Blowing machine

The Ultimate On-Line Whistle Community. If you find one more ultimater, let us know.
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 »

I agree with Hans's tuning suggestions.

The model forecasts the following relationship between air speed and frequency for the highest frequency of each note:

0.26 * air speed = frequency * window length

The coefficient 0.26 is open to debate, but it should be somewhere in that neighbourhood. At lower frequencies for each note, the coefficient is higher, up to 0.5 or so, and depends on the acoustic response of the whistle. If your Feadog mouthpiece has the same dimensions as mine, windway exit 1.8 mm high and 8.2 mm wide, window 5.25 mm long, your numbers give coefficients of 0.25 for low G and 0.24 for high G, which is right in line.

Now, Hans, this linear relationship between air speed and frequency doesn't mean that the blowing pattern for best tuning must have a regular increase going up the scale, but it does lean in that direction.

Regarding turbulence in the windway, it looks like the Reynolds number in the windway is roughly 200 times the air speed. This suggests the flow in the windway will be laminar for the lower notes, but the higher notes will definitely be up in the turbulent zone (Reynolds number of 2900 or more).
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 »

Terry McGee wrote: Sun Jan 22, 2023 3:40 am
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.
Two places to look would be loop gain and Q factor. Maybe the note sounds best when the loop gain is highest, or when the Q factor is highest. Unfortunately, the theory predicts that both will be highest just before the note breaks into the next higher register, which doesn't correspond to observations. Back to the drawing board.
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 11:36 am 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).
It has been touched on here. I brought it up in viewtopic.php?f=1&t=90943&p=1093777, and it came up again later viewtopic.php?f=1&t=103912&start=45.
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: Sun Jan 22, 2023 11:26 am If your Feadog mouthpiece has the same dimensions as mine, windway exit 1.8 mm high and 8.2 mm wide, window 5.25 mm long, your numbers give coefficients of 0.25 for low G and 0.24 for high G, which is right in line.
Hmmm, my whistle has windway entrance of 1.77 x 8.31, an exit of 1.44 x 7.97 and a windway length of 24.8mm. The window is 5.3 long and 8.1 wide. Are we talking the same instrument?
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 »

Terry McGee wrote: Mon Jan 23, 2023 1:39 am
Tunborough wrote: Sun Jan 22, 2023 11:26 am If your Feadog mouthpiece has the same dimensions as mine, windway exit 1.8 mm high and 8.2 mm wide, window 5.25 mm long, your numbers give coefficients of 0.25 for low G and 0.24 for high G, which is right in line.
Hmmm, my whistle has windway entrance of 1.77 x 8.31, an exit of 1.44 x 7.97 and a windway length of 24.8mm. The window is 5.3 long and 8.1 wide. Are we talking the same instrument?
Definitely. I would attribute any discrepancies to crudeness in my measuring technique. However the windway exit on mine definitely runs the full width of the window. I tried measuring the windway height again, taking an impression and measuring the impression, and got 1.6 mm. How did you measure the exit on yours?

The revised measurements give a coefficient of 0.19 instead of 0.26, which is definitely smaller than I would expect.
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'm using my "back of a twist drill as a feeler gauge" method we talked about elsewhere. I reckon it's good, and reveals details we would otherwise be unaware of.

I went and redid the windway and window measurements more carefully and in a bit more depth (more samples).

At 0mm, 2.08 high, 8.48 wide
@ 2.15, 1.82
@ 14, 1.7
@ 16.32, 1.61
@ 21.9, 1.53
@ 24.81 (windway exit), 1.5 high, 8.23 wide

Window 5.3 long, 8.23 wide

I can be pretty confident about the exit height as I have a 1.5mm drill that just goes through into the windway, and resists slightly being moved side to side. And note that a 1.53mm drill stops just under 3mm short of there. These height numbers are the diameters of the drills as measured, not their nominal diameters.

Note how caliper measurement at the entry suggests 2.08mm height, but by only 2mm in, that has dropped to 1.8mm high. It's a funnel! The rest of the way is a fairly straight and mild taper down to 1.5mm high at the exit.

I should mention in all of this that we are measuring a moulded product made from fairly soft plastic. I can see variation in window length across the width of the window for example. You have to go pretty lightly to avoid denting the tip of the ramp.
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 »

Interesting to look at the windway exit heights of the various whistles I've measured so far. Going from small to large, I see:

1mm, from a new Killarney and a very old Generation. They start very differently, at 1.2 and 1.7mm respectively.
1.14mm, a fairly recent Feadog, starting at 1.87mm
1.26, starting at 1.93 from a more recent Generation
1.28 from 2 (almost the same as above!) from a Mellow D
1.5 from 2.08, the Feadog Mk 1 under question

It would be nice to gather more windway height info if anyone is equipped to measure them.
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 am aware I'm dragging the chain here a bit on the flow vs frequency measurements Tunborough has been seeking. Apologies, it's all go here, on multiple fronts.

I did get to clean up the disaster area known as "the workshop", to make it safe to enter, if not linger. I had hoped to progress the measurements today, but ran into some remaining issues. Having reduced the "brown noise" generated by air flow in the T junction at the whistle head, frequency measurements were looking more stable, at least in the middle of the regimes. But as you approach the edges of the regimes, or from outside the regimes into their edges, the frequency as indicated by TTtuner goes pretty twitchy. Fair enough - we're not listening to a single clean tone, but to a mix of tone and chaos. It seems I'm prepared to call it a really flat G5, when TTtuner would rather call it traffic noise. It's to some extent easier to measure in pitch (rather than frequency) terms, eg F#5 plus 35 cents, but as you approach the F#5 +50 cents/ G5-50 cents border, the tuner goes all wibble-wobble on you again. Part of the problem appears to be that the tuner uses a very short sample period, giving a number of readings per second. What we really need here is a tuner with a long sample period - say 5 seconds - recalibrated to read out in Hz. I might have such a tool among my Electronics junk. Or there might be some App that offers that flexibility. Some digging needed.

Sheesh, I'm reminded of how we measured frequency back in my early days (late sixties). We had devices called Variable Frequency Oscillators. We would tune them by ear (and then zero beats) to equal the pitch of the signal to be measured, then read off the frequency from a big scale. That would work here, but I'm hoping for a simpler solution!
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 »

My Android tuning app of choice is Instrument Tuner, by Gebauer Matthias. Just spent a couple of bucks to upgrade to Instrument Tuner Pro, which offers different temperaments (not yet including JI).

If you can get a Windows machine into the lab, the Shakuhachi Auto Tuner by Tatsuaki Koroda offers different sampling rates, a spectrum analysis, and a history that you can scroll back through, so you don't have to scribble down numbers while adjusting the flow. Someone named McGee did a nice job on the English instructions for it. :wink:
trill
Posts: 685
Joined: Thu Aug 17, 2006 8:44 pm

Re: Blowing machine

Post by trill »

I've had good luck with this one:

https://shaku6.com/soft/s8tuner_e308.zip
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 »

trill wrote: Wed Jan 25, 2023 2:52 pm I've had good luck with this one:

https://shaku6.com/soft/s8tuner_e308.zip
So have I. Call up the help screen, and look at the note on the bottom for the name of the person who wrote the English operating instructions.
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 »

Oh dear, they're on to me! And yes, the "Autotuner" is a definite possibility but would require me to clean up more to make some room for the laptop. (Yeah, I know, First World problem!)

I've just downloaded the Instrument Tuner Tunborough mentions. It seems to offer the well-damped response I'm hoping for. I lashed out on the massive $2.99 for the Pro version, as I see it offers potentially useful feature of scale range. This probably only makes sense in Manual mode where you decide in cold blood to explore say G5, and the 200 cents range allows you to explore 100 cents each way without need to press any buttons. But nice not to have the tuner switching back and forth between say F# + 45 cents and G -45 cents when at the halfway point.

Amusingly, while trying it out in the lab, the compressor came on to top up its tank. Couldn't care less about whistles any more, Instrument Tuner says, but that compressor is playing 48.9Hz. Given our mains frequency is 50 Hz and induction motors run just a little behind synchronism, that looks just about right! But it's a reminder that we need nice quiet surrounds for reliable, easy measurements.
User avatar
David Cooper
Posts: 142
Joined: Thu Mar 10, 2022 5:24 pm
antispam: No
Please enter the next number in sequence: 8
Tell us something.: I'm about to have a go at making wooden flutes based on a quena - I want to experiment with changing the hole sizes and locations to make one that's more comfortable to play. I just received an auger through the post today, and there are blown-down trees in the garden waiting to be repurposed, so I'll try to make a start on my first prototype at the weekend.

Re: Blowing machine

Post by David Cooper »

trill wrote: Wed Jan 25, 2023 2:52 pm I've had good luck with this one:

https://shaku6.com/soft/s8tuner_e308.zip
That's brilliant - I'd given up hunting for that kind of program as everything seemed to run on phones instead of the PC. Here's a page in English that links to it (third item) for those like me who don't want to dive straight into running an undescribed program: https://shaku6.com/software.php
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 can certainly vouch for the Autotuner. I still use it on my Office PC for quickie tuning tasks (I have Flutini for the more in-depth analyses). Tatsuaki Koroda was really helpful when I first contacted him, many years ago, and made a number of changes to the Autotuner at my request. Chief among those was the range of the reference pitch. It started out as a few Hz each side of A 440, which was fine for modern players. It now covers 375 to 470Hz to handle anything from below Old French Pitch to above British Philharmonic Pitch. Great tool.

I always think it ironic that we who cling to our ancient traditional tunes are so well supported by and make such good use of modern technologies. Brilliant tuners, the ABC system, The Session database, TunePal & FolkFriend, The Chiff & Fipple forums, etc, etc.
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 »

Still struggling to work out how best to tune the whistle connected to my air-blowing rig. I'm accepting Hans' suggestion to tune to the note G (It's nicely placed in the middle of the tube, and generally a fairly solid note). But at what pressure/flow to tune it to G?

Where it sounds best? (That to my ear is pretty close to the breaking point to the next regime. Probably exacerbated in this case by it being a pretty quiet whistle)

Halfway through the low octave regime? (That requires making subjective decisions about where that regime starts and ends. And because the whistle will whisper weakly at very low airflows, it also lies quite a way below where I would instinctively play it.)

Where I would instinctively play it? I decided to investigate what happens if I try that approach.

I started by playing xxx ooo with the whistle in the mouth, until it stopped warming up and I could tune it reliably to G5. Instrument Tuner's good damping making that easy.

I then plugged it back into the blowing rig and went to feed it with enough flow for standard G5. I could almost get it there but it was on the edge of breaking to the next regime. And as I watched, the pitch dropped back a little.

Back in the mouth, and the pitch was now a little flat, but warmed up as I played. We all know that experience.

Back in the machine, and I watch the pitch drop back as the room-temperature air from the compressor cools down the warmed up whistle. The difference is about 25 cents, but you have to wait a while on both warming up and cooling down to see that full range of drift.

I then wondered if it would mean that when I first turn the compressor on, I would get a warmer (sharper) result. The act of compressing air generates a lot of heat. The unit has a heat exchanger located between compressor and storage tank to try to lose some of that, but you wonder if a newly filled compressor tank is warmer than one that has just topped itself up.

So I stopped the compressor, dumped all the air from the tank and restarted it, giving plenty of time to allow any new warmer air to make it though the system to the whistle. No noticeable change, so we can rule that complication out, at least with this compressor. Depending on the source of air, this might be a matter to investigate and allow for in any similar rig.

But it still leaves me with a whistle that is about to break into the next regime at G5. But maybe that's OK, because that's how I would play it. Thoughts anybody? Should I alternatively build in a little safety margin, and tune the whistle so that the break out of G5 sits say at G5 + 20 cents? Or am I just being hyperfussy, and it doesn't really matter?

This does serve to remind us how good we get at playing a whistle "close to the edge".
Post Reply