FROM: rwphillipsidaho (Bob Phillips)
SUBJECT: Re: Digest Number 907
[ Attachment content not displayed ]
FROM: kwbradbury (Keith Bradbury)
SUBJECT: Mouthpiece air flow
--- In MouthpieceWork@yahoogroups.com, Bob Phillips 
<rwphillipsidaho@g...> wrote:
>
> If we get turbulence at the entry to the mpc from the high air 
speed in the
> "jet" between the reed and the mpc tip, what will dissipate that 
turbulence
> in the next few centimeters of the mouthpiece passage?
> 

The cross sectional area is much larger as soon as you move away 
from the tip.  So the velocity will fall into the laminar range.  
The vibrating reed makes the cross sectional area at the tip very 
small.  At loud volumes the reed even closes off momentarily as part 
of the cycle.  Right before it closes off, the velocity is very high 
and well within the turbulent regime of fluid flow.  The high 
velocity creates low pressure that gets the reed to pull into the 
tip.  Now closing the tip off also reduces the flow to a point where 
the "spring" of the reed can pull it back away from the tip and 
start the cycle again.  To get the cycle started, we need to supply 
a significant puff and often tongue the reed to get it going by 
closing off the area.  High baffles near the tip help with the 
Bernoulli or Venturi effects.  But they also provide additional 
accoustic reflections between the reed and mouthpiece.  The Ferron 
book has some diagrams of these.

> I've "always"  thought that, once tripped to turbulence, a 
boundary layer
> will stay turbulent --and be very difficult to recover.  In wind 
tunnels, a
> long section of the tube is dedicated to steadying the flow to get 
rid of
> the turbulence induced by the fans that propel the air through the 
tunnel.
> 

There is a transitional area where the flow can be laminar or 
turbulent depending on a number of factors.  So you may be quite 
correct that the tip turbulent flow hangs on a while longer than I 
envision.  But as interesting as this is, and as much as I'd like to 
know what is going on... so what?  I have not yet seen where 
theoretical understanding of all this will be more useful than good 
old trial and error (or structured testing).