Ok folks, for some time now I've been trying to understand some of the "stuff" around how the modern ECU controlled petrol engine works. Not the mechanical stuff, like pistons, crankshaft, valves, camshafts etc, but the way in which the electronic units interact with the engine. So I thought I'd run this past you all in case I've got it all wrong? Please do feel free to comment freely.
As I see it, in general, There will be an engine. an ECU (sometimes more than one). Sensors (coolant temp, O2 sensors, MAP/MAF, Knock, TPS, etc) which send signals to the ECU. and Actuators (Fuel injectors, throttle body stepper - if fly by wire, etc) these are devices actuated by the ECU to directly influence the engine running parameters ie. RPMs, mixture, etc.
The ECU will have a fueling map program (let's forget about ignition at this time) embedded in it's memory which is approximately right to make the engine run. It will allow enrichment for cold starts and a basic mixture control for hot running. This is why the engine will continue to run if you pull the plug off the MAP sensor (for instance). What it can't do though is to "fine tune" the parameters so that maximum economy and minimum emissions are achieved.
This is where the "clever" stuff starts. From a cold start the ECU will be "watching" the signal from the coolant temp sensor to indicate normal working temperature has been achieved. Up 'till now the ECU has been controlling fueling from the embedded fueling map alone - this is "open loop". When the ECU detects the coolant temp sensor signal is satisfactory it will go to "closed loop" operation. In open loop the ECU will largely be deciding fueling from the throttle position (via it's sensor - TPS) and the signal from the MAP (or MAF if it has one) sensor. The TPS tells it how fast you want it to run the engine and the MAP/MAF allows it to calculate the volume of air being ingested so how much fuel is needed to be injected to achieve this. In closed loop it's still doing this but now it also looks at the O2 (Lambda) sensor signal which allows it to "fine tune" the course settings.
So how does it do this? First thing to understand is that the Oxygen sensor (I'm not looking at wide band fuel ratio sensors here - still learning about them in case I ever get my Civic!) is like a wee battery which can produce about one volt of electricity and it does this depending on the amount of unburnt oxygen it senses in the exhaust gasses (it is located upstream of the catalytic converter so is "sniffing" unadulterated exhaust gasses as they leave the cylinders. The sensor needs to be REALLY hot before it will work) A richer mixture will send it towards the 1 volt and a weak mixture sends it towards 0 volts. In practice it fluctuates around once to twice a second between around 0.9 and 0.1 volts. Difficult to pick up on a multimeter, you really need to graph it with a scanner (MES is ideal). So the ECU is trying to achieve an air/fuel ratio of around 14.7% air to fuel (this is by weight so you're talking a lot of air to very little fuel). This ratio uses all the oxygen to burn all the hydrocarbons in the fuel so minimum pollution occurs.
But why does this voltage fluctuate? Well, it's the ECU trying to achieve the 14.7% ratio. Lets say the mix starts to drift slightly rich? the voltage from the O2 sensor will start to rise, the ECU "sees" this so cuts back the fueling by a very small amount on the next injection and it continues to do this miniscule cutting back on each subsequent injection until it starts to "see" a weaker mixture indicated by the voltage falling. If the voltage falls too much the ECU will enrichen the mixture again and so it goes on with the swing being as I said in the above paragraph (around 0.1 to 0.9 volts).
Now we come to fuel trims and this is where I am in my climbing of the ladder of knowledge at this time. I believe these very small fueling adjustments (measured in %ages) are what present as "Short Term Fuel Trim" (STFT). and you can watch them real time to get an idea what's going on with the fueling. There is also "Long Term Fuel Trim" (LTFT) and my understanding of that is that it's the baseline from which the ECU is calculating it's fueling so you would expect to see it at or very close to zero on an engine in good mechanical condition which is working well. However if the STFT is consistently, just for example, tending towards the rich side and the ECU can't pull it back with the very small fuel variation it can achieve undet STFT it will then, after trying for a while, reset it's LTFT to the average richer figure and the STFT settings will then be zeroed on the new LTFT?
So, form a fault finding point of view I think you can say that LTFT should be around Zero on a good engine. If however it's showing, again for instance, around 15, maybe 29% or whatever, you've got a problem. If it's +20% then you're rich, if -20% then you're weak. so you can start looking for maybe leaking injectors, inlet gasket leak etc?
So what do you think everyone? am I deluding myself or have I got the "stick" by the right end?
As I see it, in general, There will be an engine. an ECU (sometimes more than one). Sensors (coolant temp, O2 sensors, MAP/MAF, Knock, TPS, etc) which send signals to the ECU. and Actuators (Fuel injectors, throttle body stepper - if fly by wire, etc) these are devices actuated by the ECU to directly influence the engine running parameters ie. RPMs, mixture, etc.
The ECU will have a fueling map program (let's forget about ignition at this time) embedded in it's memory which is approximately right to make the engine run. It will allow enrichment for cold starts and a basic mixture control for hot running. This is why the engine will continue to run if you pull the plug off the MAP sensor (for instance). What it can't do though is to "fine tune" the parameters so that maximum economy and minimum emissions are achieved.
This is where the "clever" stuff starts. From a cold start the ECU will be "watching" the signal from the coolant temp sensor to indicate normal working temperature has been achieved. Up 'till now the ECU has been controlling fueling from the embedded fueling map alone - this is "open loop". When the ECU detects the coolant temp sensor signal is satisfactory it will go to "closed loop" operation. In open loop the ECU will largely be deciding fueling from the throttle position (via it's sensor - TPS) and the signal from the MAP (or MAF if it has one) sensor. The TPS tells it how fast you want it to run the engine and the MAP/MAF allows it to calculate the volume of air being ingested so how much fuel is needed to be injected to achieve this. In closed loop it's still doing this but now it also looks at the O2 (Lambda) sensor signal which allows it to "fine tune" the course settings.
So how does it do this? First thing to understand is that the Oxygen sensor (I'm not looking at wide band fuel ratio sensors here - still learning about them in case I ever get my Civic!) is like a wee battery which can produce about one volt of electricity and it does this depending on the amount of unburnt oxygen it senses in the exhaust gasses (it is located upstream of the catalytic converter so is "sniffing" unadulterated exhaust gasses as they leave the cylinders. The sensor needs to be REALLY hot before it will work) A richer mixture will send it towards the 1 volt and a weak mixture sends it towards 0 volts. In practice it fluctuates around once to twice a second between around 0.9 and 0.1 volts. Difficult to pick up on a multimeter, you really need to graph it with a scanner (MES is ideal). So the ECU is trying to achieve an air/fuel ratio of around 14.7% air to fuel (this is by weight so you're talking a lot of air to very little fuel). This ratio uses all the oxygen to burn all the hydrocarbons in the fuel so minimum pollution occurs.
But why does this voltage fluctuate? Well, it's the ECU trying to achieve the 14.7% ratio. Lets say the mix starts to drift slightly rich? the voltage from the O2 sensor will start to rise, the ECU "sees" this so cuts back the fueling by a very small amount on the next injection and it continues to do this miniscule cutting back on each subsequent injection until it starts to "see" a weaker mixture indicated by the voltage falling. If the voltage falls too much the ECU will enrichen the mixture again and so it goes on with the swing being as I said in the above paragraph (around 0.1 to 0.9 volts).
Now we come to fuel trims and this is where I am in my climbing of the ladder of knowledge at this time. I believe these very small fueling adjustments (measured in %ages) are what present as "Short Term Fuel Trim" (STFT). and you can watch them real time to get an idea what's going on with the fueling. There is also "Long Term Fuel Trim" (LTFT) and my understanding of that is that it's the baseline from which the ECU is calculating it's fueling so you would expect to see it at or very close to zero on an engine in good mechanical condition which is working well. However if the STFT is consistently, just for example, tending towards the rich side and the ECU can't pull it back with the very small fuel variation it can achieve undet STFT it will then, after trying for a while, reset it's LTFT to the average richer figure and the STFT settings will then be zeroed on the new LTFT?
So, form a fault finding point of view I think you can say that LTFT should be around Zero on a good engine. If however it's showing, again for instance, around 15, maybe 29% or whatever, you've got a problem. If it's +20% then you're rich, if -20% then you're weak. so you can start looking for maybe leaking injectors, inlet gasket leak etc?
So what do you think everyone? am I deluding myself or have I got the "stick" by the right end?
Last edited:
