As a physicist and one time computer programmer, I’ve been thinking about the comparison of manual vs computer mpg calculation…..

The calculation is simple Miles travelled divided by fuel used

Many people have expressed opinions that the computer calculated value is inaccurate, so, where are the discrepancies between the two?


Miles travelled
Noted from the odometer when calculating manually.
I presume everyone zero’s there trip computer when brimming the tank, hence there should not be any significant errors in mileage travelled.

I cannot imagine the computer calculation takes mileage data from anywhere other than the same memory that is used by the odometer.
Presumably a snapshot of the mileage it stored when you reset the computer calculated MPG. This is then subtracted from the current mileage and divided by fuel used (see below) every time the computer re-calculates the MPG displayed on the screen.

There may be very small rounding errors if the computer readings use 1 or more decimal places, compared to a manual calculation, where I would assume mileage is rounded to the nearest whole number, but overall the impact would be very small:

e.g. if we assume worst case rounding errors at start/end readings, we still only get a max an error of 2 miles. Over a 300 mile test and assumed 25mpg = 24.83mpg (if calculated using 298 miles)

Other errors (e.g. running larger or smaller wheels) would be present in both calculations, hence can be discounted.


Fuel Used
Pump
I am assuming everyone measures fuel used by brimming the tank (i.e. fill until the fuel pump cuts off), running the vehicle for X hundred miles, re-filling to fuel pump cut off and noting the fuel from the pump reading.

There are at least two potential errors here – differing sensitivities of fuel pump cut off and accuracy of the pump reading.

I can’t find any data re. accuracy of fuel pump cut off. From personal experience, it does seem to vary a bit, but not massively.

A quick Google brings back lots of Trading Standards stories re. fuel pump accuracy testing. I have not read in detail, but it seems on the whole they find few cases of inaccurate pumps and most confirm to (within 1% above or 0.5% below) prescribed error limits.

e.g. if we assume fuel delivered is underdelivered by 0.5% max permitted error, over a 300 mile test and assumed 25mpg = 12 gallons. This would read as 12.06 on the pump and produce 24.88mpg


Computer
Assuming people correctly reset their computer calculated MPG at the point of brimming the tank, the calculation should be accurate over the life of that tank of fuel.
(As an aside, my 07 TDV8 has two MPG calculations – one I reset every time I fill the tank to monitor MPG over the life of that tank, the other never gets reset and gives me a long term MPG over many, many tanks and 1,000’s of miles – both usually read around 26mpg)

I cannot believe the computer uses anything other than injector data to calculate the fuel used. If anyone knows better, please correct me as this is an essential part of my reasoning….

Injectors have a very accurate flow rate. The ECU uses a range of sensors (MAF, Lambda, fuel rail pressure, temp etc.) to determine exactly how much fuel to supply to each cylinder every cycle. The amount of fuel supplied is determined by the duration of injector opening.
It follows that the ECU can calculate EXACTLY how much fuel has been used over the distance travelled.


Conclusion
If the assumptions I am making about how the computer calculates MPG are correct, then I cannot see how the computer calculated MPG can fail to be anything other than remarkably accurate.

However, if we also believe that most pumps are accurate to permitted thresholds and fuel pump cut off error is negligible, it raises interesting questions re. why there seems to be discrepancies between the manual and computer calculated values…..

Thoughts please… 2007 Zermatt Silver TDV8 Vogue SE - now sold but was a great car!

. . . . GraemeC - there's only one way to put this.

You need to get a shed, and go and do something in it -



PS I, too, used to think about fuel consumption when I had my Td6. I've a V8 petrol now and don't bother anymore. Where do you go after one of these . . . ?

I thought the OBD sampled a figure over time so gave an average over that period.
For example if you zero it and then blast up a hill it will read very low
If you zero it and coast down the same hill it will read very high.
However if you do not reset and repeat the exersice say after a week of normal motoring the up or down hill trip will not effect the OBD.
If you want an accurate figure for a given time or trip you would have to use the milage covered divided by the brim to brim tank liters and if you could use the same pump you would eliminate any disrepancies.

thats my 10p anyway Mine was a blue one!

Tug I'm with you I always thought the computer on my TD6 was unerringly accurate and now with petrol I don't give a monkeys  2008 MY Supercharged
Rimini Red / Jet
four zone climate
remote park heater
and no ugly kid windows.
magnus satis quod turpis satis

Andy, that's my point - if you zero the OBD when brimming the tank and take the reading when filling up with your next tank, that should be a very accurate average over the life of the tank

You will only see big variations in the OBD calculated values when you have done very few miles. Try blasting up the hill 300+ miles after resetting it and it wont make a difference... 2007 Zermatt Silver TDV8 Vogue SE - now sold but was a great car!

I have also wondered this,

I reset mine every time I fill up and the OBD MPG ranges from being absolutely spot on to sometimes 2 mpg too generous and other times 4 mpg over generous.

I brim the tank and do the correct calcs and also use the iphone APP Road Trip and Gas Cubby APP - my calcs are also 100% accurate...

long steady trips at 34 mpg were accurate, the short journeys where I was hoofing it were inaccurate ...

On the basis of company mileage and expense reports I have always found the LR product in general to be rather optimistic on the dash, whereas my old Merc was the other way. it actually did more mpg that the dash was leading you to believe.

The accuracy of fuel card stats against reported mileages and over a significant time period is more likely to be accurate purely based on the amount of input data being averaged I would think. Happy to be contradicted though if anyone knows better.

on a long journey my readout was very accurate when checked against a tank to tank fill.. ... - .- -.




Y. O. L. O.
.

Having run a 4.4 V8 (averaging about 16mpg) and now a 3.6TDV8 (averaging about 27mpg), I have come to one indisputable conclusion ref fuel economy. If you want economy, buy a Prius.

"Happiness is directly proportional to carbon footprint"

 2012 4.4 TDV8 Westminster
2007 3.6 TDV8 Vogue - gone!
2003 V8 HSE (LPG) - gone!
2018 F-Pace 25t Portfolio
1974 LR Series III Lightweight FFR
2014 BMW 640i Gran Coupe

I had my OBD computer inaccurate when i bought my FF, but i calibrated and now i check every time i fill up the tank, and i always get the same reading of the OBD than the manually calculation.

You can calibrate your OBD fuel consumption!

READ HERE MY POST here HOW TO:

http://www.fullfatrr.com/forum/topic19494.html Ex-FULLFATTY now downgraded

L494 RRS P400e Dynamic Narvy Black

-------------------------------------- gone
6 | FF-RR 2007 TDV8 Vogue JAVA BLACK
5 | FF-RR 2003 TD6 Vogue Bonatti grey
4 | Ex-RR 2000 P38 automatic
3 | Ex-RR 1994 300tdi classic
2 | Ex-RR 1993 200tdi
1 | Ex-RR 1991 turboD VM engine

I would dearly love to believe that the computer was woefully low versus reality, although to be honest would probably rather not check in case it was actually worse. Morbid Curiosity leads me to depress myself by looking at the figure regularly.

I genuinely wish I could actually remove the readout and just enjoy driving it is blissful ignorance. 4.2 Supercharged Vogue SE

I got to york and back at the weekend (176 miles)

used £35 worth of fuel - lpg

My car is so cheap I only put in a tenner each time and every time

what about the tyres?

Over a journey a new set of tyres would result in a lower mileage covered, compared to worn tyres due to less wheel revolutions required to cover the same distance.

Then of course there is fuel temperature when filling to think about.

Not forgetting air temperature and wind speed and direction to take into account.

Alternatively, you could enjoy the journey and forget about any + or - mpg figure.

Yes, but am error due to worn tyres would equally affect a manual mpg calculation and the on board computer calculation as both are using the same source for distance travelled - the odometer

My curiosity is not really about mpg - I knew exactly what to expect when buying a RR and I am totally comfortable with it - as has been said before, if bothered about mpg then buy a Prius!!

My interest is in the technical workings of the on board computer mpg calculation and why people think it is less accurate than a manual calculation. If it all works as I've summarised (and my assumptions do come from a decent degree of technical experience) then it should be extremely accurate.... 2007 Zermatt Silver TDV8 Vogue SE - now sold but was a great car!