
Here's a fun little cross-section; this whistle visually looked indistinguishable from the really good ones, and my caliper didn't show any significant differences externally.
This whistle made no sound other than a thin, quiet series of overtones.
You can see in the highlighted areas that a layer alignment issue created roughness early on in the windway I couldn't sand out completely, but the real killer is the divergence at the windway exit, caused by poor orientation strategy while printing. Having the crispest possible 90deg angle at the windway floor and windway exit face seems to be essential for this configuration of whistle.
It's by far the strangest problem I've run into. Usually even a deformed whistle plays (and sometimes quite well), but not even getting a BAD tone out of a whistle that visually resembles the ideal is a new one for me.
I changed the printing orientation to fix the warpage, and outset the windway face slightly so I can sand it beyond the rounding of that corner caused by the loss of resolution.
I use a convergent-divergent air path originally inspired by what I remembered of the DeLavel rocket nozzle (I was a big Rocket Boys fan, if you recall that memoir by Homer Hickam), and later discovered this approach was also used by some recorder makers. I've seen whistles use convergent windways but I've never seen a convergent-divergent one.
Anyway, what it does is increase the speed of the air at the windway exit relative to the speed of the breath, which makes for very responsive player feedback. My theory on the divergent portion (the airblade cut) is that it creates an area of lower pressure at the distal portion of the airblade, which helps to quickly dissipate spent vortexes after they are shed from the splitting edge.
(This, as with everything, may be entirely incorrect.)
What I do notice, though, is the combination of the extended ramp and the divergent air path seem to create a "two stage" effect in the player feedback. The first octave is loud and resonant at relatively low pressure, but the high octave seems to engage an additional element that stabilizes at a relatively higher pressure than I'd expect. It's still nowhere near a stiff blower, but it has the effect of helping to correct the octave disparity inherent to cylindrical instruments. The best way I can describe it is that where a typical whistle has an "ideal tone pressure" and an "ideal tuning pressure" in the second octave, where the tuning pressure may be greater than the tone pressure, in this case the gap between the two is quite reduced, making for both better tuning and better volume balance.
Again, I can't be
sure that some other combination of factors is causing this, but it's what I immediately noticed after the addition of the ramp and divergence. One of the articles I'm going to post is about my "whistle testbed," a whistle where I could put in different mouthpieces, blades etc on the same head for comparison of factors.