Is it true that after installing a Y-Pipe or CAI it takes your ECU 200 miles worth of driving to recognize them and accomodate the changes needed to make them actually increase power? I heard this somewhere and wasn't sure if it was true or not.

Thanks in advance for any feedback!

 

No, it is not true

 

Why's that Mike? I've heard the same thing too from multiple sources that the ECU takes about that long to adjust to the mods done to the engine.

 

I don't know about closed loop operation, but for open loop operation (above 50% thorrle or so) the ECU is programed to run on set maps. The MAF meters the air entering the motor and the postition of the TPS and then supplies X amount of fuel. You mod your car to allow more air to enter and the ECU simply adds the programed amount of fuel. It is very simple. Nothing to learn. The ECU will keep doing this till either the MAF or injectors and maxed out.

In addition, I also know this because Nealoc187 dynoed with different ECU's back to back and got the same #'s. So if there was any "learning" going on that would not have happend.

 

Because if you took a bone stock max, dynoed it - then installed y-pipe, and dynoed it again (without driving it) - it would make more power. The ecu adapts over time, but you still get immediate results.

 

This is very interesting, because right when I installed my Y-pipe it didn't seem to pull all that much harder.

A week or so later, I noticed it pulled a lot harder after 4000 rpm.

 

It was either just in your head or it could have been different weather conditions, gas, etc...

 

Ok, i was just wondering because after doing minor mods to this car and my friends cars, i've noticed the difference in power only after a week or more of driving. But it does make sense what you said. But this being said, when you buy a car off of an older person who babies the car and does their Sunday driving and thats about it, then why does the car have to readjust to your driving habits?

 

Hmm...interesting.

Thanks guys!

If anyone else has any more info, it would be greatly appreciated!

 

Who said it adjusts??? What I think is funny is those who say the ECU has to "learn" or the car "adjusts" but then offer no explanation...just that it does, or that they thought the car felt faster/slower. Those are not reasons, those are feelings.

 

well said..

 

In addition to having tested different ECUs back to back to back (one in the car for a few thousand miles, and then one in the car for like 4 minutes) and seeing no difference between the two (evidence that ECUs don't learn, or adjust, or do any other sort of magical trickery) I installed a Ypipe in the morning, drove 30 miles to the track, and made passes, gaining 3-4mph right off the bat. ECUs don't learn, or adjust, or have artificial intelligence, or come from the 19th moon of Jupiter, or any of the other stuff that people say they do.

 

Im With No Ajusting Theory Its B.s Look At All Car Magazines They Put On Different Exhaust Intakes All That Stuff And Switch It Back And Forth And Dyno The Car Immideatly It Would Make No Sence Those Guys Know What They Are Doing And Idont See A Reason For An Ecu To Adapt....

 

While the car may not adjust it's A/F tables themselves, doesn't it have a conservative map for engine knock?

If so, it does learn and adjust...just not to the extent that everyone claims.

(sorry...just nitpicking )

 

That not the same. It is actually called the high and low octane program...why exactly I don't know. I have seen the bin files on a lap top before and there is two seperate maps.

I guess you could say the computer will "adjust" to different altitudes as well since it automatically compensates if you drive to Colorado. But that is not what we are talking about here. What is being claimed is that the ECU takes time to recognize a mod and the full power is not realized until that happens. I see it posted all the time and it is plain not true. Same as with resetting the ECU....some will say to reset the ECU after a mod so it can "learn" it. Total BS. In open loop operation our computers simply supply X amount of fuel for X amount of air at X TPS position. More air = more fuel = more power.

 

don't you think if our ecus could "ajdust" over time to differenct situations, we'd be a lot farther along with ai programming? especailly give a couple years of cars and modding being around...

so mike, the fuel amound (f) is directly proportional to the air intake (a)? what is the constant of variation? j/k im about to go to algebra...

 

Why would you put an adaptive ECU in the same category as AI? We're not talking about something that learns new concepts here...just changes in the engine's behavior given certain situations. In a limited sense, our ECUs do this already (uses a different a/f map for altitude as Mike said). Would you consider this AI?

Adding a new mod is detectable by an ECU, so there could exist an ECU that adjusts appropriately. Same goes for driving style.

Let's pretend we could make our own ECU easily...we could simply keep track of throttle position over the last X hours of driving...and maybe use a different a/f map or adjust timing based on the aggressiveness of the driver. It could be done very easily with today’s technology. Now getting an ECU to truly 'tune' an engine (totally adapting) might be getting closer to AI.

 

i know that ecu doesn't adapt... mike helped a lot on that as well... i was talking about others thoughts...

when more air comes in, more fuels is applied... i got that...

i am in no means "classifying", and i am in no way knowledgable on this matter, but from what i understood, i made that comment.

i know computers are programmed with set instructions, but do you not think it is possible for a throughly written program to adapt, if thats what its programmed to do?

like in irobot? not saying they will take over the world, but lets say i wrote a program... as a small example.

it would allow you type in your name, and birthdate... it would then, add the data to a vector, array, or file, whichever... it would then ask you to enter your name, and it could derive your birthdate from the vector...

in a sense, no it's not learning, but it does store more knowledge to call on later...

that's why i associated it with what others thought...

 

That's one definition of 'learn'...gain knowledge.

There are adaptive algorithms for chess that 'learn' and make better moves based on games it's played. Adaptive algorithms are not that difficult to write and work quite well (I have written quite a few myself).

 

Its most absolutely is 100% true. Here is my proof: I failed for knock sensor/ emissions on my old 92 se. I installed the KS myself and immediately went to get the car stickered (after a 10 mile highway run a 75mph) . I still failed even with the check engine light now off, 60 days later when my rejection was due I went again and passed with flying colors. I Cut my co2 nox2 inhalf. I did nothing to the car since the KS not even a oil change just gas. Explain that

 

i hope to be where you are in a few years... i love chess, and looked into chess programming, algorithms and whatnot... phew! i have a hard time understanding binary trees... but that was like a year ago when i touched c++...

 

Emission testing is carried out at a very low throttle opening which is closed loop operation.

Everyone should go out and buy themselves an OBD-II scanner so they can see for themselves what the ecu is really doing.

 

1. That does not prove anything

2. We are talking about open loop operation, ie; flooring it...otherwise the point of going faster at 15% throttle is pointless

 

Thank you for correcting me.


according to nissan performance magazine:

....OBDII codes in the maxima for emissions use remain on until the system that failed passes a self-test on 3 consecutive trips....

Sorry my friends but that sounds to me the ecu adjusts to the changed dynamics of the engine within the stock parameters.

Who knows maybe next somehow the ecu in your maxima will electronically withdraw money from your checking account and donate it to maxima.org.

 

All it does is respond to given parameters per its programming. It's not learning anything new when it does that. It responds with more fuel when more air is present (mods), etc.

 

When you PHYSICALLY deplug the intake or PHYSICALLY deplug the exhaust you are making it much easier for the engine to bring in air, and also push it out. That means instant gains in terms of reduced pumping losses which is something you'll see instantly, regardless of what any ECU is doing in the background.

 

damn...i never knew mike was so damn smart.

 

Why does it take so long for a ecu to lower emission on a vehicle? I cant figure out why putting in new sensor (inconjunction with resetting the ecu) doesnt solve the problem instantly.(or enough to pass inspection) This still perplexes me...although I am impressed with everyones answers

 

I don't think you understand much of what is going on here. The fact that the ECU checks its systems means nothing. That is its purpose.

The knock sensor, along with many, many other sensors is monitored by the ECU. It knows if the knock sensor is working or not. The ECU does not know there is a y-pipe...it has no idea what a y-pipe is...so it cannot learn a damn thing about it.

The ECU does not know what an air filter is. The is no "air filter sensor". If you have a dirty air filter you do not get a CEL. You change to free flowing air filter and a little more air enters the cylinder. The ECU cannot do anything about that. It cannot resrtict the air flow. All it can do is change the amount of fuel that is added in porportion to the increased amount of air. If if did not do that then you would run lean would you not? It is not like the ECU say "Whoa..hold on here. There is some fancy new air filter on the car. I better not let so much air into the cylinders. I need to take some time to adjust to this. I will slowly let more air in so that I can adjust to this new air filter."

 

Nope. It's just reporting whether there's an emissions related problem or not. It's not "changing" anything.

 

Wow, thanks for all the info guys!