a redesign of both the EEC hardware and EEC software. The end result was a much cleaner, smoother running car. You may have heard of problems getting them to start, idle, and run like a 1987 - 1993 5.0 after doing some modifications. The following information was taken directly from Mike Wesley’s article "EEC vs. EEC".

"In the 1993 and older EECs, spark advance at WOT (Wide Open Throttle) was based purely on RPM. When you went WOT, the EEC jumped to a separate spark function to give what Ford thought was the best spark curve. In the 94-95 cars, Ford made a major change to the spark calculations. The WOT spark function was deleted and now the car uses the same spark tables for both part throttle and WOT spark calculations. The problem with this is the spark table is based on RPM and Load. The formula for Load is basically the amount of incoming air, ratioed against how much one cylinder can hold at standard pressure/density. You can think of Load as volumetric efficiency. The EEC uses the MAF to determine Load. It is a direct measurement of how much air is entering the engine. You might be wondering what the big deal about Load is. Well, since Load is used in the spark calculations, any change in Load will affect how much advance you get. Since the EEC uses the MAF to determine Load, any change in the MAF will change the Load calculation. Changes in the cam and anything that puts more air into the cylinders will also affect Load. Change the flow characteristics and you change Load.

When you go into WOT, the EEC will pull spark values from the top two rows of the table. In a bone stock 94 GT as RPM increases, Load decreases and spark advances a bit. The decrease in Load is due to volumetric efficiency of the engine dropping off. A typical 5.0 is not the most efficient engine around. If you compare the spark values calculated by the 95 EEC to the 93 values, you'll see they are not too much different. Change something on the engine and it's a whole new ballgame. Let's take a MAF change as an example.

An aftermarket MAF's curve doesn't always follow the stock curve. When the EEC looks at the aftermarket MAF and converts the voltage into airflow using a built-in lookup function, it will calculate differing amounts of Load as RPM increases. In our example, we'll use 5500 RPM and 4.6 volts out of the MAF. In our baseline car at 5500 RPM, the EEC calculated a Load of .78. Input that into the spark table and we get advance a bit over 26 degrees. At 4.6 volts the aftermarket MAF 'fools' the EEC into thinking MORE air is entering the engine than what really is so it calculates a higher Load. With this particular MAF on our baseline car, we saw a Load value of .91 and the resultant spark advance was 25 degrees. Wow! By changing the MAF, we lost about 1.4 degrees of spark advance!

On the dyno, this particular car lost about 14HP when the aftermarket MAF was installed. Some of this was due to the loss in spark and fuel. It looks like there is a simple fix to this by bumping the distributor up. Sounds good, and it will help get back the loss in top end spark. However, there could be a catch if we now look what happened to this car at 2500 RPM. The baseline car had a Load value of .75 @ 2500 RPM and spark was 24 degrees. After putting on the MAF @ 2500 RPM, the Load was .65, and spark was roughly 26 degrees. Hmm. We got more bottom end spark with the MAF since it 'fooled' the EEC into thinking LESS air was entering the engine. Looks good so far and the car did make more power at 2500 RPM than it did stock. Now to fix the top end spark loss, we bumped the distributor up 5.5 degrees. This included the "normal" 4 degrees everyone puts into the car plus the extra spark to compensate for what we lost with the MAF. Now the car made up the lost power plus some on the top end and lost 10 HP at 2500 RPM. Why?

Due to what the MAF was telling the EEC, and how it changed Load and advanced the spark 2 degrees, our bump in the base advance made the total spark advance at 2500 RPM a bit over 30 degrees. Way too much advance at 2500 RPM. The very top point in a torque curve is called MBT or Maximum Brake Torque spark. It's basically the amount of spark advance that produces the maximum amount of torque. If you go above or below the MBT point, you lose torque. As you continue to advance the spark, you'll reach a point in the curve called BDL spark or Borderline spark. This is the amount of spark advance where the engine just begins to knock. On a normally aspirated, low compression engine, BDL occurs at advance values higher than MBT. On high compression or supercharged engines, BDL can occur at advance values lower than MBT. Raising the octane value of the fuel moves the BDL point higher up. Going back to our low RPM example, further testing found the MBT point at 2500 RPM to be 28 degrees. Now came the question of what to do. Should we lower the base advance to get the bottom end power back and sacrifice some top end power, or leave it alone? However, should we go for the top end power, but sacrifice the bottom end? That choice is up to the owner. We opted to actually re-calibrate the EEC to run MBT spark at all RPM points by changing the spark tables. Now of course all this might go the other way depending on the MAF. It can "fool" the EEC into lowering the Load at higher RPM, which will give you more advance. It's really hard to tell what you are going to end up with.

Another thing the '94 - '95 cars do with spark is retarding it during a shift. The automatic cars REALLY pull out the timing, but the 5spd cars do it also. Inside the 5spd EEC calibration is a thing called Tip-in Retard. Any time the throttle is moved from a more closed position to a more open position, it can pull out some timing. When you shift a 5spd car, most people lift off the gas during the shift. The EEC senses this, and when you push on the gas again it pulls some timing out. The older EEC didn't have this 'feature'. You lose more torque during a shift on a 94-95 car than a 93 or older car. Why did Ford do this?? We think warranty. Ford had to replace a zillion T-5's in the older cars. A lot of them broke due to overshifting and power shifting, but a lot of them broke as a result of too much transient torque. If you could reduce the torque output of the engine during a shift, the transient torque would be lower. The trans wouldn't break as easily, and thus the Tip-in Retard was born. How much timing is pulled out during a shift varies, but it can pull out as much as 15 degrees. There is not much you can do to 'fix' this except by re-calibrating the EEC.

The 94-95 automatic cars have a torque modulation strategy installed in them to vary spark during shifts. When the EEC thinks it's time to change gears, it can pull out massive amounts of timing so the shift is nice and smooth. As far as we know there are two reasons for this. First is warranty. The AODE trans is not all that strong in stock form. By reducing the torque during shifts, you can extend its life. Second is shift feel. For some reason Ford doesn't want you to 'feel' the car shift. Ever notice that just about all Ford vehicles with electronic transmissions shift like Town cars, smooth and sloppy. Even performance vehicles like Mustangs and Lightnings have weak kneed shifting. What fun is that? The downside to this smooth, sloppy shifting is increased wear. Ford tends to slip the trans too much during gear changes slowly burning it up. Manually shift your car, and you'll see it shifts much better. During manual shifts you run through different sections of the trans control strategy. The older EECs with AODs (sometimes called DOA's) really had no idea a trans was attached to the engine. They did not spark retard during shifts and could actually shift quite well. They broke more often, but shifting was better. Adding a shift kit to a 94-95 car can help the shift feel by increasing fluid pressures inside the trans, but this does nothing to the engine torque loss during a shift.
1991 GT white/ Titanium. 5.0. mild H,C,I, built Astro T5, built 8.8 w 3.73s, FRPP 31spline termi diff, 31 splines axles, plenty of MMS stuff both suspension and braces, Bullitt springs& agx adjustables, Ford racing wheels, Dunlop star spec tires & plenty more stuff.

Her resto thread:

2011 Lincoln MKX elite

2008 Ford explorer EB

2000Ford contour. modded, dyno tunded

skunk is offline   Reply With Quote