What are flex fuel systems doing?

So, I'm in the process of researching ethanol for my f150 (2010 4.6 2v). I know lots of people talk about changing stoichiometric AFR when running fuels like E85. However, my truck came stock with 2 wideband sensors which read lambda, which to my knowledge doesn't rely on a preset stoichiometric ratio due to measuring oxygen content in the exhaust. Within HPTuners I have a master enable/disable for a flex fuel module which other models of my truck have.

What are these flex fuel systems doing? Is it mostly timing changes? Or is there stoichiometric settings HPTuners can't access on my ecu? Would an aftermarket flex fuel kit have any fundamental differences from an OEM system?

I haven't used the HP TPTuners software, but having been involved in the development of the Flex Fuel system for a major ECU manufacturer I have an idea of what is happening generally.

The ECU will usually have internal non linear translation tables that range between 0% and 100% Ethanol, and you will also have Ignition Timing, Boost Aim, Lambda Aim and (potentially) fueling tables that are tuned for your target Ethanol content. As the Ethanol content changes, the scaling between the two tables changes based on the translation tables. As an example, from E0 to E20, you don't get much of a change, but between E20 and E70, you have a large change needed in the fueling and ignition timing, but you may not want to add as much boost at this point.

Some systems will also have calibration information for the different fuel properties, like the stoichiometric values for the different fuels.

So it seems these systems are mainly modifying the timing and fuel. How do Wideband sensors come into play? are they still using a stoichiometric preset from the ECU? If it's not based on a stoichiometric preset then I should be able to set my target lambda to a desired value, and then change timing based off which fuel I run (like e90)?

It depends on the ECU and how the closed loop fueling is configured. If there is enough range available in the Closed Loop control strategies to adjust the fueling to accommodate the extra 35%~40% fuel needed when using E85 in comparison to E0, then you possibly could get away with using the closed loop fueling to maintain the Lambda Aim. Most Closed Loop systems will only allow around 10%~15% trimming before they go into a fault mode and ignore the lambda sensor value as being faulty.

You will also be leaving power on the table as the ignition timing needs to be advanced to accommodate the different behaviour of the Ethanol fuel.

Without flex fuel if the vehicle is tuned for E10, lambda 1.0, and you fill the tank with E85, lambda target will stay at 1.0, but lambda actual will become very lean, perhaps 1.2 despite large positive fuel trims, because the amount of fuel delivered is no longer sufficient to achieve lambda 1.0 since E85 requires far more fuel to air to achieve a given lambda than gas does.

Ford uses inferred flex fuel on some vehicles, by assuming the vehicle is 100% stock and mechanically sound. With that assumption, when fuel level changes and fuel trims change significantly, they can infer it's because of a change in fuel stoichiometry. I don't know all the back end details of their system, but perhaps based on the total fuel to achieve a given lambda value with a given calculated air mass, the inferred ethanol level updates. I'm sure there is a lot more to it.

Then some Ford vehicle have sensor based flex fuel which is more common. On my 19 F150, sensor based and inferred options are both available in software despite it not being a flex fuel truck.

At the OE level, flex fuel is generally focused on accounting for the change in stoichiometry, and startup needs. Often there is no change in ignition advance at mid to high engine loads, and changes to power enrichment are often smaller than you'd expect.

There is generally a lot of room left to generate a performance improvement from calibration changes while ethanol content is significant.

Hi Mike,

When I spoke to the Ford engine guys from Brazil, they explained how their inferred flex system worked. They had modeled the knock characteristics of the engine with different levels of ethanol content (they have E98 from the pump there) and would advance the timing under light load in one cylinder (once the engine was in the correct operating window) until knock was induced, and based off of the amount of ignition advance, they set the ethanol content of the fuel. This only occurred when the fuel gauge detected an increase in the fuel level of the the tank.

We tested doing this here, but didn't continue with the development as it required too much work from the end user to get working correctly.

Hey mike,

When you said the "lambda target will stay at 1.0, but lambda actual will become very lean, perhaps 1.2 despite large positive fuel trims" does this mean there are limits to how much the fuel trims can adjust in both open and closed loop? And are the wideband sensor reading based on a stoichiometric ratio of 14.7? In this case, due to not having stoichiometric settings could I offset my lambda target based on the percent error my sensor would display?

BlackRex, will open loop fueling have the same limitations that the closed loop side has in terms of fuel trim limits? Due to no ability to change closed-loop fuel once I supercharge my truck, I planned to force the system to stay in open loop. I am also not too worried about leaving power on the table because I plan to run 93 on the street and ethanol for track only, this way I would only need to worry about keeping 2 tunes (and changing the timing in one based on Ethanol content).

Hi Justin,

Open Loop will have even more limitations as this just runs on the preset tables, so if they do not have any compensations being applied, then they will just run as per the table.

BlackRex, if Open Loop has just as many restrictions what are people doing for fuel control on an aftermarket boosted application. because it seems like closed loop will always try to target a 1.0 lambda which would not be good with boost. Or do most tuners only tune the WOT fuel on slightly older applications? iirc I believe there is fuel correction based off boost on newer ECU's. However what do they do with systems that don't have this?

Closed loop fueling can have targets other than Lambda 1.0 Perhaps the ECU/PCMs you are familar with don't but many do.

Justin,

Yes you can view everything on a gas E0 14.7 stoich scale, but I really suggest working in and monitoring wideband in lambda. It avoids a lot of pitfalls in the tuning process.

Stephen,

That's neat thank you. I suspect my Ford may be a bit different since I've seen it update inferred ethanol sitting at idle at the gas pump before I've even driven off, and I wouldn't think they'd be executing that sort of knock based cycle at idle. Perhaps it's a scenario of different strategies in different markets, or perhaps a change their methods over time, but I always enjoy hearing about, deciphering, and designing control strategies.