[QUOTE=aerodave]
But for a drip from a faucet? Probably not. I’d say a slow drip tracing the underside a convex curve is a adhesion- or surface tension-driven phenomenon, whereas the Coanda effect is a pressure-driven phenomenon. The Wikipedia article uses a stream of water as an illustrative example, but I say that’s only valid of it’s enough of a stream to be considered a slow jet. If it’s a trickle–or especially if it’s in discrete drops–the curving of the water will no longer be Coanda driven, but rather dominated by adhesion. It just happens they produce qualitatively similar results.
[/QUOTE]
This seems intuitively true, but in fact it is not an adhesive effect, or at least not dominated by adhesion. CalMeacham mentioned Jearl Walker’s The Flying Circus of Physics (an excellent read for anyone who wants more than very simplified explanations for physical behavior) and he has a very similar example:
4.118: Pouring water from a can[indent]When I pour my beer why does it insist on running down the side of the can instead of falling straight down from the lip? What determines how far it adheres to the can? How fast must I pour the beer to prevent such “sticking”?
Your first impulse will most likely be to attribute the phenomenon to surface tension or adherences of the liquid to the container. However, neither is responsible for the spilt beer. What is, then?[/indent]
Walker’s answer to the question is thus:The turn of the stream around the edge of the can is stable because of the pressure difference across the width of the stream. An ideal incompressible fluid in a circular path has greater velocities at smaller radii. Thus, but the Bernoulli principle, there is less fluid pressure at smaller radii. Here the atmospheric presure outside the stream is greater than the fluid pressure near the edge and therefore holds the stream to the edge. At some point on the side of the can the stream detaches because it is unstable to small perturbations.
Now, I think that there is certainly some contribution by adhesive mechanisms, at least insofar as strong adhesion will negate the pertubative effects that cause detachment of the stream, but “adhesion” isn’t a discrete mechanicsm but rather a catch-all observation for a host of tribological phenomena involving atomic or molecular scale interactions between the glass and the fluid. Similarly, surface tension of the fluid no doubt has some significant effect on how much pressure effects can dominate before forcing the fluid to seperate into multiple streams.
But the primary effect is a pressure differential, although not the one described by the Coandă effect (which is hypothetically caused by a lateral pressure vortex trying to form between the surface and a moving fluid.) The effects are qualitatively different in terms of the direction of the effect; the former is due to internal pressure differences in a narrow stream at low Reynolds numbers (R[sub]e[/sub]<<1, where viscous forces dominate, also known as Stokes flow), whereas the Coandă effect can occur evenly over a broad surface like a “flying disk” at high Reynolds numbers without any lateral pressure differences. The Coandă effect is poorly described by the Navier-Stokes formulation, whereas the Stokes flow is a linear simplification of the generalized Navier-Stokes equations.
As to why it defies gravity, this is simple; despite its pervasiveness, gravity is a very weak force; so weak, in fact, that the electrostatic molecular forces keep you from collapsing into a formless pool of protoplasm. You can defeat gravity by simply rubbing a balloon in your hair and allowing static attaction to stick it to a wall. The only reason gravity is even held in such high esteem is because you can readily observe its effects at normal scales.
Stranger