Flex Fuel AFRs

I am tuning a Subaru STI on the Cobb Flex Fuel ECU.

I have recently started logging with high ethanol content and have noticed that my AFRs are leaner by 5-10% than my e10 tune during ramp runs to redline.

I am running ID1300x injectors and the ID provided injector characterization data. Scaling linearly between e0 and e100. During idle and cruise conditions in Closed Loop, my fuel trims are excellent. Cranking and warm up are also working well.

I am logging fuel pressure and have differential fuel pressure compensations active so fuel system hardware issues seem to be ruled out. My injector duty cycle is 75% at redline.

My intake temperature can vary my MAF calibration by a bit, but my intake temps have been fairly constant, not enough to cause this kind of variation.

What else am I missing? I'm thinking of adjusting my High Ethanol OL fuel target tables to offset the variance since my MAF scaling was working fine on my pump gas tune...

You can make scatterplots of injector pulsewidth vs AFR with E10, E85, and if possible E40 to confirm that the relationship between the injection pulsewidth and the AFR has shifted according to the ethanol concentration.

Most likely it is real and the most important root cause is most likely differences in wall wetting. You have more fuel sticking to the intake port with higher ethanol content. At a given ethanol content, there is a relationship between wall wetting and coolant temperature (really intake port temperature but we only have an ECT sensor). If your flex fuel software can't directly account for that, you need to do some kind of offsetting/fudge factor as you've described.

On the flip side, E85 runs cooler than E10, so you may not need as much enrichment anyway.

Hey thanks for the reply!

That is an interesting point about the wall wetting, but I wasn't really seeing the lean issue during transitive conditions, where manifold wall wetting conditions were active. Rather it was when the target AFR and load were relatively constant. In addition, the MWW tables aren't able to blended between ethanol contents on the Cobb ECU.

At first I tried adjusting my scalar up to correct the issue, but as expected, this threw my fuel trims during closed loop. Since my fuel trims were so close during closed loop, I chalked it up to an error in my MAF scaling in the open loop voltage ranges. I made the required MAF calibration changes and now things are looking as they should, and have continued to look good as my ethanol content has increased.

One consideration is that you'll find that due to the differing viscosity of the two fuels (pump vs E85 for example), the injectors physically flow less fuel volume for the same pulse width on E85 compared to pump. This seems to vary dramatically from injector to injector so I can't offer an across the board value you can apply. The ECU manufacturer Link did some testing locally on the Bosch EV14 based 1000 cc injectors and found that on E85 the injector flowed close to 10% less fuel.

That could definitely be it. Unfortunately there doesn’t seem to be a compensation I can apply in the stock ecu. At least with my maf scaling set up for e85, it will run rich on pump gas, if this is what in fact what I’m dealing with.

Thanks for for the reply!

Some of the aftermarket ECUs now offer a compensation for injector flow vs E% but this is still rare.