MSefi.com read only forum (Jan.15,2018)

Hey guys,

This is more of a theory question then a MegaSquirt question (unless MegaSquirt can fix this)
My brother and I have both noticed in the cold winter months our gas mileage drops off about 10%, Cars are '98 Altima, '99 Taurus, '88 325SI.All stock ECUs. We are in the Northeast, NY to be specific and we do have an additive in the fuel, up to 10% of ??something?? I will check the pumps next time I go fill up. I am pretty sure it is seasonal, from November to April +/-. I do not think the additive is MTBE. I believe that was banned a while back, not sure. I think it is some sort of alcohol. My theory is that this "blend" has a different Stoich point, and since the cars ECU were programmed for gas (none of them are flex fuel) during the winter months they are running a little rich. Would leaning the fuel table a little fix this? Would the Flex Fuel sensor compensate for this?

Rusty

It's probably ethanol, but a small enough amount that the EGO feedback compensates for it and richens up the mixture a little.

Matt Cramer wrote:It's probably ethanol, but a small enough amount that the EGO feedback compensates for it and richens up the mixture a little.

Soooo.... if this were a MegaSquirt install, we could adjust the AFR target, or use a flexfuel sensor and get this runing right.
Rusty

I believe most gas stations now use a 10% ethanol/ gas mixture year around. But if there is more alcohol introduced in winter then you will consume more fuel because alcohol has a lower energy density than gas. Also wide bands calibrated for pure gasoline may not give accurate readings when alcohol is added - but NB will.

grippo wrote:I believe most gas stations now use a 10% ethanol/ gas mixture year around. But if there is more alcohol introduced in winter then you will consume more fuel because alcohol has a lower energy density than gas. Also wide bands calibrated for pure gasoline may not give accurate readings when alcohol is added - but NB will.

Hmm.... all the cars are NBo2 Yes 10% Ethanol, that's it, but I think the tags read "Up to 10%" so maybe less in the summer. Why would the NB read correctly where the WB does not? And can a flex fuel sensor adjust for the % Ethanol?

Thanks!

Rusty

rusty105 wrote: Hmm.... all the cars are NBo2 Yes 10% Ethanol, that's it, but I think the tags read "Up to 10%" so maybe less in the summer. Why would the NB read correctly where the WB does not? And can a flex fuel sensor adjust for the % Ethanol?

Rusty

The NB tells you there is or is not free O2 in the exhaust, period. You can use it for control by trying rich or lean pw and on average you will be correct. But its kind of like driving blind in a tunnel, every time you hit the wall you get bounced back, so on average you are in the center, but its not a smooth ride. With WB, when the fuel mix changes it alters the correctness of the WB output, so if it says 13.0 the ECU will put you at 13.0, but 13.0 may not be the actual AFR - it could be considerably off compared to a wb that was calibrated for the actual fuel mix.

A flex fuel sensor is designed for exactly that purpose, so yes it senses the % ethanol in the tank, but of course you have to have an ECU that can read it and then you have to give the ECU a table of what the corrections should be vs %ethanol.

grippo wrote:With WB, when the fuel mix changes it alters the correctness of the WB output, so if it says 13.0 the ECU will put you at 13.0, but 13.0 may not be the actual AFR - it could be considerably off compared to a wb that was calibrated for the actual fuel mix.

Unless you set your system up to work from Lambda values instead of AFR ratios. Then it should be "right on" all the time.
My 2c.

That only depends on whether you can tell the WB what mix of gas ethanol it is using. The WB will always be right at stoich, because there is no excess oxygen, but as you go off stoich either way (and if you don't why use a WB) both AFR and lambda will start showing errors. So if the WB control software asks for what fuel you are using and recalibrates accordingly, then it should take that into account for both lambda and afr, but if you are going to have varying mixes then you will be constantly recalibrating your WB.

grippo wrote:So if the WB control software asks for what fuel you are using and recalibrates accordingly, then it should take that into account for both lambda and afr, but if you are going to have varying mixes then you will be constantly recalibrating your WB.

But, the WBO2 is calibrated in a free air situation and outputs Lambda values, regardless of fuel type. So, if you have your AFR table set up to "tune in" to these Lambda values instead of AFR's, and you do not program the WBO2 for any particular fuel (AFR), how would there be any deviation?
I'm not doubting you, just trying to understand.

The WB will give you a lambda, and it will look reasonable, but it won't be correct as you go off stoich, at least not on the rich side which needs to be calibrated for hydrogen and CO. If the fuel composition is changed, the hydrocarbon ratios change and the WB controller needs to know about this and take it into account. Now as long as it's consistent, you can tune for this, but your tune for a mix of gas and ethanol will change between 10% and 85% ethanol. I'm not completely sure about the lean side, but I am about the rich side.

That's cause for concern since I never programmed my WBO2 for any particular type of fuel, and I run E85 only. The only programming I did to my WBO2, (LC-1), was to output a nominal voltage for 1 Lambda to report 14.7:1 to MS2, so my tables would look like gasoline AFR tables. Originally, I tried tuning with a Lambda table in my AFR table, and running my logs through MLV in order to tune my VE table. After a few e-mails with Phil, however, I discovered that MLV uses 14.7:1 as a calculation for Lambda values, which makes a "direct" Lambda reading from the WBO2 ineffective for E85.
So, with this new information, maybe I should reconsider tuning with the WBO2 since it's not accurate with anything other than gasoline, even though E85 emits CO and Hydrogen as well due to the gasoline in it? How can I program my WBO2 to the type of fuel I am using?

In running some numbers through the equations, it looks, on a preliminary basis, that the change in lambda between gasoline and ethanol is not worth worrying about in terms of lambda (it is of course huge in terms of AFR). This means that at least for these fuels the lambda curves are linear so everything gets normalized out when afr is divided by afr_stoich. But I have seen non-linear curves where lambda will change. It also depends on the sensitivities of the wideband controller - I don't have hard numbers for these.

Thanks for the effort put into this thread.
So, does this mean that:

devastator wrote:Unless you set your system up to work from Lambda values instead of AFR ratios. Then it should be "right on" all the time.
My 2c.

Or, should it read "pretty darned close" instead of "right on"?