Blowing machine

The Ultimate On-Line Whistle Community. If you find one more ultimater, let us know.
User avatar
Terry McGee
Posts: 3330
Joined: Sun Dec 12, 2004 4:12 pm
Please enter the next number in sequence: 1
Location: Malua Bay, on the NSW Nature Coast
Contact:

Re: Blowing machine

Post by Terry McGee »

stringbed wrote: Sat Jan 21, 2023 3:25 am Have you encountered any similar problems in the measurement of air pressure? Despite my curmudgeonly whining about it, the only reason thus far given for measuring air flow is that velocity can be calculated from it more elegantly than it can from air pressure.
My concern about measuring the air pressure is that I can imagine every whistle type will vary due to the dimensions of the duct,etc. But maybe we need to prove that. I'll have a little play and report back.
User avatar
Terry McGee
Posts: 3330
Joined: Sun Dec 12, 2004 4:12 pm
Please enter the next number in sequence: 1
Location: Malua Bay, on the NSW Nature Coast
Contact:

Re: Blowing machine

Post by Terry McGee »

Now, I'm trying to get my head together on what a "full report" might comprise. Let's focus on what a series of events might look like, and what parameters we would want to record about each event.

Let's imagine we want to know all about the fingering xxx ooo on a particular whistle. Starting at zero flow and increasing, would these be the events of interest?
Low G establishes
Low G reaches correct pitch
Low G breaks to instability or next partial

High G establishes
High G reaches correct pitch
High G breaks to instability or next partial

(then, reducing flow again)
High G re-establishes
We probably don't need to remeasure High G at correct pitch, or?
High G fails to instability or lower partial

Low G re-establishes
We probably don't need to remeasure Low G at correct pitch, or?
Low G fails.

And, for each of these events, we would want to measure and record:
Pitch of note (eg G5 -35cents), airflow into duct (eg 15L/Min), air pressure (eg kPa) at opening to duct. Plus sound level in dBA @ 1M? Other?

And for all the notes? That's a hell of a lot of data! And then there's all the other whistles....

And, perhaps special case. With xxx xxx, when we go to the 2nd partial, do we do xxx xxx or oxx xxx?

I think it's past my bed time.....
User avatar
stringbed
Posts: 188
Joined: Tue Apr 12, 2022 9:36 am
antispam: No
Please enter the next number in sequence: 8
Tell us something.: Playing woodwind instruments for over 70 years and deeply interested in their history, manufacture, technology, and performance practices.
Location: Sweden
Contact:

Re: Blowing machine

Post by stringbed »

Terry wrote: My concern about measuring the air pressure is that I can imagine every whistle type will vary due to the dimensions of the duct,etc. But maybe we need to prove that. I'll have a little play and report back.
How is measuring airflow at the source not subject to precisely the same variations in duct architecture?
Tunborough
Posts: 1419
Joined: Sun Dec 05, 2010 2:59 pm
antispam: No
Please enter the next number in sequence: 10
Location: Southwestern Ontario

Re: Blowing machine

Post by Tunborough »

stringbed wrote: Sat Jan 21, 2023 3:25 amAnd we haven’t been talking about using the derived velocity for any specific purpose beyond noting that, as in any edge-blown system, frequency varies directly with it.
For my purposes, that is sufficient, maybe all that matters. I want to observe the relationship between air speed and frequency to refine a model of it. I don't plan to do that on every whistle, just enough to be comfortable with the model. I'm not even asking for great absolute accuracy; I'd be satisfied with good absolute accuracy and enough relative consistency to be able to say things like if G is in tune at some air speed, then A will be in tune when the air speed is y% higher.

After that, what I would want to do on every whistle is measure frequencies at which (paraphrasing Terry):

Low G establishes
Low G breaks to instability or next partial
High G breaks to instability or next partial
(then, reducing flow again)
High G falls to instability or lower partial
Low G fails.

And, for each of these events, we would want to measure and record:
Pitch of note, in Hertz. Maybe airflow into duct, in mL/sec, but given the prior detailed study I don't think I'd bother, as long as I know the air flow is steady during each measurement.

And, yes, a special case: For OXXXXX, the frequency where something that sounds like a flat low C-nat breaks to middle D.

Others may have different objectives, but for my purposes, that's sufficient.
User avatar
hans
Posts: 2259
Joined: Fri Nov 01, 2002 6:00 pm
antispam: No
Please enter the next number in sequence: 8
Tell us something.: I've been making whistles since 2010 in my tiny workshop at my home. I've been playing whistle since teenage times.
Location: Moray Firth, Scotland
Contact:

Re: Blowing machine

Post by hans »

Tunborough wrote: Sat Jan 21, 2023 8:51 am I'd be satisfied with good absolute accuracy and enough relative consistency to be able to say things like if G is in tune at some air speed, then A will be in tune when the air speed is y% higher.
Why would it be higher, why should it be higher? I think it should not. I try design my whistles, so that I can go from the lowest note up the scale to the cross-fingered flat seventh (C natural on D whistle) without a change in breath pressure, up and down in fact. So with air speed staying the same for all notes in the first octave (C# being a special case, and rather flat without extra push/air speed, as I care more for in-tune C nat). I increase the pressure/air speed only for the second octave notes, and I aim for an even play-ability, without pressure changes, for the second octave too. And as you know, the problems on a cylindrical instrument are only appearing for the top second octave notes (with a head bore restriction helping enormously there). A whistle should jump, for every note, from first to second octave, smoothly (no squeaky tones in the transition), and with the same amount of additional pressure, and it should jump back with the same amount of lessening pressure, ideally. A whistle, which would need constant changes in pressure beween notes, would be not much fun at all to play.
Tunborough
Posts: 1419
Joined: Sun Dec 05, 2010 2:59 pm
antispam: No
Please enter the next number in sequence: 10
Location: Southwestern Ontario

Re: Blowing machine

Post by Tunborough »

hans wrote: Sat Jan 21, 2023 9:50 am
Tunborough wrote: Sat Jan 21, 2023 8:51 am I'd be satisfied with good absolute accuracy and enough relative consistency to be able to say things like if G is in tune at some air speed, then A will be in tune when the air speed is y% higher.
Why would it be higher, why should it be higher? I think it should not.
Of course that is the whistle designer's choice. As a player, my preference is for a regular increase, so there isn't a sudden jump going across the octave boundary, B-C#-D-E and back. I've seen a graph that suggests that's what recorder players prefer, too. But that's a preference, not an acoustic necessity. My ultimate target in all this is to find guidance, whatever your preference, how you might get there.
User avatar
stringbed
Posts: 188
Joined: Tue Apr 12, 2022 9:36 am
antispam: No
Please enter the next number in sequence: 8
Tell us something.: Playing woodwind instruments for over 70 years and deeply interested in their history, manufacture, technology, and performance practices.
Location: Sweden
Contact:

Re: Blowing machine

Post by stringbed »

Tunborough wrote: As a player, my preference is for a regular increase, so there isn't a sudden jump going across the octave boundary, B-C#-D-E and back. I've seen a graph that suggests that's what recorder players prefer, too. But that's a preference, not an acoustic necessity. My ultimate target in all this is to find guidance, whatever your preference, how you might get there.
Recorder players may have a slightly different perspective on register shifts, having the benefit of the thumbhole serving as a speaker device. Either way, individual preferences indeed do vary among its makers and players. They can be — and commonly are — represented by plotting frequency against air pressure in the oral cavity. Both quantities are under operator control and can easily be measured directly.
User avatar
hans
Posts: 2259
Joined: Fri Nov 01, 2002 6:00 pm
antispam: No
Please enter the next number in sequence: 8
Tell us something.: I've been making whistles since 2010 in my tiny workshop at my home. I've been playing whistle since teenage times.
Location: Moray Firth, Scotland
Contact:

Re: Blowing machine

Post by hans »

Fascinating! I have to concede that there are different preferences, and there is no one right way. I can't recall that this issue has cropped up here in discussions. I think I never heard of particular "pressure curves" attributed to particular makes of whistles (just low to high back pressure in general). I am trying now to sharpen my awareness as how the pressure in my mouth feels playing various whistles. I don't think it needs objective mouth pressure measurements in order to gain some feeling on this issue of playing across the octave etc. But yes, great if objective measurements can contribute to better modelling and thus perhaps design improvements.
User avatar
Terry McGee
Posts: 3330
Joined: Sun Dec 12, 2004 4:12 pm
Please enter the next number in sequence: 1
Location: Malua Bay, on the NSW Nature Coast
Contact:

Re: Blowing machine

Post by Terry McGee »

I promised to look at the question of pressure testing and have a bit of progress to report. It concerned me that putting a T junction in the tubing running up to the mouthpiece would incorporate a significant amount of additional resistance between the measuring point and the start of the duct. This is because the T joiners have a bore of around 4.75mm. But I remembered that in the Gardening Department next door to the workshop, I have larger diameter T joiners used for irrigation. Their bore is 9.5mm. Further it just happens that the OD of my poly tubing fits just nicely inside them, making it easy to join up. A few turns of insulation tape keeps it all airtight and together.

And after soaking the dead corpses of the mud wasps jamming the U-tube manometer for a few days, I was able to blow them out. Urggh, they didn't smell nice. But the U-tube manometer is back in business. I also tried my Digital Manometer, but it didn't seem to agree well with the U-tube, so I'll run on the latter for now. It's not that easy to read the smaller values to any level of precision, but we're at the bottom end of the Digital unit's range also.

And I remembered that the two 20L/Min flow rate monitors each come with a needle valve. So I have the main monitor valve wide open, and the second one shut down, effectively giving me one monitor. If I need more, I open the second monitors needle valve, allowing me to read up to 40L/Min.

And I've tidied up the rig in general to make it more convenient to work with. Still a mess, but manageable for our current investigations.

And I went looking through the pictures of other peoples' whistle collections and reckon the test whistle is a Feadog Mk 1.
User avatar
Terry McGee
Posts: 3330
Joined: Sun Dec 12, 2004 4:12 pm
Please enter the next number in sequence: 1
Location: Malua Bay, on the NSW Nature Coast
Contact:

Re: Blowing machine

Post by Terry McGee »

So, I did a test run on one fingering....
The figures following each event title are: Flow (L/Min), Pressure (MM H20), Freq. (Hz), Pitch (Note, Cents)

Whistle: Feadog Mk 1
Fingering: xxx ooo

Low partial establishes 8.5, 5, 769, -35
Low partial reaches correct pitch 9.5, 7, 783, 0
Low partial sounding best 11.5, 9.5, 800, +25
Low partial breaks to instability or next partial 15, 18, 812, +60

High partial establishes 17, 21, 1549, -25
High partial reaches correct pitch 19, 27 1569, 0
High partial sounding best, 23, 33, 1570, +17
High partial breaks to instability or next partial 31, 60, 1620, +55

(Reducing flow again)
High partial falls to instability or lower partial 18, 25, 1564, -12

Low partial re-establishes 15.5, 18, 812, +60
Low partial fails 8, 8, 757, +65

Note I've slipped in a "sounding best" figure for each partial. That's a "twiddle the knob with your eyes closed" test. Wouldn't it be nice if this matched the "in correct tune" setting!

I will concede that all of this is awfully subjective. What I accept as the ends of the range, you might not.

And we may still have some development work to do. For example, in the high partial of this whistle, I've very aware of the grumble of rushing air. When I pull the whistle out of the tubing, I can still hear the noise. And when I play the whistle myself I can't. I imagine I'm hearing noise being generated by the regulators, the flow reducer, the balls dancing in their windstreams, and any other sharp edges in the system. Perhaps I need to make a reservoir in the tubing running up to the whistle head? Or even a baffle followed by a reservoir? It would slow down setting but might clean up the sound. I'm also seeing noise in the cents and Hz measurements (I'm using TTtuner, but I should try some other apps.) That could well be improved by smoothing the air supply.

Now, a word of warning from one who has suffered enough for science. OK to pull the whistle out of the tubing while the rig is running. Not OK to then put your thumb over the end of the said tubing. All the pressure now runs on to the u-tube manometer, which shoots its load of water + food colouring up onto the ceiling of the lab. Where it then sploops down on to the operator. I do hope I got all those mud wasp corpses out.

I must write a tune, The Mud-Wasp's Revenge. It will feature numerous octave jumps, and then end suddenly.
Tunborough
Posts: 1419
Joined: Sun Dec 05, 2010 2:59 pm
antispam: No
Please enter the next number in sequence: 10
Location: Southwestern Ontario

Re: Blowing machine

Post by Tunborough »

Thank you, Terry. Interesting stuff. Comparing this to my own measurements of a different Feadog Mk 1 on a different experimental setup (whistle stuck in my mouth), we're not that far off.

Low G: your minimum 757, my minimum 745 (very subjective here, as you point out), your maximum 812, my maximum 818.
High G: your minimum 1564, my minimum 1560, your maximum 1620, my maximum 1644.

I like the idea of a "sounding best" measurement. I only wish I knew how that would show up in an acoustic model.

I do notice some anomalies in your numbers. Your frequency and pitch measurements don't quite agree, particularly in the second octave, and the last measurement where 757 is -60 cents of an ET G5, not +65 (which I'm guessing is a typo).

You've got one measurement with 8 L/min at 8 mm H2O and another with 8.5 L/min at 5 mm H2O. I think that first pressure measurement is a typo. Other than that, flow is very close to linear in sqrt(P) as we would expect, although maybe a bit lower than projected for the top two measurements.

On the way up, the high partial appears at 1549 Hz; on the way down, the high partial falls at 1564 Hz. Hysteresis would suggest that the number on the way down should be lower than the number on the way up. Can you get the low partial still playing at 16 L/min on the way up, and the high partial still playing at that flow on the way down? Along with the grumble you describe, this makes me wonder if the compressor isn't introducing a low-frequency wobble in the flow.
trill
Posts: 685
Joined: Thu Aug 17, 2006 8:44 pm

Re: Blowing machine

Post by trill »

Terry McGee wrote: Sat Jan 21, 2023 7:00 pm . . . which shoots its load of water + food colouring up onto the ceiling of the lab. Where it then sploops down on to the operator. I do hope I got all those mud wasp corpses out.
Ohhh, I can picture it . . . :boggle:

I am litterally spraying my keyboard with "laugh-saliva"

Thank you for the little story :)

trill

and of course, for all the test data !
User avatar
Terry McGee
Posts: 3330
Joined: Sun Dec 12, 2004 4:12 pm
Please enter the next number in sequence: 1
Location: Malua Bay, on the NSW Nature Coast
Contact:

Re: Blowing machine

Post by Terry McGee »

Yes, these are the kind of mishaps that somehow never make it through to the pages of the peer-reviewed journals.....
User avatar
Terry McGee
Posts: 3330
Joined: Sun Dec 12, 2004 4:12 pm
Please enter the next number in sequence: 1
Location: Malua Bay, on the NSW Nature Coast
Contact:

Re: Blowing machine

Post by Terry McGee »

Tunborough wrote: Sat Jan 21, 2023 9:10 pm Thank you, Terry. Interesting stuff. Comparing this to my own measurements of a different Feadog Mk 1 on a different experimental setup (whistle stuck in my mouth), we're not that far off.

Low G: your minimum 757, my minimum 745 (very subjective here, as you point out), your maximum 812, my maximum 818.
High G: your minimum 1564, my minimum 1560, your maximum 1620, my maximum 1644.
Sigh. I spend days assembling cutting edge technology, you poke it in your mouth and get the same results. So unfair! (Heh heh, good to know I'm not off the planet though....)
I like the idea of a "sounding best" measurement. I only wish I knew how that would show up in an acoustic model.
It may need me to look at a spectrograph to see if I can see what is happening. And it might be imaginary. Let's see where it leads, if anywhere.
I do notice some anomalies in your numbers. Your frequency and pitch measurements don't quite agree, particularly in the second octave, and the last measurement where 757 is -60 cents of an ET G5, not +65 (which I'm guessing is a typo).
I'm not at all surprised - instability in the system was really messing with the tuner, so I was struggling to pick out numbers on the run.
You've got one measurement with 8 L/min at 8 mm H2O and another with 8.5 L/min at 5 mm H2O. I think that first pressure measurement is a typo. Other than that, flow is very close to linear in sqrt(P) as we would expect, although maybe a bit lower than projected for the top two measurements.

On the way up, the high partial appears at 1549 Hz; on the way down, the high partial falls at 1564 Hz. Hysteresis would suggest that the number on the way down should be lower than the number on the way up. Can you get the low partial still playing at 16 L/min on the way up, and the high partial still playing at that flow on the way down? Along with the grumble you describe, this makes me wonder if the compressor isn't introducing a low-frequency wobble in the flow.
It won't be the compressor - it only kicks in for about 20 seconds every 10 minutes or so, the rest of the time its off. It's somewhere else.

I went looking for problems later and think I've made progress. Interestingly, a lot of the spurious noise seems to have been coming from the T junction I'd located at the mouth of the whistle to allow pressure measurements. I'd fed the whistle from the stem of the T, the manometer and the source of air via the cross arms. That meant the air had to take a sharp 90 degree turn inside the T-junction just before the whistle. I've rejigged the arrangement with the airflow into the whistle across the top of the T, and the pressure take-off via the stem. So no bends. Noise is clearly less, and I note that the TTtuner readings appear to be much more stable.

I think given that improvement, I should redo all the numbers. Let me ask a question or two first though.

Setting up the whistle's tuning - should I use low A = 440Hz?
Should I tune the whistle on the compressor dry air or mouth's wet air?
If the compressor, tweak airflow for the best sounding A = 440Hz, or the middle of the low partial range = 440Hz, or other?
And once, I get it tuned and the system working, which fingering should I do first? Any preference?
Any other thoughts?
User avatar
hans
Posts: 2259
Joined: Fri Nov 01, 2002 6:00 pm
antispam: No
Please enter the next number in sequence: 8
Tell us something.: I've been making whistles since 2010 in my tiny workshop at my home. I've been playing whistle since teenage times.
Location: Moray Firth, Scotland
Contact:

Re: Blowing machine

Post by hans »

Terry McGee wrote: Sun Jan 22, 2023 3:40 am Setting up the whistle's tuning - should I use low A = 440Hz?
Should I tune the whistle on the compressor dry air or mouth's wet air?
If the compressor, tweak airflow for the best sounding A = 440Hz, or the middle of the low partial range = 440Hz, or other?
And once, I get it tuned and the system working, which fingering should I do first? Any preference?
Any other thoughts?
Using the blowing machine I think it should be tuned to it, not your mouth blowing. And I prefer always to tune to low octave G, not A. And I would use the best sounding G. At least that's how I tune: blow a nice G, then adjust the slide to get the correct tuning.
Post Reply