Transients

[Tony Bryant]

> Is this race car, or street car, how much cam?.  Lots of 
> things come into
> play.
> I'm addressing a street engine, as opposed to streetable or racing.
> 

It is a race car with a biggish cam, but I trying to develop a general 
understanding of fueling requirements, and the point of this 
exersize is to develop an ECU that will as much as possible make a
hot cam behave less like a dog, when using in street conditions.

Race engines bring out the nasties inherent in all engines, so it
make sense to consider the worst case first.

> | I run across bazillions of different theories, all of which I've
> | decided are crap (feel free to correct me though).
> 
> Everyone is entitled to their opinion, weither you call it 
> crap or not is
> immaterial.

Wrong choice of words, I'll admit, but I'm interested in facts,
just the facts. And I KNOW that some of the opinions ventured to 
me in the past have no basis in fact.
 
> | 2) The intake runners have columns of air that need to be
> | accelerated to the new conditions.
> 
> | Bollocks. That would also make it run rich, and besides the
> | flow always tops between intake cycles anyway.
> 
> Why then is cross sectional area so important to throttle response?.

Because a smaller diameter runner will have less self-EGR effect, due
to the higher velocities. e.g.

e.g.

Assume 100% VE requires 5ms duration.

Engine A:

At idle (assume 15" vac) with a small diameter runner 
	assume 30% VE -> 1.33ms

Engine B:

At idle with a large diameter runner, with more self-EGR effect 
	assume 20% VE -> 1ms

So, when the throttle is snapped open, lets say next cycle you've 
got 50% VE worth of air, but either 40% or 60% of the fuel required for this
mixture.

Engine A may be able to fire this lean mixture, Engine B probably won't.

The same principles applies to changing cams.

> | 3) There is a delay getting the fuel in the port to the cylinder.
> | Bollocks. Its squirted directly on the back of the valve, at low rpm
> | it'll just run in under the influence of gravity, when the valve is
> | opened.
> 
> Gravity ain't fast enough for cylinder filling of puddled fuel.
> 
> | 4) Manifold wall wetting effects
> | Bollocks. This is port injection, the walls don't ever see fuel.
> 
> Look again, where does the fuel go when sprayed against a 
> closed valve?.
> The back of the valve, and splashs the walls, not the entire 
> lenght like a
> carb set up, but still some.

My experience with squirting injectors at things, did not generate any
amount of splatter. It just wets what you point it at. There may be some
fuel temporary tied up wetting the valve, I can't believe the walls 
get involved with this story at all. I'll concede a small amount extra may 
be required for the first post transition cycle due to valve wetting
effects.

> 
> | 5) The vapourised fuel in the port condenses because of the pressure
> | increase.
> | Maybe, but it seems that this would not cause a major 
> change in fuelling,
> | as its hard to see how the fuel in the cylinder will not be burnt.
> 
> If the droplets are large enough it won't..

If the fuel condenses during the first 100% VE cycle, it'll condense 
during the second,third, and fourth etc too, and it'll run like crap
on all WOT cycles. I can't see how this effect would only affect 
transitions.

> 
> | 6) A richer mixture (i.e. more powerful) is needed to accelerate the
> engine.
> | Bollocks. Running a seriously rich mixture at say 
> 1300RPM/0" vac, does not
> | help
> | the problem. Neither does serious amounts of dMap/dt. 
> dTPS/dt helps a
> | little,
> | but not much.
> 
> Then try going very lean, ahh, see, problem gets worse

That because on your first 100% VE cycle after the transition, the 
fuelling will still generally be set-up for the pre-transition 
conditions. Leaner there, gives much leaner post transition.
 
> | ----------------------
> | Having stated what I think its not, here's what I think it is:
> | 1) Fueling delays. The ECU not responding to changes in MAP as
> | quickly as the engine. The engine will respond on a cycle by cycle
> | basis, whereas an ECU and MAP sensor most likely will not.
> 
> If you've changed cams, or got too much volume in the MAP 
> sensor line, then
> the filtering rate may be off.  The ecm can play a chess game between
> ignition cycles.  Time/response not an issue or no high 
> performance engine
> could run one.

It all depends on the compromises you mind making. High performance
engine generally run a (very) high idle. Why? 1. Self-EGR at low rpm
makes the required mixture hard to predict. 2. Take it out of the 
low rpm stumble "danger" area.

I'm trying to push the boundaries of what possible with a high-performance
engine. e.g. Silky smooth, low rpm idle, with instant throttle response
everywhere, all with reasonable emissions. Difficult - of course - thats 
why I'm trying to develop my understanding of the phenomenens involved.

> | 2) The fact that the fuel is injected before the IVO 
> (intake valve open)
> | event, but the engine "samples" MAP at IVO. Therefore it may not be
> | predictable what the fuel will need to be, when the fuel is 
> squirted in,
> | because the MAP can change significantly then and IVO.
> 
> For vaporization, (ie street) you want the fuel to hit the 
> back of a closed
> valve,
> also, the shearing action against the fuel as the valve opens.

Obviously there's a trade-off here between vapourisation, and
accurate fuelling, by being able to delay your fuelling calculation
until the last possible moment.

I'm currently modifying my ECU, to allow calculation and updating of
injector durations at least every cycle (at <2500RPM). I will post to 
let you know if this helps.