Tapered intakes, port velocities etc.

[Todd Knighton]

Beaubien, Matthew wrote:

> The
> way I saw it was that you have the smallest pressure differential when
> the valve is first opening, and thus, the least potential to flow a/f.

	Unless your exhaust scavenging and tuning is working well, you may well
have an appreciable pressure differential there.

> What I didn't take into consideration was the flow through the valve
> while it was closing, when the interia of the a/f mixture is probably
> close to its maximum. It sounds like a large part of the cylinder
> filling may happen when the valve is closing/near closed.

	With camshafts too advanced, maybe, but typically your trying to get
the valve to be closing just about the same time as the flow is stopping
to get the maximum permissible air/fuel in the chamber without
reversion.

> Todd, you mentioned the velocities seemed high compared to what you work
> with. You mainly work on turbo flat 6's, correct?

	Yes.  More twin-turbo's than single lately.

> The article seemed to
> be more bike orientated (obviously) which would mean multivalve engines
> that spin about twice as fast as your used to. I would assume those
> factors would have some bearing in the velocities.

	He was making many quotes from Cosworth, which was auto related, though
also very high rpm stuff.  It should still be fairly relative though,
they had 50mm strokes on a lot of the Cosworth stuff, we have 75mm, so
at a proportionately lower rpm we should be obtaining similar
velocities.  They were also typically running 3.0 v-8's and we're
running 3.5 to 3.6l flat 6's, so again, versus valve sizes and such, we
have a much higher displacent versus valve sizes and port velocities
should be pretty good.

> However, the pressure
> differential you're dealing with is greater (ie. turbo boost) which
> should speed things up.

	Not really.  Turbo stuff generates just about as much back pressure as
intake pressure, typically the most we'll see positive is about .5bar or
7.5 psi.  So the whole thing is under pressure, unlike supercharging.

> When talking about velocities, are they measured
> or calculated?

	We've calculated them, big-budget companies may have the tools to
measure them.
	Figure a 3.5L engine at 6500 rpm's pumps out about 398cc normally
aspirated at 100%VE, for a 6 cylinder that's 66.3 cfm.  Our ports are
40.5mm or about 1.6inches or about 2 sq. inches, or .0138665 sq ft., or
about 4,781.3 ft/min  divided by 60 equals 79.6 ft/sec, almost 25% of
the value they're stating.  Now this is steady state, not peak, but at
high rpm's I'd assume the stuff is pretty steady state in flow.  We'd
need about 26,000 rpm to see the port velocities he's talking about.

> 
> Interesting stuff. I just hope we don't have another "blood-letting"
> with a DIY-EFD (engine fluid dynamics) group
> forming ;-). Maybe I'm wrong and there would be enough interest/support
> for it.

	If oil pumps can be here, why not engine design fundamentals, after
all, ports, manifolds, and throttle bodies are all a good part of the
fuel injection system.

Todd Knighton
Protomotive Engineering