Are computers fast at learning varying octanes?

[mjoekingz28]

Let's say you run 93 for 10 or a hundred tanks and then start pumping 87. I have kind of noticed the engine can become used to certain things (like wear rates and grooves.... New owners and tossing on a bunch of new parts at once). I dont really know why or how, but with modern fuel injection, do you think changing things up often is okay- for experimenting for the best setup or is a wear in (aka find something, stick with it, and then leave it alone) the best. Ex. Buying a little old lady's grocery getter for a teenage boy.

 

[Rand]

They can learn you put in lower octane fast via the knock sensor, not so sure about higher octane.

Probably different on every car model/year.

Something like a ecoboost turbo gdi escape that requires 87 octane is going to be different from a
80000 bmw that requires 91

 

[WyrTwister]

As far as octane rating , here is a guess .

The computer does not know what the octane of your gas is . Most engines , I think , have a knock sensor . I further guess the computer runs the max timing advance & maybe cam timing , up to the point where it starts sensing a knock ( fuel mixture may be adjusted too ? ) .

The computer probably also uses information from the TPS , Mass Air Flow , coolant temp , outside air temp and probably other sensors that do not come to mind .

At that point , the computer starts tapering some or all of the variables back , until the knock goes away . I am sure the computer stores those parameters for future use .

Now you fill up with lower octane gas . I suspect the computer goes through this learning and adjustment process again .

I do not know how long this takes and / or how often the computer " tests " the current conditions .

All of this assumes the engine has warmed up to the point where the oxygen sensors are functioning properly & the system is in closed loop .

As far as the rest of your question , I have no answers .

 

[mightymousetech]

Originally Posted By: WyrTwister
As far as octane rating , here is a guess .

The computer does not know what the octane of your gas is . Most engines , I think , have a knock sensor . I further guess the computer runs the max timing advance & maybe cam timing , up to the point where it starts sensing a knock ( fuel mixture may be adjusted too ? ) .

The computer probably also uses information from the TPS , Mass Air Flow , coolant temp , outside air temp and probably other sensors that do not come to mind .

At that point , the computer starts tapering some or all of the variables back , until the knock goes away . I am sure the computer stores those parameters for future use .

Now you fill up with lower octane gas . I suspect the computer goes through this learning and adjustment process again .

I do not know how long this takes and / or how often the computer " tests " the current conditions .

All of this assumes the engine has warmed up to the point where the oxygen sensors are functioning properly & the system is in closed loop .

As far as the rest of your question , I have no answers .


Yup. The BMW tests the fuel every time it sees the fuel level rise, and then assumes it does not change. Can be problematic at times, like when you store your car for the winter and use Seafoam as a fuel stabiliser. Doh. Learned that one the hard way. Ran like a bag of [censored] until I got some fresh fuel in there.

It will take advantage of much higher octane race fuel, but takes several WOT runs before the computer adapts.

 

[VeryNoisyPoet]

The computer may relearn faster if the battery is unhooked for a few minutes, since this will make many settings revert to default values.

My car took the better part of a tank to wake up when I started using 89 instead of 87. Going to 91 didn't seem to have much of an effect because I only run a few psi of boost.

 

[Yah-Tah-Hey]

Originally Posted By: WyrTwister
As far as octane rating , here is a guess .

The computer does not know what the octane of your gas is . Most engines , I think , have a knock sensor . I further guess the computer runs the max timing advance & maybe cam timing , up to the point where it starts sensing a knock ( fuel mixture may be adjusted too ? ) .

The computer probably also uses information from the TPS , Mass Air Flow , coolant temp , outside air temp and probably other sensors that do not come to mind .

At that point , the computer starts tapering some or all of the variables back , until the knock goes away . I am sure the computer stores those parameters for future use .

Now you fill up with lower octane gas . I suspect the computer goes through this learning and adjustment process again .

I do not know how long this takes and / or how often the computer " tests " the current conditions .

All of this assumes the engine has warmed up to the point where the oxygen sensors are functioning properly & the system is in closed loop .

As far as the rest of your question , I have no answers .
I think you are right twister. If the 87 octane doesn't cause the engine to "knock," the computer won't know what fuel you are using.

 

[WyrTwister]

Back in the day , I had a 66 model Comet with the FOMOCO 289 2 barrel & 3 on the tree .

Timing was set by rotating the distributor . I ran the timing advanced to the point where it knocked / pings just a tad , when I was getting on it in 3rd gear .

Fast forward to the present , I grive like an old man , because I am and old man , trying to stretch MPG's .

Under those conditions , the computer may have all the perimeters advanced all the way , or nearly . Who knows .

I am also cheap and buy 87 octane gas .

 

[SHOZ]

It depends on the programming for the timing and whether the knock sensor is used as a continuous feedback loop or not. On the late 90s fords there was a table that would be used for decay when knock was sensed and one to rebound after the knocking stopped. Rebound was quite slow but within a couple minutes.

Newer ECUs such as my Gen Coupe will actually boost timing until knock then stay near that value so it almost instantaneously works with a higher octane. It's just a matter of the time values and degrees added or subtracted.

 

[WANG]

Originally Posted By: SHOZ

Newer ECUs such as my Gen Coupe will actually boost timing until knock then stay near that value so it almost instantaneously works with a higher octane. It's just a matter of the time values and degrees added or subtracted.


This is consistent with my research and observation.

My truck will rattle away on 87, and generally run flaky, but is still "operating as designed". It runs much better on 93, counter to what I've (sometimes passionately) told people for years concerning selection of octane rating with respect to owner's manual recommendations. It is my belief that the engine controllers of the latest models have much more aggressive calibrations/strategies than in years past.

 

[badtlc]

Yes. They had to when they started producing flex fuel vehicles.

 

[Silk]

A lot of vehicles will show a percentage of knock retard on a scanner, you could use this to compare fuels. My car doesn't show this, and it's the only one I change fuel on.

 

[CrAlt]

I think this is going to be a vehicle specific thing.

My Jeep is already set in its ways. It learns and remembers the fuel/timing values during the 1st 50 warm up cycles. But it will back off timing if it sees knock. So if it learned on 93 and you go to 87 it will see knock and pull timing. But if you learn on 87 and go to 93 it will never know to add timing. There is a procedure to "reset" the computer and make it relearn how much timing it can get away with.

On a L98 vette i had the computer tested the fuel every single time you started it. The 1st time you jammed the throttle you would hear crazy spark knock for a second then never again until you restarted the car. So going between 87 and 93 or even race fuel had an instant effect.

 

[MNgopher]

The Ecoboost engines in the F150 do adjust rather quickly to changes in fuel octane. You can see this by monitoring the Octane Adjustment Ratio in Torque. Typically within 5 miles of filling up with a different octane, the OAR will adjust up or down accordingly. The OAR varies from 1 (bad) to -1 (good). Running an 87 octane E10 fuel will typically net an Oar of 0.2 to -0.2. I've seen as high as -0.9 using shell 91 octane E10 fuel.