True Flex Fuel Tune

[K4fxd]

Just out of curiosity what parameters did you temporarily lose?

I didn't document them. I just counted. It was 16 or so that were not there. The new version has the same as the 2.11 version. I must have downloaded the 3.1 version right as they uploaded it.

I tried reading my car with 2.11 and it says it cannot read tri-core processor.

Sponsored

 

[engineermike]

I changed my min time and min ego switches to 20 and 20, doubling the default/stock values in order to prevent it from maturing too soon after a fillup. I tested it today and it at least worked this time (1 whole data point).

I went from E85 sitting on 0 miles to E and filled up with Shell 93. At 1:30 minutes from start-up cruising at 55 mph, the fuel trims dropped down to -21% while learning was "paused". I thought this was too soon for gasoline to make it to the engine and oddly enough, returned back to around 0% trim at 2:30 minutes. I can't explain this. It "unpaused" about that same time. At about 3:10 minutes, it started learning rapidly (I max'd out the max rate of learning). At 3:45, the ethanol % dropped below my 55% "toggle" and the borderline timing returned to gasoline values. At this point, it is "safe", the way I have it set up. The A/F gauge was around 11.5 at steady-state, and rising. I tried to keep the MAF between 2.4 and 3.0 lb/min in order to slow the fuel purge rate while also being in the range that would allow maturing (torture test, of sorts). At 11:30 minutes, the alcohol had learned down to 30% but not yet matured when I shut it off.

Just an update…it hasn’t matured yet but it made its way down to 16-18%, alcohol learned which is about right since there was probably 1-2 gallons e85 still in the tank when I filled with e10.

 

[K4fxd]

it hasn’t matured yet

This is the problem I ran into. I believe mine is now set to 15 on the time. I'm going to play with some of the added parameters in PCMtec and see if I can get it closer to how a F150 works.

 

[engineermike]

Ok, this is something I just learned that may be leading us down the wrong paths in testing. I think I discovered that the learning windows are actually different based on the reason for learning. This would be captured in the "Inferred AFR Source", which shows up as "fault exit" if you did a KAM reset, or "refueling" if you added fuel to the tank. I believe in "fault exit"/KAM reset, the learning windows are much wider, like it will learn at MAF flows down to <1 lb/min. It appears as though in "Refueling"/refilled tank, it uses the Airmass min (typically 2.4 lb/min) as a min threshold to allow learning. This actually makes a lot of sense, because a KAM reset wouldn't need the lines to be purged so it should learn as fast as possible. But after a refuel when you need to allow for the lines to purge, the min airflow (and fuel flow) rate must be higher which would better allow for a line purge to happen.

 

[K4fxd]

Compared a F150 to a GT mustang using PCMtec and all is the same except for fuel line capacity, timer and LAMBSA avg filter constant.

 

[engineermike]

Compared a F150 to a GT mustang using PCMtec and all is the same except for fuel line capacity

This is huge! It's zero'd out in the Mustang cal and .8 lb in the F150. I think this is the parameter that we need to prevent it from maturing too soon. We should be able to put the ego switches and time to mature back to stock values with that corrected. Wonder if the big name tuners are aware of this....

Which timer are you speaking of?

Also, how on earth did you find that fuel line capacity parameter? I'd been through the GDI stuff several times and never would have thought to search for that.

 

[K4fxd]

Which timer are you speaking of?

Time to mature. The mustang file was set to 5 seconds and the F150 10

Also, how on earth did you find that fuel line capacity parameter? I'd been through the GDI stuff several times and never would have thought to search for that.

Sometimes a blind squirrel finds an acorn.

 

[junits15]

Time to mature. The mustang file was set to 5 seconds and the F150 10


Sometimes a blind squirrel finds an acorn.

This is huge, do you have the ID's for those parameters?

 

[engineermike]

So with time and ego switches set to 20, it took four 10-15 mile trips for it to mature. But when it did it came in at 16.9% ethanol. My math says actual should be around 17.6% based on residual ethanol. But I’m also sure of my injector data and spent time to fine tune my maf curve.

 

[engineermike]

This is huge, do you have the ID's for those parameters?

I know you understand this, but to clarify for others…..junits15 posted a while back that ford described waiting an amount of time after a fillup before ethanol learning can start, but no such parameter seemed to exist and it seems to start learning pretty much immediately which can lead to prematurely locking into an incorrect alcohol %.

However, if a known amount of fuel exists in the fuel line and the pcm knows the consumption rate, it can determine exactly when the new fuel hits the injectors and, thus, know exactly when to start looking for the change and learning. There are 3 fuel line volumes defined in pcmtech (not in hpt), the fuel line from the tank to the T, from the T to the port injectors, and from the T to the direct injectors. Now, I’m not 100% side ford logic does this but it certainly has all the info it needs to calculate when the new fuel hits each type injector even if it’s at different times for port vs di.

In pcmtech, those 3 parameters are populated with reasonable values for the stock flex gen3 f150 calibration but they are all set to zero in the equivalent mustang calibration. I can’t help but wonder if this is why most big tuners have given up on flex tuning in so many ford applications. This, if corrected and with working logic, should take all the risk out of the ford flex strategy.

 

[Grimreaper]

is pcmtech able to read/define roush files? or are they limited like Whipple files?

 

[engineermike]

Don’t give up on Whipple just yet. They have something in the works.

 

[junits15]

PCMtec ID	Name	F150 stock value	Mustang stock value
auF61356	Additional LBM of dfi fuel that must be consumed to allow stabilization of AFR - any size refuel.	0.8​	0​
auF61359	Additional LBM of dfi fuel that must be consumed to allow stabilization of AFR - large refuel.	0.1​	0​
auF61358	Additional LBM of fuel that must be consumed to allow stabilization of 0x0p+0lcohol - any size refuel.	0.1​	0​
auF61361	Additional LBM of fuel that must be consumed to allow stabilization of 0x0p+0lcohol - large refuel.	0.1​	0​
auF61357	Additional LBM of pfi fuel that must be consumed to allow stabilization of AFR - any size refuel.	1​	0​
auF61360	Additional LBM of pfi fuel that must be consumed to allow stabilization of AFR - large refuel.	0.1​	0​
auF34825	Ethanol Calculation Airmass Timer	10​	5​
auF21178	Inferred AFR EGO Learn Events	10​	5​
auF21166	Fuel Type Detection Enable	1​	0​

Took me a second but I located them! These were buried under there! Thank you @K4fxd for finding these this will be very helpful!


In addition I saw this fun little thing:

In brazil, cars are made to be able to run on 100% ethanol, I wonder what this does.

 

[vaeevictiss]

I feel like this is a double edged sword. On one hand you can look at this like a lot of people look at water meth systems. Any little thing can go wrong like a pump failure and you can blow an engine. On the other hand, i feel like this can also happen with an after market e85 return style fuel system, or even the stock system. At least with a stock system id like to think there are built in failsafes that will cut power if something like that happens. but is it faster than that one detonation that gives your piston a nice view of the strut tower?

 

[junits15]

I feel like this is a double edged sword. On one hand you can look at this like a lot of people look at water meth systems. Any little thing can go wrong like a pump failure and you can blow an engine. On the other hand, i feel like this can also happen with an after market e85 return style fuel system, or even the stock system. At least with a stock system id like to think there are built in failsafes that will cut power if something like that happens. but is it faster than that one detonation that gives your piston a nice view of the strut tower?

This is the only real risk with an inferred type system, that your ethanol percentage could be wrong so your MBT/borderline timing could be much higher than intended.

I haven't been able to test this yet because I'm still waiting for my dongle to arrive, but if these changes work like we expect them to, it should eliminate a lot of the uncertainty and issues with incorrect learned concentrations. That in itself will reduce most of the risk you're (rightfully) concerned about.

The other thing we have working for us is the KR of 10 from the factory, and the way that timing is brought in slowly not all at once. If you add less than 10 degrees of timing over the stock gasoline timing, the stock KR system will be able to pull back to stock gasoline timing. Timing is not added in all at once, if the fuel is low octane, the knock will start immediately, and timing will be reduced by KR immediately at the beginning of the pull. So a borderline of 25 degrees at redline (as an example) will never happen if your fuel cannot support it. Over that, excessive knock will trigger a code and limp mode.

there is another failsafe that we could theoretically activate, which is the octane adjust system, with that activated you would have a secondary protection that would reduce timing if knock is excessive. So then, even if ethanol content is wrong, you would be safe.

There are a lot of failsafe's in this system, because it's factory.

Sponsored