fuel consumption/mpg calculation

computer, #1 injector on time + all parameters IE. speed, ambient temp, engine temp, gear selector position, rpm, throttle position, intake air temp, O2 sensor output, air pressue, etc. gives a fairly accurate fuel/mileage reading.

 

There are several threads on this. It is strictly the #1 injector on time. Now this can be very precise, but unless you can guarantee the amount of fuel squirted during this on time, not accurate. I believe the injectors have a margin of error of 5%. The closed loop sensing of the O2 in the exhaust makes sure the right amount of fuel is entered, but the computer can't know if the error is due to operating conditions, or incorrect determination of fuel dispensed.

 

The only way it could calibrate the flow rate from the injector is if it could determine the fuel used at a macro level, and compare it to what it 'thinks' it used at the micro level. It's kind of like measuring gallons with an eye dropper.

What measurement of fuel useage do we have at the macro level? That would be the fuel guage, and we know how accurate THAT is.

The other inputs: temp, pressure, power demand, etc,etc, etc, is so the computer can guess at how much fuel to deliver. It then checks the O2 content of the exhaust, and adjusts the fuel delta to get the right O2 content in the exhaust that indicates optimum burn. This error it finds could be caused by an error in any of its assumptions or inputs, including the error of how much fuel it delivered. It can't determine where the error is, only the end result(s): poor burn (O2 content) or incorrect power (too fast or slow).

That's the beauty of closed loop feedback, they are self correcting, not self guessing. That actually is why we can stand straight on two or even one leg.

 

The real question is... is the system's error larger or smaller than the error made by calculating mpgs manually?

 

there are probably several areas's where error could be introduced, temp degrees per microvolt output from the O2 sensor though small could change all the values in the closed loop also the air intake temp sensor, also the engine temp sensor, margin of error in the discharge of the injector, etc. That's probably why there is a difference in the mileage from car to car. Now what is the % 1,2 3, 5, 9% hard to determine. Also the variables could go either way,IE: cold O2 sensor but a lean injector = normal output. it'd be like picking fly s**t out of pepper. I wouldn't loose sleep over it in any case. The closed loop system handles it pretty well, check your mileage

 

I thought that's what this topic was about; determining accurate MPG.

 

I think the manual method is much more error prone since you can't be sure you fill the tank to the exact same level each time. The bladder might not stretch as much in cold weather. Different pumps will have different errors in their flow rates. etc, etc, etc

 

the amount of gas used to go x amount of miles is the mpg. LTT is the only real mpg

 

You're right!!!

The accumulating (pulse on time - injector dead time) is very proportional to the fuel usage.
I just made my own fuel mileage display instrument and tested last holiday.
I drove 502.10km and my measured usage number was 19.675 liter.
The MFD display was 502km and 26.0km/L, which was 19.31 liter used.
I know the MFD always shows 2% less consumption, therefore adjusted usage was 19.70 liter, which was only 0.12% more than measured usage.

Regards,
Ken@Japan

 

LTT - Long Term Testing???

I agree that the miles/actual amount used is the real MPG. But for a specific tank, I don't trust the "did I fill it to the same level as last time" calculation. Now that it is colder, my car seems happy to kick the pump off about a gallon short of my expectation based on the MFD MPG number. I then have to squeak the next gallon in. I have had absolutely NO hint of spit back so I don't think I'm overfilling.

However your comment made me think. Now that I have nearly 4K miles on it, I should be able to get a fairly accurate lifetime MPG calc based on miles and fill as there will be less 'single tank' variation.

OK, did that. At 3765 miles on the odometer, the MFD lifetime calc for the last 4 tanks seems to have settled and averages 0.32 MPG higher than the "gallons put in" calc.
Those 3765 miles cost me $181 less than if I were still driving my Odyssey. Savings: about 5 cents per mile.

 

My car spit back at me last fill up. Pump seemed to kick off early, so I put much more than normal. Based on previous fills, that shouldn't have been a problem. The difference was that I didn't pull the nozzle out a few times after the first kickoff. That may be the key, burping the baby when nearing the end of its feeding, or she'll spit up.

 

LTT=life time total
tank to tank has to many variables in it to be accurate. As you will notice the cold and bladder shrinkage and different pumps all affect the tank to tank.