There have been a good few posts about fuel consumption being much lower than the official figures suggest. As someone pointed out in a previous thread, the tests are done on a rolling road at 20- 30C. The urban test begins with a "cold" start, but there are few mornings in the UK when the temperature is 20C, so they hardly represent UK conditions in terms of warming up.
I have also read of a marked variation in mpg on cars of the same type in a fleet. They may have been run in differently and will clearly get driven differently, but I wonder how much may be due to the differences between cars in the sensors that the ECU monitors (air temp, mass flow, coolant temp, elec system voltage among others). These are obviously made to fairly strict tolerances, but if they are slightly out of calibration, how does the ECU know? If the ccolant sensor reads 5C low, then the fuelling may be affected because the ECU thinks that the engine is not fully warm - as an example.
Can any ECU experts out there enlighten us?
|
There will be differences, albeit reduced by the closed loop fuelling control which is required for petrol engines since 1993 or so. However, these differences will be far less marked than the difference between a poorly and properly set-up old fashioned engine with points, tappets, and carburettor.
The lambda sensors typically have a sharp change in their output when the concentration of oxygen in the exhaust changes - once the engine is warm, and is running at steady-ish state, the lambda sensor is the dominant sensor in deciding engine fuelling.
The main point about fuel consumption tests isn't to try to get the same figures in real life, because the tests **cannot** replicate or simulate real life, but to use the test results to benchmark one car against another. Just because car A delivers 40mpg in the tests doesn't mean it will do in real life - but, if car B delivers 45mpg in the tests, it will probably deliver better economy than car A in real life, even if the real life figures are only 35mpg!
Number_Cruncher
|
The reason for a standard test for fuel consumption is so that vehicle MPG can be compared between different vehicles.It is not a realistic figure that will be achieved when on the road and should not be used as such. The government in thier nievity use these laboratory figures to fix the allowance for company car riembursement.So at the moment for a diesel up to 2 litres 9pence a mile is the figure before it attracts tax.Urban driving on the C5 I drive means Iam subsidising the company I work for.
--
rustbucket (the original)
|
|
I seem to remember HJ mentioning that different tyres on the same model car can have a dramatic effect on fuel economy.
|
|
|
|
NC, thanks for the info. Is the lambda sensor also the dominant sensor in a diesel (common rail turbo)? Do you know what happens in the warming up period? Is the lambda sensor less important then? The fuel consumption of my Colt diesel is strongly affected by ambient temperature - presumably affecting warm up time. I guess a dodgy stat could contribute too, by opening too early.
I get the impression that some engines are inherently nore economical than others (or can be driven to greater economy). My 2002 Vauxhall Combo 1.7 DTi van returns similar mpg figures to the Colt despite being having a worse drag factor and being taller, wider and heavier!
Re tyres, I have seen manufacturers mpg figures for the same model of car on different tyres - lower profiles being worse.
|
|
My two tear old Almera was doing close to 45 MPG when I first bought it. The Lambda sensor was replaced after failing, and I have a job achieving anything like that now.
|
|
|
I don't know much about modern diesels - I'm blissfully ignorant of all things common rail!
On a petrol engine, the lambda sensor does nothing when cold - it sits there, registering a weak mixture, whatever is coming down the pipe. During this phase of operation, the engine relies on look up tables, based mainly on engine temperature, speed, and load. The ecu monitors the lambda sensor signal, and when the lambda sensor suddenly switches state, the ecu then begins to switch into closed loop mode, where the lambda sensor is used to keep the average mixture close to stoichiometric.
Owing to the non-linear response of the lambda sensor, and the delay between injecting fuel and the resulting exhaust gas appearing in the pipe, smooth, continuous control can't be achieved - instead, the system ping-pongs between rich and weak, with the average been about right. As the engine speed is increased, the time lag between injection and the gases being detected reduces, and the frequency of the rick/weak switching of the lambda sensor increases - typically they switch at about 1 cycle per second at tick over. When lambda sensors fail, they can sometimes become blocked, and this increases the time lag, and makes the switching really sssssssslllllllllloooooowlwwwww - so slow, that you can sometimes feel the car speeding and slowing if you hold the throttle pedal really still.
Number_Cruncher
|
|
|
|
|
