Forum

[Rattlebattle]

Does anyone have the definitive answer to how the EFI system times the injection of fuel into the inlet? As I understand it the ignition is wasted spark, induced by the 24th coil on the alternator passing the crank position sensor, presumably using the Hall Effect. But what about the fuel injection? Is this wasted too? If not, what is it that tells the ECU what stroke the engine is on? I have read the workshop manual and, of the sensors in the system the only possible candidate is the MAP sensor, which I suppose could indicate the stroke because of fluctuations in the air flow. FWIW BMW oilheads had two sensors close together. Using the Hall Effect, the rotor passing the first one caused the fuel to be injected, followed shortly afterwards by the second sensor firing the spark plug. On these the injection as well as the ignition was wasted. How does the RE EFI system work - does anyone know? I'm just curious really.

[jefrs]

It do believe it is indeed 'wasted spark', one spark per revolution when it only needs one every two. I gather the alternator timing pulse is longer than the others; I do not think Hall, the book says "inductive pulse generator" sending an AC signal to the ECU. The crank sensor most likely tells it when to inject but it cannot inject against a closed inlet valve and would waste fuel, it may use the MPS* to tell the ECU when not to spark or inject (no inlet flow, don't inject or spark). The advance-retard is in the ECU's ignition mapping table, I cannot see this but can see how the Power Commander V modifies that table. Nor can I deduce whether it advances on acceleration like the sprung weights on a distributor do.



The Workshop Manual lists and describes in poor detail - ECU, Alternator, Throttle Body Assembly (Throttle Position Sensor and *Manifold Pressure Sensor), Temperature Sensor, Fuel Injector Assy, Crank Position Sensor, Fuel Pump Module, Rollover Sensor, Lambda HEGO O2 Sensor*, Malfunction Indicator Lamp (MIL), Low Fuel Sensor.



The Crank Position Sensor scans 23 short and one long pulse from the alternator, the long pulse is 5° before top dead centre. It is also used to determine engine speed. Although we might advance the spark with a PCV, I see no way it can produce a spark before the crank sensor pulses. But it may use all 24 pulses in some way, it is a Keihin system.



The Lambda Sensor attempts to provide a stoichiometric mix of 14.7:1 but this takes 90 seconds to warm up before it does anything.

[jefrs]

If it injected once every revolution it would inject one against the closed inlet and fuel would sit in the manifold and be swept into the cylinder on the next induction. That would not work at all well! It must do something else to only inject once every two revolutions, and if it can do that then it does not need to do a wasted spark.

[Rattlebattle]

Thanks for that. As I see it the fact that the crankshaft position sensor sends a signal to the ECU (and you are right, it isn't likely to be Hall Effect) every two strokes doesn't per se mean it is wasted spark because unlike a magneto or flywheel generator it only triggers the low tension side and it wouldn't take much to programme a high tension signal for every other low tension pulse, if you see what I mean. On another forum the consensus is that it is wasted spark (easy enough to check I suppose) but in that case what on earth triggers the fuel injection to "fire" before alternate sparks? I don't really see how there can be two injections of fuel per 4 stroke cycle, but in this case there must be some means by which the ECU knows which stroke the engine is on. Like I said of the sensors you list the only candidate I can see is the MAP sensor. Perhaps I'm missing the blindingly obvious but it's got me baffled....

[Nobbi1977]

The injector will fire once every revolution and the fuel will cool the inlet and start to atomise in the heat so it is better when drawn in to the cylinder.

I have made and programmed ECUs and read up a lot on timed injection. The consensus seems to be there is a little to be gained onemissions and economy aiming to fire directly in to the valve as it opens but no power is to be gained. Rover actually timed it to inject on the closed valve to cool it.

If you think this is all happening at up to 60 times a second and the intake is opening 30 times a second then it is all just one constant flow rather than individual events.

[Rattlebattle]

Thanks. That is how I understood the BMW system to work. In effect, wasted spark and "wasted" injection. In the absence of a camshaft position sensor or some other sensor that would tell the ECU what stroke the engine is actually on I couldn't really see how else it could work. Thankyou for confirming that.

[Thack]

I wouldn't say it's confirmed, yet. The only way to know for sure is to instrument the system and get that data.



Both injection regimes (timed injection and "wasted spark" injection) are valid and widely used, and it's quite common to use both in the same engine: timed at lower rpm and "wasted" at higher rpm. One reason some engines use that regime is when the injector can't inject enough fuel in the time the inlet valve is open, such as happens at very high rpm and large throttle openings. In those cases the injector is fired again on the power stroke. The fuel builds up in the inlet tract behind the valve, and has extra time to vaporise, so it's actually quite a good thing.



The MAP does provide the information the ECU needs to decide which stroke the engine is on, although whether it makes use of it only Keihin will know (and us too, once it's been fully instrumented).



So we shouldn't consider this question answered until we've actually been in there with a 'scope and looked. My "black box" data logger is almost in a position to do that, and after a little more development I'll be able to say for sure just what happens.

[Thack]

I wouldn't say it's confirmed, yet. The only way to know for sure is to instrument the system and get that data.



Both injection regimes (timed injection and "wasted spark" injection) are valid and widely used, and it's quite common to use both in the same engine: timed at lower rpm and "wasted" at higher rpm. One reason some engines use that regime is when the injector can't inject enough fuel in the time the inlet valve is open, such as happens at very high rpm and large throttle openings. In those cases the injector is fired again on the power stroke. The fuel builds up in the inlet tract behind the valve, and has extra time to vaporise, so it's actually quite a good thing.



The MAP does provide the information the ECU needs to decide which stroke the engine is on, although whether it makes use of it only Keihin will know (and us too, once it's been fully instrumented).



So we shouldn't consider this question answered until we've actually been in there with a 'scope and looked. My "black box" data logger is almost in a position to do that, and after a little more development I'll be able to say for sure just what happens.

[Thack]

That's odd..... Sorry for the double post.

[Rattlebattle]

Thanks Thack; I was hoping you'd chip in. I look forward with interest to what you find. My guess (just that, obviously) is that it will be "wasted injection" (for want of whatever the correct terminology). All academic really; I'm just curious, though I suppose it would help understand tuning the ECU better if we knew exactly how it worked.