Brendan Hughes

Wanted to ask this for a bit, but have been prompted, so here goes;

Just how accurate are these computers?

Basically, do they measure the fuel via the amount going through the injectors (which you'd expect to be hyper accurate) or by the tank sender reading (also known as the Wet Finger In Wind calculation). Or something else?

Anyone know?

(Coz my car seems to drink fuel on some trips, yet the computer suggests all is normal).

Ghostdog

I have also wondered this as when i put my foot down in my BMW it says I am doing -1000 mpg

brickboy

The ones in VAG cars do it by a calculation involving the MAF sensor, throttle position, ambient temps, injection timing etc etc to end up with a MPG figure ..... they must derive it from the fuel maps in the ECU.

THe ECU will know how much fuel (i.e. what mass of fuel will pass through the injectors) should be injected at any given time, and multiply this up.

On my Passat diesel the fuel calculations from the computer are only about 5% optimistic, which is more accurate than I thought it would be.

MattOz

My average fuel consumption on my computer tells me I'm doing 41.4mpg. I get between 500 and 550 miles from approx 55 litres of diesel. Seems to be pretty accurate. My 2nd mpg reading is at just over 44mpg, because I reset it the other day before a motorway run. It was showing 47 ish, but then I've done some urban driving since.

Overall the computer is pretty accurate and the figures are not bad for a 3 litre, 6 cylinder, turbocharged car!

Matt

Brendan Hughes

I've now started to take mileometer and fuel pump readings to do a few calculations. I reckon it's the long motorway runs that it doesn't like, at 150-160kph. Uses far too much fuel. Either there's a mistake, or the crap aerodynamics really make a difference.

brickboy

Brendan, what car is it?

On the Passat it makes only about 5% difference to economy cruising at a steady 90 cruise rather 80. At a steady 90 cruise the computer will read 52-53mpg over the journey. At a steady 80 cruise, it will show 55-56mpg.

Maintaining more than 90, consumption rises quite sharply to the low 40s mpg. Presumably aerodynamic drag, and the engine's out of its "sweet spot" at 90+.

Brendan Hughes

2.0 diesel RAV4 - hence my "drag" theory.

It took well over a tank (55litres - so maybe 70-75 litres) to do 550km down to the Algarve and back again. I thought diesels were a lot more economic than that.

brickboy

Brendan, perhaps the Toyota diesel has an engine "sweet spot" like the Passat for economy?

As mentioned above, the MPG computer shows the same economy (within 2%) whether I cruise at 85-ish, or at 70 on the nail. So I tend to cruise at the higher speed

But above 90, consumption rises very sharply -- the extra 5% speed increase carries a penalty of using over 20% more fuel.

Brendan Hughes

Cheers mate, will try that out as well

Brendan Hughes

Looked at the computer while driving yesterday - damn thing seems to be a random number generator.

But brimmed the tank, noted the mileage, and subsequently got 910km out of 45 litres, so that's not so bad. (Is it?) Obviously not much 150km/h driving there, then...

ALi-B

iTrader: (1)

I got the trip computer on a Pug 307SW HDi really confused the other day in Spain, whilst cruising down a very steep 6mile long hill, not touching the pedals and maintaining 70mph

Went to zero litres/100 km for a while, then couldn't make its mind up if it was using fuel or not, before finally giving up and reading "---/--- l/100 km"

Brendan Hughes

Ali - sounds like normal behaviour for mine

Puff The Magic Wagon!

iTrader: (2)

Recently tested my XC70 over 500 miles.

Trip computer read 40.3

Pump fill/mileage 42.5

The things you do on your daily commute Thing is I shouldn't worry about it 'cos its on a fuel card

R1916v

my old ren underreads by about 2 or 3 mpg. Comp says approx 27mpg between fillups, milage/fill count puts it bang on 30.

dnb

The trouble with all these fuel computer things is that they measure very small things (the amount of fuel used each cycle - only 1 or 2 ccs) and the distance covered each cycle (again, only a hand full of metres, and this is probably from integrating the measured speed w.r.t. time) to try to work out something very big. (ie miles per gallon)

The possibility for errors here is huge - imagine there's a 0.1cc error in the injector reading. Quite small in the grand scheme of things you may think, but over a journey of say 10 mins with a constant engine speed of 4k, this will become an error of about 2 litres.

BTW, I wrote a fuel meter for my MY96 Impreza, based on the select monitor port... It worked quite well, but under read by about 3 to 4 mpgs. Not too bad for a home lash up... I'd say that they can be made more accurate (I know a few tricks to try on mine) than this, and it should be possible to get to an accuracy of 1 to 2 mpg or better.