Enginuity - LTFT with Methanol

[dynamix]

Just a query with this.

I have been running methanol directly within the fuel tank at a concentration of 10%. Works superbly and the extra power/torque is stunning but....

Logging shows that the long term fuel trims are learning to add extra fuel to compensate for the leaning effects of the methanol in the fuel. This is around 5.5% at the point at which it switches over to Open Loop.

Two thoughts cross my mind:

1) reduce the LTFT limits down from the +/-15% that are set in the ECU ROM
- This will still mean that it will trim the fuel using the fuel correction in closed loop, but when it comes time for power - it wont have values stored in the RAM for it to apply in open loop.

2) alter the Front O2 sensor scaling to kid it into thinking the fuel is the right AFR for its conversion. Methanol leans mine by about 0.5:1 AFR so changing the scaling for this sensor 'should' do the job.

What do you guys think - am I barking up the wrong tree ?

ps - when can we have the enginuity forums back as I am sure this has been answered there already ?

* Registered users of the site do not see these ads.

[gabedude]

Just decrease the injector scaler by 5%. That should do it.

Gabe

[dynamix]

tried that Gabe - it just ended up at the same place in the end. ie. the wideband AFR was the same as with the LTFT. Sure it didnt have the LTFT but the net AFR was the same which was too rich.

[gabedude]

When does your WB logs match your fuel targets?

After learning or before learning?

If it is before learning, then the scaler is set correctly. If it does no match after a reset, adjust the scaler. Then adjust the MAF scaling in the learning ranges (below 80 G/S). You can adjust the voltage or the G/S values. However, first thing you want to do is make sure your fuel targets match your WBO2 right after a reset (when AFL D range is 0).

[gabedude]

Here is more info. This applies to ethanol, not methanol, but it is similar.

http://www.innovatemotorsports.com/resources/news6.php

An engine tuned to 12.5 gas AFR will run at the equivalent of 13 gas AFR with a 10% ethanol blend. This is what these people were seeing.

Of course, when running in closed loop, the engine will run at 14.13 AFR instead of 14.7. O2 sensors (incl. widebands) don’t measure AFR, but Lambda. Lambda is defined as actual AFR/stoich AFR. It's a ratio. In closed loop part throttle the engine is just running at Lambda 1.0, regardless of fuel. The same would be true for other Lambda values when running closed loop at WOT using a wideband. The engine would run at the tuned Lambda and everything would be fine. Open loop systems would need to be retuned for alcohol blends though.

The front O2 measures in Lambda. Your AFRs should be richer than mormal. You should be able to custom set your WBO2 to read correctly with that gas. I know that with the innovate WBO2, you have to set the fuel you are using. There is a custom setting for mixed fuels to set the stoich AFR. After that, tune it.

Gabe

[gabedude]

Read the E85 fuel FAQ too. It will help you.

Gabe

http://forums.nasioc.com/forums/showthread.php?t=803341

[Freon]

It might be easier in the long run to just think in terms of lambda and just drop the idea of air to fuel mass ratios.

Really, changing injector scale should fix up changes in fuel stoichiometry.

[dynamix]

Originally Posted by gabedude

When does your WB logs match your fuel targets?

After learning or before learning?

If it is before learning, then the scaler is set correctly. If it does no match after a reset, adjust the scaler. Then adjust the MAF scaling in the learning ranges (below 80 G/S). You can adjust the voltage or the G/S values. However, first thing you want to do is make sure your fuel targets match your WBO2 right after a reset (when AFL D range is 0).

WB matches prior to learning or is marginally leaner than target. When learning is applied it is richer.

I am thinking lambda but it seems that the front O2 sensor thinks in terms of AFR due to the O2 Calibration table (hence why thinking of that as a fix)

[dynamix]

injector scaling it is then

I will trim the fuel targets using the fuel tables separately.

[crazymikie]

Originally Posted by dynamix

WB matches prior to learning or is marginally leaner than target. When learning is applied it is richer.

I am thinking lambda but it seems that the front O2 sensor thinks in terms of AFR due to the O2 Calibration table (hence why thinking of that as a fix)

The front O2 should be working in lambda. The whole purpose of the front O2 sensor and the learned trims is to try and account for variances in fuel. You should be able to calibrate the injector scalar so that learning goes back to 0.

I would not change the front O2 cal in this case. It shouldn't be necessary to fix your problem.

[dynamix]

I am convinced that the front O2 does work in lambda but with the calibration table it translates to an AFR for petrol of 14.7 and then uses this translated figure through the rest of the ECU. I dont know enough about the internal workings of the ECU rom to tell myself that this is rubbish it is certainly what it looks like to my untrained eye

[gabedude]

It is just learning to add 5% fuel as it should which gives an equivalent Lamba value as it would with straight gas. Just adjust the lower portion of your maf scale (below 80 G/S) to make it as close to 0 as possible. Remember that your stoich AFR is now offset (ehtanol and methanol have a lower stoich AFR), so tune it a bit richer in the fueling tables so it does not run lean (measured in lamba).

By that I mean 11:1 on gas is not the same as 11:1 on gas+10% methanol. The AFRs from your WB will differ, but lamba would be the same. You just need to figure out the equivalent AFR (it will be richer) to put into your fueling tables. Also, get your MAF scaling adjusted. There is a ton of info in the e85 fueling FAQ.