Discussion
OK, so nowadays we have adaptive ECUs which can adjust themselves to cope with different fuel, engine wear, and stuff. But in what terms do they define optimum running? Is it measured in emissions terms, or BMEP? And if the latter, how do they measure engine output - average deflection of engine mounts?
The TCCS toyota ecus are adaptive.
They compensate for different components being fitted, engine wear, different fuels,altitudes etc etc.
As an example, if the ecu is disconnected, it "forgets" all the previously stored settings and has a poor idle for quite some time, lower power etc until it has been driven for some distance.
Only after its been driven does it re learn the engine parameters.
They compensate for different components being fitted, engine wear, different fuels,altitudes etc etc.
As an example, if the ecu is disconnected, it "forgets" all the previously stored settings and has a poor idle for quite some time, lower power etc until it has been driven for some distance.
Only after its been driven does it re learn the engine parameters.
Sure; advancing until the knock sensor signalled problems is I believe the first adaptive function that was introduced. What I'm interested in is: in the case of things like, but not limited to, the Toyota TCCS ECU, are they optimising purely on monitoring combustion in some way (knock sensors, lambda sensors, and related devices) or are they actually measuring the mechanical output of the engine, and if they are measuring mechanical output what kind of sensors are they using to do it?
I reckon it's lambda's they are monitoring but on some ecus's that go open loop you get some extra fuel when you put your foot down. So the they are reading the throttle pot and vacuum. This is why I remarked about the sport car or suv. The manufacturer is thinking about what the customer wants and emmissions.
My saab 9000 was a bit flat when I had had the battery off for a few days. It didn't seem to want any boost. I've fixed it now though
Gave it a darned good thrashing.
My saab 9000 was a bit flat when I had had the battery off for a few days. It didn't seem to want any boost. I've fixed it now though
Gave it a darned good thrashing.Theyre monitoring boost pressures, fuel pressure, throttle position, O2 contents, rpm vs road speed, previous iscv settings required to maintain the current idle speed vs any new ones necessary, egt, oil/air/water temps.
The tccs system has a zillion and one in built safety modes to limit boost pressures also, so if you disconnect the map sensor to raise the boost level (its not used in Map mode btw, its used purely for overboost cutout or in toyota's case fuel cut) youll get a rise in boost, but the remaining sensors will only let it go on for so long until the advance gets cut back, and thats before any knock even occurs.
All in all, adaptive systems are extremely clever pieces of hardware/software combos, which is why i want an aftermarket version of the factory unit....
The tccs system has a zillion and one in built safety modes to limit boost pressures also, so if you disconnect the map sensor to raise the boost level (its not used in Map mode btw, its used purely for overboost cutout or in toyota's case fuel cut) youll get a rise in boost, but the remaining sensors will only let it go on for so long until the advance gets cut back, and thats before any knock even occurs.
All in all, adaptive systems are extremely clever pieces of hardware/software combos, which is why i want an aftermarket version of the factory unit....
OK, right. So they're basically monitoring combustion and all the factors that may affect it, but don't actually monitor the mechanical output of the engine.
What's behind my question is the desire to make an adaptive ignition advance controller for my MZ, which as standard just has fixed ignition timing. The four-stroke methods of monitoring combustion are unlikely to work properly on a two-stroke, and even a knock sensor wouldn't be much use, as tweaking the timing by hand reveals that it loses power well before it runs into knock. So I was hoping that there existed a car ECU that measures actual output, which would give me a good starting point to research the sensors it used to do this. But apparently there isn't. Oh well. Bugger. Thanks for your answers, anyway!
What's behind my question is the desire to make an adaptive ignition advance controller for my MZ, which as standard just has fixed ignition timing. The four-stroke methods of monitoring combustion are unlikely to work properly on a two-stroke, and even a knock sensor wouldn't be much use, as tweaking the timing by hand reveals that it loses power well before it runs into knock. So I was hoping that there existed a car ECU that measures actual output, which would give me a good starting point to research the sensors it used to do this. But apparently there isn't. Oh well. Bugger. Thanks for your answers, anyway!
Pigeon said:
OK, right. So they're basically monitoring combustion and all the factors that may affect it, but don't actually monitor the mechanical output of the engine.
What's behind my question is the desire to make an adaptive ignition advance controller for my MZ, which as standard just has fixed ignition timing. The four-stroke methods of monitoring combustion are unlikely to work properly on a two-stroke, and even a knock sensor wouldn't be much use, as tweaking the timing by hand reveals that it loses power well before it runs into knock. So I was hoping that there existed a car ECU that measures actual output, which would give me a good starting point to research the sensors it used to do this. But apparently there isn't. Oh well. Bugger. Thanks for your answers, anyway!
Hey Pidgeon boy!
"Theyre monitoring boost pressures, fuel pressure, throttle position, O2 contents, rpm vs road speed, previous iscv settings required to maintain the current idle speed vs any new ones necessary, egt, oil/air/water temps."
This is essentially correct, there is often barometric compensation too. However more critical is that they're measuring cataysts brick temperatures and adapting fueling to keep these cool in extreme conditions. Air flow is measured-which is similar to Volumetric efficiency, strictly speaking an engine can be mapped without this, using look up tables, but the strategy is much more flexible and adaptive if the air flow is measured.
I know of no mainstream production ECUs that monitor actual cylinder pressures to get IMEP although THAT would be ideal.
A patented system on Jaguars, is that the INSTANTANEOUS crank speed is measured using flywheel speed sensor. This is helpful at idle, for instance, as if this is held in reference to which cylinder has just fired the ECU can deduce which cylinder explosion is proucing greater output (bascially equating instantaneous crank speed to power each cylinder produces) and therefore reduce power of that cylinder using ignition. Adaptive ignition to control idle speed fluctuations is a much quicker response then using air flow variations with the throttle or an idle speed control valve- so represents a great short term control strategy- while the air flow control using the throttle is still in place as a long term strategy. The biggest issue with using the throttle ALONE (in NON port throttled cars) in idle speed control is the big throttled volume (from throttle to intake valve)- so it takes quite a while for the engine to realise a target change.
The other new feature which is coming in, is monitoring idle speed flare on start up to try to ascertain what kind of fuel is being used in the engine and thus altering the characteristics of the ECU. It is a known fact that Low volatility fuels cause a much lower flare on start up as they're producing less torque! The challenge is for the ECU to successfully recognise start up flare fluctuations from this fuel effect rather then one from something else like leaky injectors or the like. Now you start to see the challenge that a modern EMS engineer has to cope with!
Pigeon said:
OK, right. So they're basically monitoring combustion and all the factors that may affect it, but don't actually monitor the mechanical output of the engine.
What's behind my question is the desire to make an adaptive ignition advance controller for my MZ, which as standard just has fixed ignition timing. The four-stroke methods of monitoring combustion are unlikely to work properly on a two-stroke, and even a knock sensor wouldn't be much use, as tweaking the timing by hand reveals that it loses power well before it runs into knock. So I was hoping that there existed a car ECU that measures actual output, which would give me a good starting point to research the sensors it used to do this. But apparently there isn't. Oh well. Bugger. Thanks for your answers, anyway!
I'll be honest, I'm NOT as intimately au fait with the 2 stroke process as I am with 4 strokes. But what you're describing is that you're carrying out an "ignition loop" when you're tweaking the ignition and monitoring power. What you're achieving is MBT ignition- or the optimum ignition for best torque before you run into knock. MBT ignition is basically when the gradient of the curve of ignition versus torque (dy/dx)=0. Four strokes do this at part load. If you loose power BEFORE running into knock, on a 4 stroke, you could do a couple of things to make the engine run straight into knock just after it achieved full power:
You could either ramp up the compression ratio or increase the air flow/Volumetric efficiency/breathing of the engine so it is effectively running at highler load/cylinder pressures. I'm not sure if this instantaneous crank speed feature can be used to monitor outout for what you need. With a flexible enough system and enough time I don't see why not.
Hey, Marquis, I was hoping you'd find this thread! Thanks.
Yes, I am thinking of making a very simple ignition loop system. The information I have suggests that maximum power occurring significantly before the knock limit is the expected behaviour for two-strokes. I don't want to up the compression ratio too much as it makes it more peaky. Experiments with altering the volumetric efficiency are taking place but slowly
The idea of measuring crank acceleration during firing is very neat - especially since the revs can change so rapidly that I would probably be having to monitor it anyway to get the timing right under hard acceleration in low gears. Using the same input in different ways. Nice one.
The reason I'm thinking of doing it adaptively is mainly that if I'm going to replace the standard fixed-advance system, the easiest way to set it up is to get it to set itself up. I don't have access to a dyno; riding it, stopping it, tweaking it, riding it etc. is a pretty hopeless method; making it so that I can tweak it as I'm riding is not much better and rather dangerous; so making it tweak itself seems the best option. The hardware is all the same, it only requires different software.
Yes, I am thinking of making a very simple ignition loop system. The information I have suggests that maximum power occurring significantly before the knock limit is the expected behaviour for two-strokes. I don't want to up the compression ratio too much as it makes it more peaky. Experiments with altering the volumetric efficiency are taking place but slowly
The idea of measuring crank acceleration during firing is very neat - especially since the revs can change so rapidly that I would probably be having to monitor it anyway to get the timing right under hard acceleration in low gears. Using the same input in different ways. Nice one.
The reason I'm thinking of doing it adaptively is mainly that if I'm going to replace the standard fixed-advance system, the easiest way to set it up is to get it to set itself up. I don't have access to a dyno; riding it, stopping it, tweaking it, riding it etc. is a pretty hopeless method; making it so that I can tweak it as I'm riding is not much better and rather dangerous; so making it tweak itself seems the best option. The hardware is all the same, it only requires different software.
Gassing Station | Engines & Drivetrain | Top of Page | What's New | My Stuff