Given the hotly debated "engine option" thread, perhaps its time to throw around some ideas about a more reliable engine controller concept.
First of all I will assume for the sake of this exercise that the engine uses port injectors and that they are fired sequentially. It is also assumed that the operator will use unleaded gasoline, which is much cheaper than leaded.
I would propose an analog (partially mechanical) solution using something akin to a distributor to handle ignition timing, timing advance based on RPM, and vacuum advance (advancing for light loads). This would not be a device to manage the HT side of the ignition, for that one has coil on plug on most modern engines, thus the output would just have to drive the circuit buillt into the coil pack.
In addition, out of phase to the ignition would be timing signals for start of injection at each cyinder. End of injection will be managed by a micro controller on the basis of a speed/density look up table. The fuel value would take into account the pressure and temperature in the intake manifold (giving density) and the engine speed. Engine speed can be determined by the frequency of the pulses on the ignition trigger in the distributor type device.
The coolant temperature (or CHT) in the case of an air cooled engine, together with engine RPM determine enrichment factors (additive to the basic fuel pulse length) to handle cold starting and ignition would be fully retarded based on minimum engine speed and partially open throttle.
The thought overall is to make the system uncritical in terms of processing speed with none of the critical timing being dependent on the microprocessor except to manage the length of the fuel pulse, instead these important signals come from an analog/mechanical device. The signals themselves from either hall type or optical type (think rotary encoder) devices. Both of these options can be very reliable and can be customized. Of course there will need to be an old fashioned centrifugal type of mechanical advance mechanism, along with a vacuum diaphragm/spring type advance system. No points or condenser..... The advance type mechanisms have proven themselves over decades as very reliable. If both advance mechanisms were to fail, if one had the means to lock the timing, one could then adjust the static timing to a median value that would allow the engine to run at a sub optimal point but still letting the engine run and not being stranded.
As inputs, we need the following:
Intake Manifold Pressure x2
Intake Manifold temperature x2
Coolant Temperature x2
As part of the cockpit controls there will be 3 position switches for the 3 input variables, in other words A, B, both
For the distributor type device we could have 2x encoders for each group of signals (Ignition timing and Injector timing).
Thus adding to the cockpit controls for Ignition and Injection there would also be 2 more switches for A, B, Both for ignition and injection timing.
Given that most engines no longer have distributors, this device would have to be integrated into the camshaft drive, which will require a modification to the engine. Not really desirable, but considered a worthwhile compromise. Ideally this would be a 1:1 coupling with the camshaft, but an alternative is to drive it externally via a 1:2 drive ratio toothed belt from the crankshaft. As a device it is barely going to apply any kind of load to the engine, but one has to consider the loss of drive to the distributor as a single point of failure that while highly unlikely, would kill the entire system.
The other potential system killer is loss of electrical power, so ideally we would have a means to power from the system 12V bus with an alternate isolated LI battery, if for some reason one was to lose the main engine electrical power (alternator failure)..
I would need to sketch out this idea and break things out into the seperate modules. I am not an electrical engineer, nor a fantastic code writer so there are opportunities for collaboration with like minded individuals with appropriate skills.
Thoughts ?
First of all I will assume for the sake of this exercise that the engine uses port injectors and that they are fired sequentially. It is also assumed that the operator will use unleaded gasoline, which is much cheaper than leaded.
I would propose an analog (partially mechanical) solution using something akin to a distributor to handle ignition timing, timing advance based on RPM, and vacuum advance (advancing for light loads). This would not be a device to manage the HT side of the ignition, for that one has coil on plug on most modern engines, thus the output would just have to drive the circuit buillt into the coil pack.
In addition, out of phase to the ignition would be timing signals for start of injection at each cyinder. End of injection will be managed by a micro controller on the basis of a speed/density look up table. The fuel value would take into account the pressure and temperature in the intake manifold (giving density) and the engine speed. Engine speed can be determined by the frequency of the pulses on the ignition trigger in the distributor type device.
The coolant temperature (or CHT) in the case of an air cooled engine, together with engine RPM determine enrichment factors (additive to the basic fuel pulse length) to handle cold starting and ignition would be fully retarded based on minimum engine speed and partially open throttle.
The thought overall is to make the system uncritical in terms of processing speed with none of the critical timing being dependent on the microprocessor except to manage the length of the fuel pulse, instead these important signals come from an analog/mechanical device. The signals themselves from either hall type or optical type (think rotary encoder) devices. Both of these options can be very reliable and can be customized. Of course there will need to be an old fashioned centrifugal type of mechanical advance mechanism, along with a vacuum diaphragm/spring type advance system. No points or condenser..... The advance type mechanisms have proven themselves over decades as very reliable. If both advance mechanisms were to fail, if one had the means to lock the timing, one could then adjust the static timing to a median value that would allow the engine to run at a sub optimal point but still letting the engine run and not being stranded.
As inputs, we need the following:
Intake Manifold Pressure x2
Intake Manifold temperature x2
Coolant Temperature x2
As part of the cockpit controls there will be 3 position switches for the 3 input variables, in other words A, B, both
For the distributor type device we could have 2x encoders for each group of signals (Ignition timing and Injector timing).
Thus adding to the cockpit controls for Ignition and Injection there would also be 2 more switches for A, B, Both for ignition and injection timing.
Given that most engines no longer have distributors, this device would have to be integrated into the camshaft drive, which will require a modification to the engine. Not really desirable, but considered a worthwhile compromise. Ideally this would be a 1:1 coupling with the camshaft, but an alternative is to drive it externally via a 1:2 drive ratio toothed belt from the crankshaft. As a device it is barely going to apply any kind of load to the engine, but one has to consider the loss of drive to the distributor as a single point of failure that while highly unlikely, would kill the entire system.
The other potential system killer is loss of electrical power, so ideally we would have a means to power from the system 12V bus with an alternate isolated LI battery, if for some reason one was to lose the main engine electrical power (alternator failure)..
I would need to sketch out this idea and break things out into the seperate modules. I am not an electrical engineer, nor a fantastic code writer so there are opportunities for collaboration with like minded individuals with appropriate skills.
Thoughts ?
