[DerAlte]

Originally Posted by Jeroen

... but essentially they are all indirect measurement. e.g. it derives the amount of fuel used from a theoretical model. The amount of fuel injected is not necessarily linear with the duration of the pulse. Engine loading, fuel pressure etc will all play a role. I'm no expert, but they probably use an approximate or simplified model to calculate the fuel injected. So that introduces an error by itself. ...

The method of fuel calculation is neither theoretical, nor approximate, really. At least not in Common Rail. With >1200bar on tap, duration for which the needle is retracted at the injector governs the volume of fuel that flows through the injector orifices, assuming a rectangular waveform (sharp on and off). Critical assumptions are lack of cavitation and vaporization inside the injector cavity.

At this level of fuel flow (picolitres per second), conventional flow measurement will not show any flow - but will add a point of failure in the injection path. One can't measure flow before the rail, as there is a return line at the rail.

Actually, the injector needle retraction does not follow a rectangular waveform due to the fact that
a. The solenoid is an inductive load, and will never produce a pure rectangular movement
b. The spring behind the needle may have a non-linearity (however small)
c. The injection pattern (waveform) is a series of nozzle opening commands for each cylinder firing

All this is type-tested (a few engines on test bed) and calibrated before the engine / variant is released for production. In calibration, one uses a graduated cylinder to correlate the fuel consumption over x million revolutions with the design figure. The design figure, modified by the calibration factor produced in testing, is used in the injection map creation. This calibrated figure is pseudo-linear over injector needle opening durations.

The big IF here is that IF the engine ECU reports the actual per firing consumption to the MID ECU, the consumption figures will match, BUT it doesn't because that will overload the communication mechanism (CAN) AND the low-MIPS-power MID ECU. What gets conveyed is an average over a couple of seconds worth of revs, and this averaging can introduce a cumulative error. The error is not expected to be constant + or -.

In actual driving, one can see the MID figure going up or down logically w.r.t. the driving conditions, i.e. higher consumption in lower gears while accelerating, lower consumption in sustained level road cruising, even lower consumption on a gradual down-hill coasting - at all speeds.

[Jeroen]

Originally Posted by DerAlte

The method of fuel calculation is neither theoretical, nor approximate, really. At least not in Common Rail. With >1200bar on tap, duration for which the needle is retracted at the injector governs the volume of fuel that flows through the injector orifices, assuming a rectangular waveform (sharp on and off). Critical assumptions are lack of cavitation and vaporization inside the injector cavity.

At this level of fuel flow (picolitres per second), conventional flow measurement will not show any flow - but will add a point of failure in the injection path. One can't measure flow before the rail, as there is a return line at the rail.

Actually, the injector needle retraction does not follow a rectangular waveform due to the fact that
a. The solenoid is an inductive load, and will never produce a pure rectangular movement
b. The spring behind the needle may have a non-linearity (however small)
c. The injection pattern (waveform) is a series of nozzle opening commands for each cylinder firing

All this is type-tested (a few engines on test bed) and calibrated before the engine / variant is released for production. .

Thanks,
I've done a bit of design work myself on diesel injection systems but it's been a while. In marine applications in the old days we actually did use flow based system in the feed and return line, so we could nett.

I've worked with system whereby we just used to calculate the theoretical volume of injected fuel. Even there we used some correction tables based on actual testing.

But as you state, the MID is likely to show an error as it shows an average reading over a certain amount of samples.

The MID on my Jaguar shows two readings when it comes to fuel flow/useage.
It shows the actual usuage, ie as discussed above, but it also shows the total useage based on the last reset of the trip computer. For the first 10000 kilometers or so I recorded all the readings of the trip computer and the actual fuel I put in the car.

It also shows the amount of fuel used and the range still left.

My conclussion were that in my case I was getting on average 10% better fuel consumption than the MID told me I was getting. So it also meant that when it told me I had zero range left I could easily drive another 30-40 miles, which used to make my poor wife very nervous.

The 10% was very consistent. No matter how I drove (sedate or pedal to the metal) or where I drove (city or highway cruising) the result was pretty much the same. Always 10% off.

I don't know if they get calibrated that way. I assume so, because it's on the safe side of cautious from the manufacturer's point of view.

Jeroen

[DerAlte]

Originally Posted by Jeroen

... I don't know if they get calibrated that way. I assume so, because it's on the safe side of cautious from the manufacturer's point of view. ...

Calibration is the wrong factor to be questioned. As I mentioned, the calculation of quantity of fuel to be injected is done *very* correctly and accurately, otherwise engine performance would be something else. This is *not* relative - quantity computed from the map is an absolute value ultimately, since the injector is an absolute return-to-zero device (needle is spring loaded).

The likely culprit is the MID itself - it is a separate system, which only reads values from the CAN bus
- the average (not instantaneous per rotation) fuel consumed value, broadcast on by the engine / injection ECU
- the distance traveled comes from the Transmission ECU (in simpler systems without a Transmission ECU, this is broadcast by the injection ECU which gets the pulses from a transducer on the output side of the gearbox)

The MID uses these values to display the consumption figures. What formula, or bias, is used in this display algorithm, and why - only the MID designer can say.

[Jeroen]

Originally Posted by DerAlte

Calibration is the wrong factor to be questioned. As I mentioned, the calculation of quantity of fuel to be injected is done *very* correctly and accurately, otherwise engine performance would be something else. This is *not* relative - quantity computed from the map is an absolute value ultimately, since the injector is an absolute return-to-zero device (needle is spring loaded).

The likely culprit is the MID itself - it is a separate system, which only reads values from the CAN bus
- the average (not instantaneous per rotation) fuel consumed value, broadcast on by the engine / injection ECU
- the distance traveled comes from the Transmission ECU (in simpler systems without a Transmission ECU, this is broadcast by the injection ECU which gets the pulses from a transducer on the output side of the gearbox)

The MID uses these values to display the consumption figures. What formula, or bias, is used in this display algorithm, and why - only the MID designer can say.

Thanks, this is what I meant to. I.e. what the MID displays based on the input from the ECU I call "calibration". Against a very precise and determined input (ie. from the ECU) the output is set.

In my case with a 10% too pessimistic reading!

Jeroen

[amalji]

Originally Posted by DerAlte

the calculation of quantity of fuel to be injected is done *very* correctly and accurately, otherwise engine performance would be something else. This is *not* relative - quantity computed from the map is an absolute value ultimately, since the injector is an absolute return-to-zero device (needle is spring loaded).

The likely culprit is the MID itself - it is a separate system, which only reads values from the CAN bus
- the average (not instantaneous per rotation) fuel consumed value, broadcast on by the engine / injection ECU
- the distance traveled comes from the Transmission ECU (in simpler systems without a Transmission ECU, this is broadcast by the injection ECU which gets the pulses from a transducer on the output side of the gearbox)

The MID uses these values to display the consumption figures. What formula, or bias, is used in this display algorithm, and why - only the MID designer can say.

So, my take away from this is

1) ECU is 'aware' of the exact amount of fuel injected into the engine.
2) The error happens at the algorithm used on the MID side.

If that's the case, if I get raw data from the ECU using ELM 327 and process it, I should be getting accurate fuel consumption figures, right ?

Also, do you recommend any android app which you feel is accurate in interpreting the data that ECU supplies through the OBD2 port ? I use torque pro. But, there is a discrepancy between

• Average mileage showed by Torque pro &
• Average mileage calculated by dividing distance covered ( shown on torque pro ) by fuel consumed ( shown on torque pro )

The weirdest thing is that the error is not always one over the other. It can be either way. sometimes the first method gives a higher kmpl than the second and it could be the reverse as well.



So, where do you think the error could be creeping in ?

[DerAlte]

Originally Posted by amalji

... If that's the case, if I get raw data from the ECU using ELM 327 and process it, I should be getting accurate fuel consumption figures, right ? ...

No, you won't. Both the MID or the OBD scanner (ELM327) are reading the same reading on CAN bus - which is average consumption over a largish period (a few seconds - equal to the update rate on CAN bus).

And, any and every App will show the same behaviour, as long as they are reading the OBD scanner which is reading average consumption over a large period.

Caution: It would be very simplistic to treat the subject with instantaneous values. The consumption / mileage calculation is based on Moving Average, and there are a few variants of the same, including the autoregressive one. The computed value from any software depends on the statistical model used.

I am not a good enough mathematician to do a deep dive in the subject - maybe there are others on the forum who are good mathematicians / statisticians who can explain the phenomenon. I am happy to just follow the trend (not absolute value) of mileage FT-to-FT every few full-tank-fillings - I only fill to full from the warning light point.

I am monitoring mileage trend, which is to say whether the mileage is increasing, constant or decreasing. I would do the same if I had an MID in the car - with values taken every 1K km. I can do something when the trend is decreasing (the other 2 cases are perfectly OK for me) - correlating with tyre pressure, reviewing driving habits, even to the extreme of checking injector health. I can do precious little with Absolute Mileage (with NA carb cars we could), so why bother - less stress on the mind.

[amazingGuru67]

Hello friends,

I own a Honda amaze S mt diesel for the past one month.

This post is regarding my trip meter problem that is common in many car companies across the world.

Honda Amaze's trip meter is a little optimistic in reading the distance. I measured 100meters on a measuring tape and have reset my trip meter to zero. When I drove across the marked line, trip meter read as 100m at the 80m mark itself.

The same happened with my friend's Mahindra Quanto.

I am worried because, ECU in Amaze calculates the average kmpl based on trip meter's readings only. Is this the reason why the kmpl readings are optimistic than actual real life data in tank full to tank full method?

This gives me goose bumps friends...

Can any one else check and let me know if this trip meter malfunction is normal or not?

My car is due for it's service next week. Will bring up this issue at the first place and let you know.

If there is any other problems found in your Amaze, feel free to post it in here.

Let other Amaze owners help you and in turn help ourselves.

Regards,
Guru.

[drsnt]

Every car i have & had and ever driven in, each had a optimistic speedometer. it varies from 10% to 4%. I had used GPS in my different mobiles Nokia/Samsung along with different GPS apps like GPS essentials.showing detailed information. I have never measured trip meter but i always thought that the error should be in same ration to speedometer.
But 80:100 is 20% and huge. I would advise to take car on a longer route and measure speedometer and trip/odometer error with help of GPS. 100 meters is very small distance. go for drive of minimum 10kms to get error ratio.

[a4anurag]

Originally Posted by amazingGuru67

Honda Amaze's trip meter is a little optimistic in reading the distance. I measured 100meters on a measuring tape and have reset my trip meter to zero. When I drove across the marked line, trip meter read as 100m at the 80m mark itself.

Guru bhai,

I just have a few doubts on this:

1) Does tyre pressure meter readings? I mean a under-inflated or an over-inflated tyre would give you different reading?

2) Have you upsized the tyres by any chance post the purchase?

If the above two problems don't affect the reading you get by the odometer then it definitely is a problem with the odometer.

Originally Posted by drsnt

Every car i have & had and ever driven in, each had a optimistic speedometer. it varies from 10% to 4%. I would advise to take car on a longer route and measure speedometer and trip/odometer error with help of GPS. 100 meters is very small distance. go for drive of minimum 10kms to get error ratio.

+1. Try all this measurement in a long distance to get correct readings.

Even I tried using GPS and ELM 327 OBD II reader for checking the speeds.

What I found was:

Speedometer: 100 kmph,
OBD II Reader: 94 kmph,
GPS via OBD II reader: 96 kmph.

Anurag.

[Rudra Sen]

Originally Posted by a4anurag

Speedometer: 100 kmph,
OBD II Reader: 94 kmph,
GPS via OBD II reader: 96 kmph.

How accurate are these GPS devices? I'm asking this because my Garmin nuvi 2565LM shows exactly the same speed as my car's speedometer.

[a4anurag]

Originally Posted by Rudra Sen

How accurate are these GPS devices? I'm asking this because my Garmin nuvi 2565LM shows exactly the same speed as my car's speedometer.

This is not a specific GPS device but a sensor that reads speeds from the ELM OBD II reader. So there is difference in all three ways of measurement.



The above one shows the GPS speeds and second one is the speed the ECU measures and reads out. At this time the speedometer shows 100 kmph.

Now is there any way to standardize / calibrate this?

Anurag.

[amazingGuru67]

Originally Posted by drsnt

Every car i have & had and ever driven in, each had a optimistic speedometer. it varies from 10% to 4%. I had used GPS in my different mobiles Nokia/Samsung along with different GPS apps like GPS essentials.showing detailed information. I have never measured trip meter but i always thought that the error should be in same ration to speedometer.
But 80:100 is 20% and huge. I would advise to take car on a longer route and measure speedometer and trip/odometer error with help of GPS. 100 meters is very small distance. go for drive of minimum 10kms to get error ratio.

I can check it with the government distance board signs and compare it with my trip meter. But, it is also not accurate.

Originally Posted by a4anurag

Guru bhai,

I just have a few doubts on this:

1) Does tyre pressure meter readings? I mean a under-inflated or an over-inflated tyre would give you different reading?

2) Have you upsized the tyres by any chance post the purchase?

If the above two problems don't affect the reading you get by the odometer then it definitely is a problem with the odometer.

Anurag.

Hi Anurag,

I have not checked my tyre pressure till today. It's been more than a month and around 1000 kms.

I have not upsized my tyres till today.

Let me check this with the dealer next week about this problem.

[whitewing]

Originally Posted by a4anurag

This is not a specific GPS device but a sensor that reads speeds from the ELM OBD II reader. So there is difference in all three ways of measurement.
Now is there any way to standardize / calibrate this?
Anurag.

Actually there would be no way to calibrate the GPS reading with the Speedo, since the GPS does not factor in the road topology. Imagine speeding on a spiral and if your GPS sensitivity range encompasses the spiral your effective speed would be 0!

Originally Posted by amazingGuru67

I can check it with the government distance board signs and compare it with my trip meter. But, it is also not accurate.

If you have access to drive in an local athletic track, then a few spins should do the trick??

[k_nitin_r]

Originally Posted by drsnt

Every car i have & had and ever driven in, each had a optimistic speedometer. it varies from 10% to 4%.

Originally Posted by a4anurag

Now is there any way to standardize / calibrate this?

The speedometer in the car is intentionally made by the vehicle manufacturers to display a higher speed than the vehicle is actually traveling at to avoid lawsuits from the variance that could occur from changes in tire pressure or other parameters. In the words of many software firms: "This is not a defect, but a feature!"

[a4anurag]

Originally Posted by amazingGuru67

Hi Anurag,

I have not checked my tyre pressure till today. It's been more than a month and around 1000 kms.

I have not upsized my tyres till today.

Let me check this with the dealer next week about this problem.

Not checked the pressure for a month / 1000 kms. That's bad bro. You'll have to keep an eye on it. Do check it minimum once a week at least. Mileage, tyre wear, suspension damage improper ride all can be saved by keeping the tyre pressure on check regularly.

Anurag.