[Escape]

I've been wondering what data the trip computer uses to calculate the economy. On Prue the computer always indicates way too much, on Piper it just seems unreliable. I always calculate myself when filling up, and this rarely matches the indicated consumption.

Obviously the distance travelled is one of the inputs. But I would think the fuel rate would be too small and with too much variations to be of use. Not sure if the GEMS actually knows the fuel rate or if it just adjusts from it's base set according to the current operating parameters.
Perhaps airflow could be used, considering the ideal mixture is 14.7kg of air for every kg of petrol.

Any thoughts welcome! :wink:

 

[RRToadHall]

Unless you reset your trip computer for each tank that you are manually calculating the numbers would always be off.

 

[waveydavey]

Mine always seems pretty good; I m fairly sure it woks on the the fuel meter but then maybe that is better with the DDE?

 

[Keijo]

Good question escape!

Distance travelled is the easy one, direct from odometer data.
I think the computer would also have a good handle on the fuel consumption though.
If you think about it, each shot out of an injector is a metered amount of fuel, so as to optimize the combustion for the current conditions. The metering is based on the duration of the injector firing pulse and a fixed fuel rail pressure.
So if the computer knows exactly how much fuel it is squirting out, it only needs to accumulate the total injected for whatever period your trip computer is accumulating the data for.

Having said all that however, the amount of fuel dispensed each shot is a function of the following:
The duration of the injector pulse.
The fuel pressure in the rail.
The injector bore size

The injector size remains the same at all times unless someone has replaced them.
If your fuel pressure regulator is out of spec' or playing up for example the data will be incorrect.
Lets say the fuel pressure is low by 10% of factory spec'. There is no fuel pressure sensor in the P38.
In such a case, your car will still run correctly, because the computer will extend the injector pulses by around 10% based on oxygen sensor data, but your trip computer will be out because it is now calculating the extended pulse, compounded with the original factory spec' fuel pressure.

Having said all that though, comparing the trip computer to a manual fill is also fraught with difficulty.
It's extremely difficult to get a consistent fill even when fully topping up, as there is always the possibility of some air entrapment in the tank at different parking angles, even when you think you are essentially level.

Then again, maybe the bowser is wrong and we are all getting ripped off at fill time

Cheers

 

[Col]

Would imagine the size of your Right Foot would have to come into the equation somewhere.

Cheers

 

[Escape]

Good question escape!

THX! :wink:

First of all, I'm not to bothered about it, just something I've been pondering about. Even if the actual figures are not correct, it still gives an indication of consumption (for example that the misses uses more fuel than I do, wether that is due to a heavier right foot or less favorable driving circumtances remains open to debate. :mrgreen: ).

I usually reset the trip computer when filling up, and if I don't, I combine the manual calaculations for that stage to compare.
These calculations do seem to be pretty accurate, I'm usually filling up at the same station and pump (not so much by design as by convenience :wink: ), thus taking some variables out of the equation.

Distance travelled is the easy one, direct from odometer data.
I think the computer would also have a good handle on the fuel consumption though.
If you think about it, each shot out of an injector is a metered amount of fuel, so as to optimize the combustion for the current conditions. The metering is based on the duration of the injector firing pulse and a fixed fuel rail pressure.
So if the computer knows exactly how much fuel it is squirting out, it only needs to accumulate the total injected for whatever period your trip computer is accumulating the data for.

Having said all that however, the amount of fuel dispensed each shot is a function of the following:
The duration of the injector pulse.
The fuel pressure in the rail.
The injector bore size

The injector size remains the same at all times unless someone has replaced them.
If your fuel pressure regulator is out of spec' or playing up for example the data will be incorrect.
Lets say the fuel pressure is low by 10% of factory spec'. There is no fuel pressure sensor in the P38.
In such a case, your car will still run correctly, because the computer will extend the injector pulses by around 10% based on oxygen sensor data, but your trip computer will be out because it is now calculating the extended pulse, compounded with the original factory spec' fuel pressure.

Distance travelled is indeed an easy one. Even when using bigger tyres, a simple check with the GPS will allow the necessary compensation.

As for fuel, I'm not convinced the GEMS ECU actually knows how much fuel it is injecting. First of all, fuel pressure is not exact, so the trip computer would have to be calibrated for the specific fuel pressure regulator used (because, as you pointed out, there is no fuel pressure sensor).
Secondly, some variables might change over time.
I believe the GEMS just starts from a base map, and shortens or lengthens the injection pulses as it sees fit, without bothering about the exact amount of fuel used. Off course it would be possible to deduce the fuel used in adapted pulse by comparing it with the amount that would be used with the base setting, but this would seem to cumulate a lot of errors.

Greetz,

Filip

 

[waveydavey]

Accordign to Rave (Page 61 of the workshop manual) there is a fuel used signal just for the trip computer.

 

[Escape]

Accordign to Rave (Page 61 of the workshop manual) there is a fuel used signal just for the trip computer.

Good one! I hadn't been looking in RAVE yet (haven't got it at work), but was planning to. :wink:
Next question would if this fuel used signal comes from a dedicated fuel flow sensor (which seems unlikely) or is given by the ECU/BECM to the trip computer. If the latter, it still needs some basis for the calculation.

 

[Hoges]

I suspect the displacement of the fuel sender from "full" is also included in the computer calculation mix... when mine was playing up due to gunk on the sender (from ethanol contamination :evil: ) and the gauge wouldn't read above half full even with a full tank, the "Distance to empty" was only 50% of what it should have been and the Av consumption was also wildly out...

 

[Escape]

I suspect the displacement of the fuel sender from "full" is also included in the computer calculation mix... when mine was playing up due to gunk on the sender (from ethanol contamination :evil: ) and the gauge wouldn't read above half full even with a full tank, the "Distance to empty" was only 50% of what it should have been and the Av consumption was also wildly out...

I thought about a link with the fuel sender, but assumed it unlikely because of the huge margin on the reading. It seems logical it is used to calculate the distance to empty, but surely the average consumption must have anothers bases?

 

[paul.adshead]

Trip economy data comes from the engine management computer. The ECM sends to the instrument cluster, something like 12,000 impulses, per litre of fuel used. The instrument cluster is supplied road speed (well, axle rpm), and thus can work out economy. Also, using the fuel tank level sensor, the distance to empty/range can be (roughly) calculated. It's all in RAVE (search for "C505").

Different: tyre sizes, fuel injectors, or fuel pressure regulator, will all affect road-speed/economy accuracy.

 

[Keijo]

Escape wrote: As for fuel, I'm not convinced the GEMS ECU actually knows how much fuel it is injecting. First of all, fuel pressure is not exact, so the trip computer would have to be calibrated for the specific fuel pressure regulator used (because, as you pointed out, there is no fuel pressure sensor).

Paul wrote: The ECM sends to the instrument cluster, something like 12,000 impulses, per litre of fuel used. Different: tyre sizes, fuel injectors, or fuel pressure regulator, will all affect road-speed/economy accuracy

Absolutely right! But the injectors will still pump out the right amount of fuel even though the above statements are correct.

All control systems use the same overall principle.
ie, Model the system and then trim with feedback.

In a typical engine management system such as the P38 for example:
The instantaneous fuel quantity to be injected is initially calculated by the GEMS from input parameters such as Fuel pressure, Mass Air Flow, Throttle Position, Coolant Temperature, Intake Air Temperature etc.
If all those parameters are accurate and the engine and combustion model is 100% correct, then nothing else needs to be done.

However in the real world, there will always be errors in some or all of the input data.
That's why the control loop is closed with feedback, via the heated oxygen sensors which then trim the overall system based on the actual emissions, to correct and optimize the model. These trim values change slowly and are retained in the system long term.

So essentially the GEMS is fault, drift and error tolerant to the limits of the feedback loop.

The same is not true for the trip computer however as there is no feedback system so a change in any input parameter will result in an error.

I personally am happy with the "approximate" values I get from the trip computer, after all it's only meant to be an aid to trip planning, and controlling driving habits.

Escape, thanks for opening up an interesting topic of discussion.

Cheers