Any automotive engineers around?

Status
Not open for further replies.
Joined
Jul 22, 2005
Messages
11,948
Location
NorthEast
I am trying to understand engine parameters on a typical OBD system. Let us take an example with 3.0 V6 engine with MAF. I can monitor bunch of parameters such as air intake in gm/sec, fuel consumption gal/hour, load in percent, rpm, throttle position sensor, fuel trims etc.

Which sensor monitors the engine load? As it turns out there is no single sensor. The load is "computed" by the ECM using other sensors such as rpm, TPS, MAF. What confuses me is that the ECM uses this computed load to determine the ignition timings and fuel injector pulse width.

For my engine, how would I compute fuel consumption? I understand that UltraGauge/ScanGauge type instrument use MAF and engine size to compute fuel consumption by assuming average 14.7 stoichometric ratio.

Engine air intake consumption is 3 liter every other revolution. So at 600 rpm idle, it is ingesting 300 * 3 liters / minute = 15 liter/sec. What is the weight of a liter of air? One cubic meter of air weighs about 1.2kg which means 1 liter of air weighs 1.2 gm (at room temp / sea level)

So my engine should be taking in (1.2 * 15) = 18 gms/seconds of air.

Unfortunately, the actual MAF reading is at 3.0 gm/sec. The shop manual says that the range is 2--6 gm/sec for the idle MAF. The 3x range also surprise me. It tells me that the ECM does not really care about the actual MAF value but only the relative value. I do not understand how that works.

I also understand that I am missing something important because my calculation is based upon *only* RPM and nothing else. We know for sure that fuel consumption shoots up very high if the engine is under load at the same rpm.

How do I account for that?

- Vikas
 
The engine is operating almost all the time at partial pressure -not at STP. Therefore you cannot calculate the airflow rate at partial throttle by using swept area.
 
I also firmly believe the "Load avg %" as reported in the generic data is a bias and not a literal thing you could measure on a dyno.

So you might see 25% load at idle. To get down near zero you need to be decellerating.
 
Airflow mass, throttle angle, RPM, O2 sensor etc is used to calculate injector pulse width. Total injector pulse width is the most direct way to measure fuel consumption.
 
Originally Posted By: ARCOgraphite
The engine is operating almost all the time at partial pressure -not at STP. Therefore you cannot calculate the airflow rate at partial throttle by using swept area.

This.

You need to have a vacuum reading as well, in order to calculate grams of air.
 
If you're looking for one sensor that gives the best indication of engine load at any RPM, it is the MAP sensor. Engine torque output at any speed is proportional to intake manifold absolute pressure.

If your engine doesn't have a MAP sensor, then it must be calculating fuel quantities based on cylinder air quantities. This would be the total airflow reading from the MAF sensor divided by the number of cylinders breathing in one engine revolution and the engine speed.

Your engine would only be breathing 18gm/sec of air at 600 rpm if it were at Wide Open Throttle, had no induction pressure losses, and had a volumetric efficiency of 100%.

If you want to calculate fuel consumption, can't you just use the fuel flow reading out of the ScanGauge?
 
Originally Posted By: Vikas
Unfortunately, the actual MAF reading is at 3.0 gm/sec. The shop manual says that the range is 2--6 gm/sec for the idle MAF. The 3x range also surprise me. It tells me that the ECM does not really care about the actual MAF value but only the relative value. I do not understand how that works.


On an OBD-II vehicle, the ECM will typically take inputs from the oxygen sensors above and below the catalytic converter and continually adjust the A/F ratio around stoichiometric - think of the way a thermostat cycles to maintain temperature in a relatively narrow range. The absolute reading from the MAF sensor is fairly inconsequential, but the ECM needs that input to have a rough starting point to know where to set the injector pulse width, and it needs to know that if the MAF sensor indicates a 50% increase in flow rate that it should bump up the fuel delivery in proportion.
 
This vehicle does NOT have MAP sensor. I have always been under impression that MAF sensors are more accurate and lately most manufacturers use MAF than MAP. As a matter of fact, UltraGauge use MAP value and computes MAF by using "several engine sensors such as MAP, RPM, vehicle speed and others".

Fuel flow readings are not available on OBD-II (at least the early 2000 implementation). ScanGauge/UltraGauge computes it based upon MAP/MAF.

So let us assume I have 20" of vacuum at idle with TPS reading 0%. Can we derive the total mass of the air intake?
 
No. You don't have enough information to calculate engine airflow. You would also need to know intake manifold temperature, and the volumetric efficiency of the engine at that operating point.

If you want to calculate fuel consumption, it is bedrock certainty in an EFI system that the final commanded fuel injector pulse width is proportional to volumetric fuel flow. (You can't rely on mass airflow because the air/fuel ratio is not always 14.7:1. During warmup, fuel enrichment is required, and at high engine load, enrichment is also required.) Even my '85 Corvette had an MPG readout that gave instantaneous and average fuel economy. It worked by totalizing the amount of time that the injectors were open to get a fuel flow reading in gallons/hour, then dividing that into the vehicle speed. Miles/hour divided by gallons/hour yields miles per gallon.
 
I don't know what kind of car you have. it may have all fancy stuff in achieving the high MPG mandated by gov. It'll be very hard to figure out the fuel consumption at idle. you need to consider if your car has VCM (Vari Cyl Mgmt), EGR, VVT ( which will make your 3.0L maybe scaling down to 1L by pumping back fresh charge thru intake valve out of the cylinder), etc.

You could probably install flow meter to your fuel line.

Good luck figuring out this puzzle.
 
Please make as many *reasonable* assumption as possible and take that to analytical calculation. For example assume sea level and 70F air temperature.

My car does NOT have MAP, so it does not measure the vacuum. Also my car has nothing fancy such as cylinder deactivation etc.

I know which parameters are available for the scanner on my car. And using them, ScanGauge/UltraGauge comes up with rough approximation of the fuel usage. It obviously makes some reasonable real life assumptions. I am interested in reverse engineering the algorithm.

The impetus for this exercise started when I realized that people had figured out a way to find if their MAF was working or not based upon the instant fuel consumption numbers spit out by the on board dash display. Even accounting for different engine size, the MAF readings of the two cars seem to be wildly different. Theoretically, they should have been proportionate to their engine size since both the engines are NA.
 
Last edited:
I agree with A_Harman, you need some more data.

Are you comparing the MAF at idle, or at some point with the car moving?

MAF is a measure of total airflow through the engine, so anything that would cause more air to flow (higher load) through it will increase the number. There are also big differences in EGR levels, engine idle speed, accessory drive load, alternator duty cycle, A/C on or off, heated seats on/off, DRL turning on, etc. These are probably throwing the readings you were looking for out of range.

All of my cars will pitch a fit if the MAF is unplugged, so I'm not sure about determining the functionality of the sensor by using the instant MPG reading, that assumes the Air/Fuel ratio is constant, among too many other things. MAF sensors are also only able to measure the air that goes through the sensor, sometimes leaky ductwork lets in "extra" air.

The older 3800 engines did not have a MAP sensor, but you can calculate it if you have everything else resulting in a synthetic MAP sensor.

If you have an OBD II scanner, why are you worried about being able to tell if your MAF is working?
 
More of curiosity than anything else; I want to learn and in general I do like to understand how a gizmo works. The maximum MAF reading that I have obtained in my car is just over 90 gm/sec; this at WOT flat out. At idle it will hover around 2.5 depending upon the exact idle speed and other accessories and/or cooling fan etc. On my scanner, I have a screen to give me MAF and gallon/Hr display and I have taken few pictures. My MAF and G/H readings have straight line relationship. At 90 gm/sec, I will be getting 10 G/H consumption (according to the UltraGauge) That is about 38 Liter/Hour at WOT in my car. The M5 folks are getting 130 Liter/Hour when their MAF is good and under 100 Liter/Hour when it is old.

I am trying to understand the difference between 38 vs 130 on two different NA engines.

Let us make some real life assumptions. I am all game. I think this is more fun than talking about why USPS needs to go down etc :)

- Vikas
 
Can somebody check my calculations? I am not getting the expected answer.

MAF readings 2.52 gm/sec = 2.52*3600 = 9.072 kg/hr
Assuming 14.7 ratio, fuel 9.072/14.7 = 0.617 kg/hr
Assuming a liter of fuel weighs 1 kg, => 0.617 liter/hr
= 0.162 G/hr

However, the UltraGauge is reading 0.28 G/hr and I believe that is the right number.

- Vikas
 
Originally Posted By: Vikas
More of curiosity than anything else; I want to learn and in general I do like to understand how a gizmo works. The maximum MAF reading that I have obtained in my car is just over 90 gm/sec; this at WOT flat out. At idle it will hover around 2.5 depending upon the exact idle speed and other accessories and/or cooling fan etc. On my scanner, I have a screen to give me MAF and gallon/Hr display and I have taken few pictures. My MAF and G/H readings have straight line relationship. At 90 gm/sec, I will be getting 10 G/H consumption (according to the UltraGauge) That is about 38 Liter/Hour at WOT in my car. The M5 folks are getting 130 Liter/Hour when their MAF is good and under 100 Liter/Hour when it is old.

I am trying to understand the difference between 38 vs 130 on two different NA engines.

Let us make some real life assumptions. I am all game. I think this is more fun than talking about why USPS needs to go down etc :)

- Vikas


MAFS, plural. The M5 has two.

Interesting why you are using that as a benchmark? The engine makes 400HP, has 12:1 compression and subsequently moves a heck of a lot of air.

A MAF is a heated element. As the element gets contaminated, the contaminants build up on the filament and act to insulate it. With the filament insulated, it requires less current to keep it at temperature, and subsequently, this skews downward the reading the MAF provides.
 
Originally Posted By: Vikas
Can somebody check my calculations? I am not getting the expected answer.


You're taking the output from the MAF sensor as Gospel - I highly doubt it's ever calibrated beyond determining the general parameters of the voltage vs. flow rate curve. It would be better if they didn't try to apply a mass per time dimension to the output of the MAF and just used the voltage output.
 
Come to think of it, MAF only measures a fraction of the air going through the intake as the MAF is situated in side a tiny cylinder.

BUT that does NOT explain how the UltraGauge able to figure out the approximate fuel consumption. The only thing I fed to UltraGauge is the engine size and the fuel tank size. Everything else it got from the car ECM and I know each and every parameter reported by the ECM. ECM does NOT report any conversion factor.

- Vikas
 
Last edited:
Originally Posted By: Vikas
Come to think of it, MAF only measures a fraction of the air going through the intake as the MAF is situated in side a tiny cylinder.


Now you're getting there...

If you were to try to precisely calibrate a MAF sensor, you'd have to know exactly what the flow profile inside the intake duct looks like, and make sure that you know exactly where across that flow profile you're locating the probe for the MAF sensor. I'm assuming that you'd generally see laminar flow inside the intake duct - turbulent flow would actually simplify matters, but you'd still have to know exactly at what engine speed and throttle opening you see laminar to turbulent transition.

The M5's intake duct is probably designed to operate in turbulent through more of the engine speed and throttle opening range - the insanely high flow rate readings are a result of the MAF probe being cooled better by turbulent flow, not of actually moving that much air. As long as the car's ECM knows throttle position and RPM, the MAF can say whatever it wants, but the ECM will adjust the actual mass flow rate in its calculation of injector pulse width.

...and the fact that you had to feed the UltraGauge the fuel tank size tells me that it's only fudging numbers. It's like saying that you have to know the size of your municipal reservoir to get the flow rate of water out of your kitchen faucet.
 
Ultragauge does not use the fuel tank size to perform any calculations beyond "fuel level (gallons)" which it probably gets from the fuel level sender, and "distance to empty".
 
Status
Not open for further replies.
Back
Top