Car MPG – Computer vs Actual, Why the Difference?

Why does the computer in my car show an indicated average MPG of around 55 but when logging mileage and calculating MPG using brim to brim fuel consumption the best I can get is about 50?

This, with different numbers, has been the case with every car that I’ve had with a computer MPG display in it.

Surely it must be possible for an on-board computer to be made that can give an accurate figure.

I know that vehicle manufacturers’ MPG figures are calculated using a standard set of conditions, don’t reflect real world conditions and are pretty much impossible to achieve in real life.

This has happened in every car I have had with this function.

At least the current one’s speedo only overreads by 3mph at 70

Speedo isn’t accurate, therefore your calculations won’t be accurate.

Car Speedos over read by 5 or 10 percent. If the odometer does likewise then that accounts for most of the difference.

Genuine question; why do people measure/monitor their fuel consumption?

My Wife’s old Nissan Qashqai was very accurate. None of our other cars have been. Worst offender was BMW.

Genuine question; why do people measure/monitor their fuel consumption?

Cause I’m a data geek (or control freak!).

My BMW does it too, it has built in sat nav, so no excuse for not having an accurate self calibrating speedo.

The speedo is usually <5% at 60 mph, but the fuel consumption figures are always between 5 & 10% out.

It’s just not in vehicle manufacturers interests to provide an accurate easily comparable figure.

IHN – Member
Genuine question; why do people measure/monitor their fuel consumption?

Because it encourages you to drive in a sensible way.

Because it encourages you to drive in a sensible way.

But I would bet that the kind of person who monitors their fuel consumption is the kind of person who drives in a sensible way anyway.

There’s so many variables, that there is bound to be an error. It’s convenient that it almost always seems to be in favour of reading higher mpg than you are actually getting….it’s almost like they do it on purpose. 😉

My old Ibiza 1.9 PD TDi was normally only around 2-3% out between calculated & actual. I had a ton of data points to know that this was a consistent amount of variation.

My current Ibiza (actually my Wife’s car) has the 2 litre TDi engine in it (one of the emissions scandal engines) and over-estimates hugely. It consistently reckons I get between 60-62mpg, with the occasional foray into high 60’s during consistently warm weather.
The reality is that getting 55mpg out of it is very good going; so it seems to over-estimate by at least 10%.

You would have thought that they would use a similar calculation method, so to me it’s surprising to see such a difference.

I monitor it on the car alone despite knowing it is a little out. Primarily as it allows me to understand if something might be wrong with the car if there is a significant drop in MPG.

It also allows me to plan roughly how much fuel I need for longer journeys so I can fill up locally instead of paying MWay service prices.

IHN – Member

Genuine question; why do people measure/monitor their fuel consumption?

Because fuel economy is the only reason to justify owning a diesel 😈

Mine seems to under-read – 29.6mpg indicated. Actual = 420 miles / 63 litres = 30.3mpg in mixed (average = 5.5m b-road, 6m motorway, 5.7m in start-stop) driving.

Not bad for a 13 year old, 3L, automatic, petrol estate car.

Old Clio 182 was remarkably (surprisingly) accurate. However, the average MPG dropped when you left the ignition on without the engine running 🙂

Speedo isn’t accurate, therefore your calculations won’t be accurate.

I assumed this too.

But the ‘average speed’ on the computer is almost bang on. If I sit on the motorway at ~65mph (sat-nav), re-set the computer and wait a minute, it’ll say 65mph.

It’s as if the car knows exactly what it’s speed is, or at least has an ‘accurate’ number before applying a +ve error to the speedo. Then uses the speedo number to calculate mpg.

That said, an old set of tyres is ~12mm smaller diameter, which makes almost as much difference again.

Car Speedos over read by 5 or 10 percent. If the odometer does likewise then that accounts for most of the difference.

See, I thought this at first. I actually started typing a reply then deleted it because I realised – if the odometer is throwing out the computer’s calculations, it’ll be throwing out the OP’s manual calculations also.

It also allows me to plan roughly how much fuel I need for longer journeys so I can fill up locally instead of paying MWay service prices.

If it’s a longer journey, would you not just fill the tank locally before you set off? Or are you trying to avoid having too much fuel and only taking the amount you need for the journey? And if so, why?

subjectively on the matter of speedo over or under reading – I guess I’d need to get the GPS or higher accuracy unit out to test against.

My old Saab 900, if I drove towards one of those “I display your speed” safety signs, indicated “30” on the sign was often pretty close to 30 on the speedo. Current Octavia – “30” on the sign, and the speedo is consistently reading 28 or so. so yes, between 5 – 10% out somewhere. Correct (as far as aI know) tyres and tyre pressures.

How do you know that the petrol pump isn’t lying to you?

Have you ever actually seen the petrol going in?

It counts litres and charges you for litres but I’d wager that it’s actually dispensing the fuel in increments of nine hundred and something millilitres.

How would anyone ever know?

Cougar – Moderator

Car Speedos over read by 5 or 10 percent. If the odometer does likewise then that accounts for most of the difference.

See, I thought this at first. I actually started typing a reply then deleted it because I realised – if the odometer is throwing out the computer’s calculations, it’ll be throwing out the OP’s manual calculations also.

I went through this thought process before I started the thread.

My thinking is that the odometer will give a consistent figure to both computer and driver calcs so won’t affect the result either way.

The maths to work out MPG is very straightforward and the same for both calculations so on this basis the only figure that can be variable (in this case) is the actual amount of fuel used over a given period/mileage.

As forecourt fuel pumps are calibrated and checked by trading standards/weights and measures people it must be the fuel useage sensor thing in the vehicle that is potentially inaccurate.

Car Speedos over read by 5 or 10 percent. If the odometer does likewise then that accounts for most of the difference.

Nope – this would only be the case if you were independently validating the mileage. Assuming that you are in fact using the odometer readings to calculate fuel economy, the error will affect manual calculations the same as the on board calcs.

Could be deliberate, I guess – another VAG-led initiative to make you think your car is better than it actually is….

Edit – dammit, beaten by two other posts while I made a coffee halfway through writing this!

I monitor mine as it can indicate a engine problem before you would otherwise notice.

Car Speedos over read by 5 or 10 percent.

See, I’d have said that’s not so true these days. Yes, they’re not allowed to under read but I don’t think they over read like the old cable speedos from years back. I doubt they’re more than 3-4% over at 70mph for the most part and probably bang on at 30.

“How do you know that the petrol pump isn’t lying to you?”

“…forecourt fuel pumps are calibrated and checked by trading standards/weights and measures people…”

They’re checked by TS, re-calibration will be done by the manufacturer/supplier of the pumps to maintain their traceability.

PP – The same applies to weighing your bananas in the supermarket, etc, these things do get checked.

on vauxhalls, there is actually a configuration accessible by using ODBC to change the ‘fuel consumption correction’ that is displayed. By default it is set to +10% iirc. You can (if you care) adjust it to be accurate

http://workshop-manuals.com/vauxhall/corsa-d/n__electrical_equipment_and_instruments/technical_service_bulletins/poor_fuel_economy/diesel_ems_system_wrong_fuel_consumption_indication_shown_in_board_computer/

I would imagine other manufacturers have a similar fudge

See, I’d have said that’s not so true these days. Yes, they’re not allowed to under read but I don’t think they over read like the old cable speedos from years back. I doubt they’re more than 3-4% over at 70mph for the most part and probably bang on at 30.

It varies certainly, and I haven’t tested every car I’ve ever driven, but I’ve yet to see anything that low.

My 2p re. speedo accuracy:

I regularly drive a Ford B-Max and recently borrowed a BMW 330e for a week.

Based on the local sign posts that display your speed, the BMW was bang-on but the Ford overestimates the speed, showing about 33mph when the sign says 30.

I’m quietly confident that my car’s Odo matches the SatNav distance. However the trip computer seems to work in SI and convert to MPG with large increments from a rounding step somewhere. Were I confident that I could get it back, I’d try running it in km and litres for a while to see.

Re: Speedo accuracy.

Because I used to have too much time on my hands I actually did some research into it.

There’s 2 main factors:

Tyres – tread depth and pressure affect the circumference of the wheel and as this is the only data source for the speedo they’re inherently inaccurate.

Cost – when car makers are shopping around for bits and bobs for their new parts they’ll be offered a huge range of versions and they may be like this “this one is £5 a unit and it accurate +/- 10%” “this one is £20 a unit and accurate +/- 3%” the car maker will say “I’ll take 20 thousand of the £5 one.

So they end up with a speedo that’s accurate +/- 10% gathering data that can vary another 2/3% depending on the tyres – so calibrate the speedos to over-read by 13% so at best they’re bang on and at worst they’re over reading. There’s no down sides for them because:

People generally cannot perceive the difference.

If they under read and their customers starting getting tickets all over the place they would have hell to pay.

It makes the cars seem faster to customers who care.

It makes them seem more efficient to customers who care.

It adds the miles onto the odo quicker so services come around quicker.

It ‘ages’ the cars quicker – so customers will change them sooner.

If questioned it’s sold as ‘safe’ because speed kills.

*figures made up for the purpose of example.

My old focus 2ltr diesel used to regularly show 50-51mpg whilst actually doing 48ish. Speedo was around 10% out against the garmin.

Current Megané 1.6 diesel regularly shows 57-59 but doing 52-53.
Speedo unerringly accurate as runs off the satnav.
Error must be in the way it measures fuel use.

FWIW, I’m always checking mine, It’s like a competition with myself to get better MPG.
My Passat used to be pretty good with the readout vs actual being very close, usually within 1-2 mpg either way. Mrs Egf’s Merc 320CLK was pretty accurate too, with a trip to Scotland showing 46mpg but was actually 45.5, currently sh’es got an Xtrail which has shown 1mpg either side of actual.
My Mondeo however is a complete fibber & is always about 4-5mpg on the wrong side of me being happy. 🙄

Mrs_OAB’s Seat Ibiza estate gets about what the official figures are 😯 The readout says more. I’m ok with a lie when it’s a good mistake…

The idea is to get the number as low as possible, right?

That depends on whether you’re working in metric or imperial

I’ve never done the computer vs manual test before but I do know when the computer says 100 miles to empty it’s on zero a damn sight sooner than 100 miles.. Does this mean the computer is over egging the 25.6mpg and I’m actually getting less?! 😕

@nickewen, pretty much yes.

In my current car I think the range is calculated based on average consumption over the previous 2k miles (or something).

I’ve also (to my shame 😳 ) been tracking my MPG using Fuelly and have averaged 49.7 Mpg over the last 40k ish miles and the on-board computer is always between 53 and 55 mpg

Oh dear.. I’ll just leave it on the time/date screen for another 17,000 miles and continue ignoring th fact that I’m at the petrol station far too often

I monitor mine as it can indicate a engine problem before you would otherwise notice.

Ditto, but I just look at miles per tank full.

It also tells you which garages sell sh*t petrol. (Sainsbury’s, I’m looking at you…).

Possible reasons for differences:

Car computers display “instantaneous” MPG, and I would assume that the reported long term figure is a cumulative average of this, rather than a running [miles traveled] / [gallons used] type of calculation.

They (generally) don’t report MPG when stationary, so fuel used when waiting at traffic lights, etc. wouldn’t figure in the overall average; also the extra fuel used when starting the engine wouldn’t figure in the overall average.

Seems there are some big discrepancies, nonetheless.

The computer calculation would require an estimate for the bleed back lines and depending on the fuel system a performance curve for fuel flow pressure etc of pump. I expect this is where the errors come from. The rest of the fuel air system has a closed control loop so can learn and improve on factory settings. The mpg calc doesn’t have this possibility.

My SMax says i’m getting near 50 mpg when it is actually about 40.

Linear fuel level sensor in a slightly conical tank?