S40RCE:Smurf

honestly, i dont like it that much, but if it suits you its okay

orangefro© pt. II

<IMG SRC="http://www.uk-car-store.com/images/neon_interior_red.jpg">

Radak

But what was it before the Milltek?

Bob W.

exactly what I was thinking....looks half finished.

Bob W.

this guy gained 18whp simply by running the car on the dyno...both numbers are stock RS4, the low number is at the beginning of the session (9:37am), and the high number is at the end of the session (1:01pm). 4 hours of ECU 'learning' = 18whp.

<img src="http://pictureposter.audiworld.com/111032/clean_dyno_small.jpg"><ul><li><a href="https://forums.audiworld.com/rs4b7/msgs/19912.phtml">https://forums.audiworld.com/rs4b7/msgs/19912.phtml</a</li></ul>

Alan Smithee

Oh my.

edwardschmedward

I have two other cars in line for complete interiors at that shop, my '41 Plymouth hot rod, and a brand new Ford P-71 Police Interceptor... Maybe when they're done I'll send it back.

RI A6

"These cars apparently don't like the dyno ...."

"I need to drive the car so that it "learns" and adapts, then they'll dyno it."

Makes the ECU sound like it's human.


Silly me, I am an engineer, and there have to be explainations. For fun, I've been doing a bit of research on the subject of Bosch ECU's. In that context, learning and adaptation have specific meanings.

First, on the subject of the word "learning". This is used and applies to the Tiptronic ECU found in Audi's. Learned driving style shift points in the transmission. Since we have manual tranmsission, there is no transmission ECU and no driving style "learning" occurs. It does not apply to the RS4.

Second, on the subject of the word "adapt". An ECU's primary function is to control the fuel/air mixture, as measured by the primary and secondary oxygen sensors in the exhaust system (lambda sensors). Lambda = 1 indicates the correct mixture for emissions. Lambda &gt; 1 indicates a lean engine. Lambda &lt; 1 indicates a rich engine.

Under normal driving conditions, a computer controlled closed loop adjusts for changing conditions, maintaining a lambda of 1 by altering the fuel injection time. In order to do this, a pilot table of fuel injection time values is programmed in the ECU from the factory. This is a 2-dimensional table of injection time (fuel mixture quantity) vs RPM and throttle angle(requested torque). It is created based on many dyno runs, under varying load and rpm conditions, which is then statistically compressed into a 2-d table. This is known as the pilot mixture, around which the closed-loop control system makes it's adjustments continuously.

Without modifications, the air/fuel/engine environment change. Air intake can become restricted. Sensors can degrade. Air leaks can occur. Deposits can form on injectors. Air pressure can change at altitude, or just do to barametric pressures. When these things happen, the closed loop control system has to move the fuel/air mixture injection times away from the pre-programmed starting points. This leads to inefficieny and emissions.

To solve this problem, additional ECU adaptations were developed back in the late 80s. Bosch patent 4,584,982 is the seminal patent reference, which describes three additional adaptation variables. One is an additive fuel injection time constant that compensates for mainfold air leaks between the mass/air sensor and the intake. This adaptation is only used at idle and low engine loads. When the ECU finds that the average value of the injection time diverges significantly from the table values at low loads, the delta time difference is computed and stored in memory. This is retained during engine shut down, and is sometimes referred to as "additive fuel trim." This adaptation value is only used at low engine loads and low rpms, under normal driving conditions. It is designed to allow the ECU to predict the correct injection mixture and eliminate most emission problems. THIS ADAPATAION VALUE HAS NOTHING TO DO WITH MAXIMUM ENGINE POWER.

Injector wear and deposits can effect the amount of fuel delivered by injectors. As injectors wear, less fuel is delivered for a specific time of opening. A second additive adaptation constant is computed to compensate for changes in mixture from the pilot baseline, due to injector deposits and wear. This adaptation constant is also computed and operates under low load conditions, but high rpms, where the effect is signficant,and is stored as a delta value. The injectors are left open for a longer time period to compensate for decreased fuel flow, to help the closed loop controller adjust quickly during low load acceleration. THIS ADAPATAION VALUE HAS NOTHING TO DO WITH MAXIMUM ENGINE POWER.

Finally, errors in the mass/air sensor due to sensor degradation, deposits, changes in atmospheric pressure, cause a multiplicitive change in the air/fuel ratio. This divergence causes a change in slope of the measured air fuel mixture (lambda) with respect to RPM and throttle angle (torque). This change is only easily measured at high load conditions (high torque). When the ECU detects long term changes in the closed loop control value, w.r.t. the pilot table values, it computes a multiplictive correction factor and stores it. THIS IS THE ONLY ADAPTATION VALUE THAT HAS ANYTHING TO DO WITH MAXIMUM ENGINE POWER OUTPUT.

Closed loop control, to keep the air/fuel mixture at Lambda=1, is performed continuously. When load or RPM changes, the pilot lookup tables are used to determine the correct initial injection time. To to this, the adaptation values are either added to or multiplied with the pilot table values, depending on rpm and throttle angle. Then, the closed loop contoller tries to control for optimal lambda=1 and makes minute corrections to the injectors. But, if it has to correct beyond the average value computed, and does this for a long period of time (10 minutes), then new adaptation constants are computed and stored.

For lambda adaptation to occur, there has to be a change from the stored values for a long enough time, to trigger a new computation. For low load additive adaptations, this is easy, since these conditions occur continuously, under normal driving. However, the multiplitive adaptation that accounts for changes in air density, and therefore torque, can only occur when a torque threshold is reached, above normal cruising torque values. (Unfortunately, what this threshold is, is hidden in the code implementation, and is not specified.)

Okay, so to trigger adaptation, you have to be in the torque and rpm range where the adaptation occurs, for a long enough time for the software to filter the deviation and "notice" that a shift has occured. For the additive values this should occur easily, within the first 10 to 20 minutes of driving. For the air mass multiplier adaptation, it is not clear where the high load threshold is for adaptation to be triggered. And, this is the real kicker, the algorithms assume that the change in air mass is linear with rpm and load, that the problem is with the mass/air sensor, and not some flow restriction in the rest of the intake or exhaust system. (Remember that the initial pilot tables were already designed to take into account any non-linearities in flow do to the intake and exhaust systems.)

If an new intake or exhaust system changes the nature of flow significantly, there will be a deviation between the shape of what the fuel metering should be, and what it is in the pilot table, which no single variable can compensate for. Remember this little factoid.

So WTF does all this have to do with exhaust systems, maximum power and dyno readings????

Well, first, you have to understand what happens when you apply full throttle. Although an air fuel mixture of Lambda = 1 is optimal for emissions, it is not optimal for power. It turns out that maximum power is produced when there is 10% to 15% excess fuel over the optimal mixture for emissions. So, when you go WOT, 10% to 15% additional fuel is injected into the combustion chamber. But, there is no closed-loop correction applied, since the oxygen sensors and the controller are not designed for this. Instead, at any particular RPM, the pilot injection time values from the table are pulled and increased by 10% to 15%. Cool, the controller knows what the optimal mixture is and just enriches it. But, the assumption is that the value in the table, along with the adapatation multiplier, is correct. If it is not, then the WOT injection times will be wrong, and the engine will not make peak power.

This is important. For an engine with a Bosch ECU to make maximum power at WOT, the adaptation multiplier must have been computed and close to the current flow and atmospheric conditions. Also, since there may be non-linearities in the intake and exhaust system, the adaptation value should be computed near the top of the torque curve, and not near the bottom or middle. This means that you have to run the engine just below WOT for a long enough period of time to allow the ECU to adjust.

So tell me guys, how many of you have the ability to run an RS4 at 90% throttle for more than a a few seconds at a time? And how many of you opt for WOT instead? If you only go from low throttle to WOT, you will NEVER cause the adaptation multiplier to be computed. The ECU will use whatever is currently in the table, add a fuel enrichment constant to it, and call that WOT. OOPS.

So, on the road, the solution is to not run WOT, but to run at just below WOT, as much as you can. IF you can do that going up a hill, that's great, otherwise, you can find a lonely highway, slow down untill traffic is way ahead, and put the hammer 90% down. Do this for enough runs, and the ECU will adjust correctly. Then, once it's done, you'll notice that the difference between 95% throttle and WOT is a boost, not something that feels like mud.

So why do dynos have problems with the RS4. My theories.

1) Adaptation at high load was never done on the street.

2) Air temperature is different in dyno room requiring additional adaptation.

3) No one runs non-WOT on dyno, so adaptation never occurs.

4) on inertial dynos, it is impossible to develop a long enough run below WOT to cause adaptation to occur ... this should not be a problem on dynos like the Mustang which place a load on the axles.

JoeyJoJoeJrShabadoo

&lt;head asplode&gt;

Joey T

Sweet Accord!