|
Howdy,
You made me dig out my books. ;-)
First, I need to correct the major mistake I made. I said, "... a fuel injector will pull 5 Amps average...". That should have read "... a bank of fuel injectors will pull 5 Amps average...".
Second, I am self-taught. If I have erroneous ideas about the way a thing works, I have had no one to correct me. Feel free to correct anything I say; it is the only way I have to learn anything new or correct anything I misunderstand.
Your Mitsu' uses peak-hold injectors. It is a very elegant design that solves some of the problems associated with saturation type injectors. It requires a sophisticated control circuit capable of reversing current rapidly. The Mitsubishi design is also a disc type injector; it has no pintle. Your Japanese injectors have several decades and many millions of miles worth of data behind them to improve the design. Finally, it is a SFI system; a whole different animal compared to the LH series.
Bosch used saturation injectors in the LH systems. To simplify matters a bit, I'll stick to the LH system; we are discussing saturation injectors.
As I mentioned earlier, the injector is basically a coil of wire that pulls on a ferrite slug. That is where the simplicity stops. The injector is one of the most highly engineered pieces ever designed for use on production automobiles. The injector in the LH series was designed over twenty-five years ago. There have been refinements over the years but it is basically the same injector throughout the series production life.
The construction of the injector is "self-limiting", meaning the electrical characteristics are designed to allow maximum inrush current followed by a saturation condition that bucks (self-limits) hold current to about 60% of the initial inrush amperage. This design was adopted by Bosch with the advent of the LH series. It eliminates the need for current limiting resistors and the attendent need to supply more current to overcome those resistors until the magnetic field is fully built. Faster reaction time. Less heat. All good.
The injector ON time is limited at the low end by the nature of coils. It takes a finite amount of time for the ferrite slug to "pull off" the seat. It only has to move about .006 inches to achieve full flow. But that small distance requires somewhere near a full millisecond to achieve. Fuel begins to flow at about .25 thousandths. But coils react in a logarithmic fashion. The first 70% of opening time is dedicated to building a sufficient magnetic field to move the slug at all. The coil has to overcome the inertia of the slug, the "keeper" spring behind the slug, and the hydraulic pressure of the fuel. That is why the minimum pulse is roughly 2ms.
The pulse duration is governed by the load, RPM, the environment, and O2 feedback. Compensation for the physical limits of the injector is designed into the computational circuitry. It is what it is and we must assume the ECU knows best what the injector ON time should be. Now, it can be wrong; but invariably the problem is related to bad data received by the ECU, not poor design.
Back to the injector. Closing time is likewise limited by the mass of the slug. It takes about 3 milliseconds for the magnetic field to collapse completely. During that time the slug will accelerate toward the seat until it closes the path for the fuel. The spring and fuel pressure will now hold it closed until the next injection event.
At the end of the pulse period (anywhere from 2-15ms) the ECU removes ground and the injector closes. The collapsing magnetic field causes a rather large inductive kick of reverse polarity at the end of the event. If you use an oscilloscope to monitor injector pulses you can see all this going on.
That is my understanding of the function of a saturation injector.
The original inquiry was regarding injector firing sequence as it is timed to valve events. I've gotten a bit off track delving into fuel injector operation but I thought it was necessary to explain why the Bosch engineers designed the system the way they did.
The Bosch batch fired system will fire twice for each 720° of crankshaft rotation (one Otto cycle). The LH2x series fire all four injectors at once. Two of the injectors will be "looking at" open intake ports for an injection event. The other two intake valves will be closed. 360° latter the ECU fires the injectors again. Now the two that were closed are open and the two that were open are closed. So what happens to the fuel squirted against the closed valve?
The valve gets cooled. That is a good thing. The fuel gets vaporized, also good. Not so good is that as the valve closed (slammed) shut, it created a pressure wave traveling backwards toward the throttle plate. That pressure wave is carrying with it some of the fuel charge we worked so hard to put right at the back of the intake valve. This is compensated by a fine adjustment carried in the map of the ECU. Each engine is different so this value was determined empirically during development; along with a dozen other factors, I'm sure.
Now consider a situation of high RPM, high load. Fuel demands are going up. But the intake valve is only going to be off the seat for a comparatively short time; less than 3ms at 6000 RPM (dependent upon the cam). The ECU knows that this load is going to require an injector pulse of (for argument) 8ms. But at 6000RPM the entire 4-stroke cycle is only 20ms long. There is barely enough time to fire the injectors twice. What if RPM increases to a point where there is insufficient time to fire the injectors twice? Will the system revert to 100% duty cycle in an attempt to keep up with demand? No! It will remove one firing entirely, reducing the fuel by half, slowing the engine speed until it falls to a trigger point some where below the redline and resumes normal operation. It is a "soft" rev limiter.
In a cold intake tract (winter morning starts) fuel will tend to drop out of the charge (puddle) creating a lean condition. That is why there is a cold start injector; to super-saturate the cold charge. Inefficient? Yes. But it was a workable solution to an engineering problem that had to meet production realities. Back then computing power was about 1/10000 of what is available to the engineers today. (Look at MicroSquirt; worlds better than the LH at just about everything in a box about 1/5 the size.) They solved the problem with relatively inexpensive hardware. Had they decided to attack the problem with computing power, there would have been yet another control box added to the system. Or, at least, another board and larger ECU than the one they settled on.
There is more to know about fuel injection. Boatloads more. But I think that answers the original inquiry.
--
Mr. Shannon DeWolfe -- (I've taken to using Mr. because my name tends to mislead folks on the WWW. I am a 51 year old fat man ;-) -- KD5QBL
|
