Hi,
I made a recorder-style whishtle out of PVC pipe yesterday. This one is a great success: It actually makes a sound, the sound is beautifully clear and fairly loud and the octive tuning is perfect.
However, only one out of every three recorder-style PVC whistles I make are able to make any sound at all; at this stage I do not know why I succeed nor why I fail.
Any ideas, anyone?
My experience has shown that the “blade” in the window is often too high in relation to the windway, simply due to the concentric construction of a plastic-pipe instrument. My solution is to make a slope on the floor of the wind-way (the fipple block) so as to direct the air up to the blade … works for me 
You might care to consider, at this point, the benefit of having interchangeable whistle heads, so’s you don’t have to redrill a complete new whistle for each experiment. Also, not glueing the fipple block if at all possible. These options give better opportunity to retain “known good” parameters whilst enabling a minimum of re-engineering when trialling a new idea.
Try swapping the fipple blocks around to see if it the block that makes a difference. I too have made whistles that don’t make a sound and usually it is the block. The “blade” of the fipple changes the characteristics of the sound but the block allows the sound to be made.
John
Thanks for helping me, you guys! When I make my next attempt, I will definitely keep you posted.
Some Basic Measurements to keep in mind.
You will find many voicing diagrams at this site that are adaptable to whistles, have a look.
http://www.organstops.org/_apps/Illustrations.html
Interesting figures! I think my own whistles do not conform to most of these in general. It depends so much on what kind of sound you are after…
For the problem of getting no sound at all: I guess the most likely reason is that the ribbon of air exiting the windway does not get divided by the “blade”.
In the curved-windway tubular-sandwich construction I use the floor of the windway is exactly aligned with the blade edge. And I need to cut a very small bevel to the exit of this floor (where the block forms the lower exit edge of the windway). That is enough to deflect the wind enough to hit the blade, and so the oscillation can start and the whistle can speak.
From my own experience, I expect you are using thin PVC pipe. How the air strikes the windway blade sounds as though it is the problem, as previous commentators have said. You might find it easier to try with thicker pipe. I use polypropylene barrier pipe, which is about 3mm thick. I’ve never made a whistle that didn’t sound. Flutes, now, that is a different matter. 
polypropylene barrier pipe
… I’ve used this as well, with some success, but had problems shaping the window etc. as I found it difficult to file and resorted to cutting with a sharp knife and/or chisel, which I found to be a little hit or miss, compared to filing pvc pipe (or bamboo). What did you use to shape yours?
Some Basic Measurements to keep in mind.
- The internal bore diameter/length ratio should be between 1/25 and 1/30
- Keep the distance between the center of the voicing window and the center of the top tonehole at 10X the internal bore diameter.
- The thickness of the tube at the voicing window must be at least 1/8th inch or poor resonance will occur.
- The average size of the voicing window should be near 1/2 of the internal bore diameter.
- A good “rule of thumb” for voicing window measurements is 2/3, length/width.
1 - does anybody know where this rule of thumb comes from? - does it only apply to D whistles? - does have any basis in theory? Any way my D whistles are not in this range. I suspect that it is area/length that should be kept constant - but I have no theory to predict the value, I am working empirical data.
2 - ??? - my D whistle is close to this figure ( ~9.8 ) but not the E, Eb, or C which share the same id. I also make a wider bore C - it is not at hand to measure as I write this. I find it interesting that my instrument is close to this, but not to #1 - I have my doubts about this number/guide.
3 - my D is about half this thickness and works well.
4 and 5 match my measurements reasonably
At this stage I drill a 5mm hole and use a nail file to shape the window.
- The internal bore diameter/length ratio should be between 1/25 and 1/30
1 - does anybody know where this rule of thumb comes from?
… it’s not really a “rule of thumb”, but basic whistle theory, though a little smaller bore than I’m used to seeing quoted. The usually quoted “magic figure” is around 1/20 - 1/25, compare any normal Generation, Walton, Feadog etc. A Walton “Mellow D” is around 1/18. Ratios much lower than this, eg. 1/15, tend to have a very “pure” tone, but struggle to get much more than an octave and a half in range, whistles with a narrower bore ratio can have poor low frequency response. Do note that these approximations are only true for parallel-bored whistles. Conically-bored instruments (eg. recorders etc), reed instruments (chanters) and possibly side-blown flutes have their own rules.
I should probably wait to see if any one else chimes in…
but
I’m playing dumb (about this rule of thumb)
I know the effect of wider and narrower tubes, and that it refers to cylindrical bores, and that it is quoted about high D whistles…
what I would like to know is:
if it applies to higher and lower whistles
if any one is aware of any theory behind it (beyond hand waving)
Perhaps someone knows if organ pipes of widely different pitches share the same diameter/length ratios - I should be able to look that up I guess, not run across it as of yet.
OK so a quick search turned up a site that I have looked at before but not read this particular page.
Interesting site -note I have found things I disagree with, maybe I was wrong can’t remember details, so as always read with a grain of salt.
Here is the url: http://www.navaching.com/shaku/didge.html
what I would like to know is:
if it applies to higher and lower whistles
… yes … to any whistle … it’s basic acoustic physics. With a fixed fipple in a parallel-bore whistle, approximately 1/20 bore/length gives best compromise between easy-to-reach high notes AND reasonably loud low notes.
Yes, the design of the voicing window and the labium etc. also affect these features, but if the high (or low) harmonics aren’t there in the first place (a function of bore/length ratio) then all the clever voicing in the world won’t make them come about.
You can see these characteristics in action if you play with TWJCalc http://twjcalc.sourceforge.net/2.10/TWJCalc2.10.html … it’s really quite accurate
well I wrote a long rambling informative post which was zapped by a storm just as I hit the post button.
So here goes the short version:
You can see these characteristics in action if you play with TWJCalc > http://twjcalc.sourceforge.net/2.10/TWJCalc2.10.html > … it’s really quite accurate
There are 3 calculators on that page - I put in some sample numbers and got three answers from 41.2 - 44.9 for the bottom hole, in my book that is a significant difference.
I know you can look at cut-off frequencies (or at least an approximation of them the formula assumes some things that are not true) but I do not see anywhere where it suggests the strength of the fundamental and harmonics or what would make one choose a particular id.
A little more research turned up a rule of thumb for organ pipes - pick a pipe, go up 17 half steps, the id of this pipe will be 1/2 the diameter of the first - the math was done in 1833 and is perhaps responsible for some boring (as in not interesting) pipe organs!
A more recent (Johan Liljencrants, 2006) paper looked the relationship of the Q of an organ pipe and its width - and suggests that the above rule of thumb is close to maximizing the Q.
a rule of thumb for organ pipes
… fine, but almost as irrelevant as details for a didgeridoo. An organ pipe is optimised for one frequency, then voiced for tone. A whistle is (usually) required to operate over at least two octaves, so compromises have to be reached
Swiss Army Knife. It is a little hit and miss, but as I don’t operate commercially, I don’t care. I do find that the barrier is a useful guide to the shape of the blade: the line of the barrier should be straight, not curved, unless I’ve chosen to make a round windway hole. Mostly I prefer to make them straight for whistles, but round for overtone flutes.
For “difficult to file” I would say “impossible, or repulsive, to file”. I strongly dislike the tendency for the plastic to form lumpy threads.
I should say that my Swiss Army knife is a small model, known as “The Waiter”, with only three blades. I also use a steel to keep it very sharp indeed. It’s my ever-present implement for all sorts of uses.
It does take a bit of care, but you can get a nice smooth windway blade with a small sharp knife. It’s worth remembering that the blade does not need to be sharp, and a square edge on a blade can give a better sound.
Swiss Army Knife. It is a little hit and miss, but as I don’t operate commercially, I don’t care
… my situation exactly, thanks for confirming I wasn’t missing anything