|
|
Ernest Christley wrote:
Bill Dube wrote:
The difference in pressure across the injectors is what matters, not the absolute fuel pressure. If you hold the difference in pressure constant, the injectors deliver the same amount of fuel per pulse, regardless of manifold pressure changes. If you run a fuel pressure that is not referenced to the manifold, you will have less and less difference in pressure as the boost comes on. This will make your situation worse, not better.
Agreed. But with a fuel pressure over 50 PSI, it will take a very significant turbo boost to cut off the flow altogether. The two question that then remain are, "Do the injectors have the time to dribble in enough fuel at full boost?" and (on the other end) "Can we turn the firehose off quick enough not to flood the engine?"
It isn't that it will cut off. It is that the manifold pressure directly modulates the fuel flow if you don't reference the fuel pressure to the manifold pressure. At high manifold pressure, you would lean out, not cut off, the fuel flow.
The stability issue is a subtle one, but important. Let's pick high manifold pressure (boost) as an example. You open the throttle. Pressure instantly goes up. The difference in pressure across the injectors goes down. The fuel per pulse goes down. The fuel flow instantly goes down before the injection computer can compensate. The engine goes lean. The engine slows down a bit. The manifold pressure is starting to go up a bit more because the engine slowed down a bit. The difference in pressure across the injectors now goes down a bit more. They instantaneously deliver less fuel. The engine goes even leaner. The computer is now catching up. It calls for more fuel. The injectors open for longer. The engine starts to speed up. The manifold pressure goes down. The difference in pressure across the injectors increases. They instantaneously deliver more fuel. The engine speed up more. The manifold pressure drops more. The injectors deliver even more fuel per pulse, instantaneously. The computer now catches up and calls for less fuel. ............
Bill Dube'
This reasoning hinges on the assumption that the mechanical manifold to fuel pressure reference system is instantaneous and that the computer will be lagging behind. I'm not sure I can swallow that pill unless there is a lot of unnecessary hysteresis built in front of the computers pressure sensor (perhaps from using an overly large hose). I don't think I've ever heard of a problem tracked back to the EC2 not being able to keep up.
You don't compute the length of every pulse to every injector every time. Since the engine RPM changes slowly and the manifold pressure changes at modest speeds, the EFI controller ramps the injector pulse width up and down slowly (a bit faster than the engine can change speed.) The change in fuel delivery rate is instantaneous with a change in manifold pressure if you don't reference the fuel pressure to the manifold, however. If you do reference to the manifold, the change in fuel pressure to compensate for that manifold pressure change is very nearly instantaneous (from the engine's perspective.)
For those with an EE controls background, it is easier to explain. The problem with no manifold pressure reference on the fuel is that it results in a pole in the right-hand plane. While it is possible, in theory, to compensate for this in the control feedback, it is difficult and can easily get out of tune.
For those without an EE background, here is the simple explanation. An absolute fuel pressure control causes "positive feedback" in the system. An increase in in manifold pressure with constant fuel delivery naturally causes a lean mixture. If you don't reference the fuel pressure to the manifold, the fuel delivery is not constant, but actually reduces with increasing manifold pressure. Thus, as you increase the manifold pressure, the mixture leans out much more than it would otherwise.
You program the EFI controller to sense the change in manifold pressure and use this "rate" signal to give an extra shot of fuel when the pressure changes rapidly. (This is why there is an accelerator pump on a carburettor.) This, "sort of" works if the pilot moves the throttle at a reasonable speed. On cars, they sense the throttle position so they don't have to wait for the manifold pressure to change before they put in the extra shot of fuel. (They do this because the throttle position changes more rapidly in cars.) If you compound the effect by referencing the fuel pressure to atmospheric, rather than manifold pressure, the EFI has a very hard time holding the system stable. It will tend to oscillate as the mixture chases the manifold pressure which chases the engine rpm.
To top it all off, if you reference the fuel pressure to atmospheric, the fuel per pulse will vary with manifold pressure AND altitude. The EFI only senses manifold pressure (because that is what matters to the engine) and would not compensate for altitude changes. The system would go way too lean at high altitude and way too rich at low altitude.
In a nutshell, the EFI is set up to have the fuel pressure referenced to the manifold pressure. If you change that, you will have to do a lot of additional R&D to the EFI system to make it work properly.
Bill Dube'
|
|