Blowing machine

[trill]

. . . this video by Phil Bleazey and thought it might be meaningful to readers of this thread. . .

That video was a joy to watch.

Such clear, simple descriptions.

Really admire the craftsmanship.

Wish I could afford one !

edited to add: hit 4 separate registers without breaking a sweat ! Magical knowledge of the shapes needed to produce the sound !

[stringbed]

It's pretty ancient magic.

Here’s how far you can get without fingerholes,

https://youtu.be/ezureVn-cJQ

and here’s how such instruments are made

https://youtu.be/gqrbwOR4KeQ

[stringbed]

Phil's approach is (I think) similar to that used by fine Recorder makers. (The results, not necessarily the methods.)

The floor of a recorder windway is normally a raised platform on the block. The block can therefore not simply be made from round stock or turned on a lathe, as appears to be the case in the video (chamfer and all). This permits the radial curvature of the windway to be set intermediate between that of the bore (at the windway exit) and none at all. This freedom is propagated to the edge and, more importantly, permits the underside of the ramp to be at an adjustable angle to the axis of the air column without forgoing the arching.

As to the method of broaching the windway (and note the tapering of its sides): https://youtu.be/lxzO5Kvo7gk?t=780

[Terry McGee]

Of course you are right, which has me wondering where have I seen a block that is essentially round but has a flattened top? And how could you guarantee it to always go in with it's flattened top parallel to the ceiling of the windway. Keep trying, I guess. Hmmmm....

Ah, yes, the verb "to broach". I hesitate to broach the subject, but I have heard that expression in regard to cutting windways before. And yet it doesn't fit with normal engineering practice does it? I think of a broach as a long multitoothed tool such as used to cut keyways, eg:



Whereas the sneaky thing about the tool we've been looking at is that it can cut a scalloped trench. Which of course rules out the kind of broach shown above. But I note that the Von Huene people use the term, eg see from 12:50 in this video:

https://www.youtube.com/watch?v=lxzO5Kv ... hop%2CInc.

Note that their cutter allows you to move the tool left and right as well to cut a lateral taper at the same time.

I guess although the tool is very different, the action of the driving machine is similar to the classic broaching machine, and so the term has followed over.

I stumbled upon another recorder making video showing another single point cutter, see at about 5 minutes into:

https://www.youtube.com/watch?v=4KH7Lj6 ... erWorkshop

Note the flashing image at right of screen showing that both ceiling and floor seem to be getting the scalloped treatment.

We shouldn't overlook the amount of stuff available about making recorders in our quests to make whistles!

[stringbed]

Ah, yes, the verb "to broach". I hesitate to broach the subject, but I have heard that expression in regard to cutting windways before. And yet it doesn't fit with normal engineering practice does it? I think of a broach as a long multitoothed tool such as used to cut keyways

That basic type of broach is commonplace in the recorder maker’s toolbox. It’s seen in motor-driven use here.

I stumbled upon another recorder making video showing another single point cutter, see at about 5 minutes into:

https://www.youtube.com/watch?v=4KH7Lj6ald4

You’ll see a broach used manually if you slow down the playback speed of that video to 25% in the segment labeled “Widen the wind-way”, here. If it’s of any interest I can describe how they’re used to cut an entire windway by hand.

[Terry McGee]

Hi all again

Sorry for lapsing quiet, too much going on! Hey, at my age, aren't I supposed to be at risk of boredom?

Still puzzling about how we define the low end of the 1st Regime. I unearthed my old Type II Sound Level Meter, and wondered if it would reveal anything. Type II is "industrial grade", whereas Type I is "laboratory grade". Probably good enough for our immediate purposes.

I've talked before about the lab acoustics not really being ideal for sound measurements. So instead of following the classic "1 Metre on Axis" approach, I just put it down on the desk 200mm (8") away from the whistle's window. And not even on axis, as that would make reading the meter harder. If we decided it is a necessary feature in future, I could play with the location of the various items to optimise ease of use and quality of results. But let's see what it tells us, if anything.

Things to note:
- couldn't measure the lowest flow rates, so just marked them <2 and relied on Pressures
- Included a zero flow line, to note the Sound Level Meter's background reading. I'm using it on "slow" and "dBA". A background reading 10dBA lower than any other reading is fine.
- Noted that it rose to 46 dBA Slow or 60dBA Fast while the compressor was running! That's OK, I'm happy to wait!
- I used the Old Gen whistle, note it too has what appears to be a higher pitched Aoelian regime below the start of the 1st, but it's much less noticeable than on the Feadog Mk1 I'd investigated previously. Probably a voicing issue, but not noticeable in use. Nobody blows that low!
- Note when the 1st Regime kicks in, it's again at a very low frequency and volume level, rising a long way in both till it snaps into the 2nd regime.
- I didn't go much into the 2nd Regime, as I was interested to see if we would see Hysteresis going back down in flow rate. It appears that we do, whereas I don't think that was the case with the Feadog. Unless perhaps I didn't investigate that thoroughly enough.
- the point 5L/Min, 8mm Pressure, 571Hz and 71dBA looks like a possible "practical" bottom end of the 1st Regime. But identifying it on the run wouldn't be easy!

Code: Select all

Old Gen D, 6 holes covered, 23 May 2023		
				
Flow	Pres	Hz	dBA	Regime
0	0	0	20	
<2	2.5	575	52	Aoelian?
<2	3	370	30	1st
<2	4	496	28.5	
<2	5	544	39.6	
2	5.5	548	42	
3	6.5	563	60	
4	7	568	63.5	
5	8	571	71	
6	10	575	74	
7	13	578	77	
8	17	581	79	
9	22	1137	84	2nd
10	29	1147	89.5	
9	22	1135	85	
8	16	1092	73	
7	12	576	76.5	1st

Thoughts anyone?

[Tunborough]

Thanks, Terry. That should be interesting when I get a chance to plot it, but that may not be until next week. Meantime, when you get a chance, I'd be interested in what happens in the next couple of registers, too.

[Terry McGee]

OK, pushed on a bit.

Things to note:
- Hysteresis noticeable at each transition
- When plotting, good to use one colour for low to max flow, and another colour for the reducing flows. Otherwise it's confusing! Are we coming or going?
- Note that the volume seems to get a bit tied up with hysteresis too, with the volume dropping just before regime failures on the way down.
- I believe I noticed some sensitivity to my position in the upper flow volume readings. This would make sense as the wavelengths there are shorter, and my broad chest (OK, rounded tummy) can be expected to mess with the sound field more.
- Not sure if the Volume (dBA) data is of much help. But if so, I can investigate the body position sensitivity futher and look for solutions.
- Interesting to note that most of the whistle range above the First Regime exceeds the 85dBA level set down as the maximum level for an 8-hour day. So don't play your whistle all day!

Code: Select all

Old Gen D, 6 holes covered, 23 May 2023, extended 26 May				
				
Flow	Pres	Hz	dBA	Regime
		0	20	
<2	2.5	575	52	Aoelian?
<2	3	370	30	1st
<2	4	496	28.5	
<2	5	544	39.6	
2	5.5	548	42	
3	6.5	563	60	
4	7	568	63.5	
5	8	571	71	
6	10	575	74	
7	13	578	77	
8	17	581	79	
9	22	1137	84	2nd
10	29	1147	89.5	
11	33	1152	93.4	
12	39.5	1155.5	95	
13	47	1158	96.6	
14	56	1162.5	98.5	
15	62.5	1723	99	3rd
16	68.5	1727.5	99	
17	79	1733	99	
18	87	1738	98	
19	97.5	1742	96.3	
20	116	1746.5	95	
22	120	1747.6	98.3	
24	143	2314	105	4th
22	112	2302	102.4	
20	113	2300	100.7	
19	96	2288	98.2	
18	85	2279	97	
17	76	2270	96	
16	69	2257	93	
15	59	1722	93.5	3rd
14	51.5	1715	92	
13	45	1704	90	
12	38	1678	80	
11	32	1151	92.5	2nd
10	27.5	1147	90	
9	21.5	1135	87.7	
8	15	1086	58	
7	11	576	76.9	1st
6	8	572.5	74	
5	6	568	70.8	

[Tunborough]

Ok, my turn to apologize for disappearing. I haven't forgotten about this thread, although it would have looked that way.

I have a request for a bit more information about the old Gen geometry: the length of the window between the windway exit and the blade, the length of the tube from the blade to the end of the tube, and the inside diameter of the tube.

- When plotting, good to use one colour for low to max flow, and another colour for the reducing flows. Otherwise it's confusing! Are we coming or going?

I'm not sure the distinction is that important. For points on the curve that are reachable from both above and below, it looks like a given flow will always give you the same frequency, whether you're increasing or decreasing.

[Terry McGee]

Ok, my turn to apologize for disappearing. I haven't forgotten about this thread, although it would have looked that way.

Ah, just remind me your name again, young man....

I have a request for a bit more information about the old Gen geometry:

- the length of the window between the windway exit and the blade: 5.23 on the sides, 5.1 in the middle
- the length of the tube from the blade to the end of the tube: 271.1mm
- the inside diameter of the tube: 12.02mm

- When plotting, good to use one colour for low to max flow, and another colour for the reducing flows. Otherwise it's confusing! Are we coming or going?

I'm not sure the distinction is that important. For points on the curve that are reachable from both above and below, it looks like a given flow will always give you the same frequency, whether you're increasing or decreasing.

I was thinking about the hysteresis of the regime changes. So, for example on the way up, flows between 16 and 22 L/Min have you in the third regime, c1728-1748Hz, while on the way down, the same flows have you in the 4th regime, 2302 to 2507Hz.

[Tunborough]

Thank you, Terry for the new numbers. After putting the old Gen numbers through the mill, the Strouhal number plots don't look too much different from those with the Feadog head. The graph of Strouhal number as a function of phase angle comes even closer to a straight line than the earlier numbers. Here are all three sets of data, the Feadog head with two different tubes, and the old Gen, plotting Strouhal as a function of phase angle. The right hand side is the high end of each playing register, and the left side is the low end. Note that there are several outliers at the low end of the first register; while it would be nice to understand those better, they are beyond where whistles are actually used.



It would take more data sets to be sure what the exact relationship is, but I think we are definitely onto something. The WIDesigner model can calculate X and R as a function of frequency. Recall that the Strouhal number is

St = f * Lw / V

where Lw is the window length in meters and V is the velocity of air leaving the windway, in m/s.

If we can estimate St from X and R, then for any playing frequency, we can estimate how fast you have to blow to play that frequency.

In particular, at the high end of every register, X is zero, phase is zero, and St is around 0.21 or so. So the air velocity at the high end of every register, is:

V = f * Lw / 0.21

Give or take 5% or so, this may work for every whistle.

[Terry McGee]

Within 5% sounds promising. If we could model to anywhere near that level of precision, the usual tweaks we make to the prototype would get us home. Wherever home is!

We've been concentrating on the old models, would it help to have the equivalent data from something more recent, or just confuse things further? The only really modern commercial cylindrical whistle I have is the Killarney treble D, which has the thin curved windway. I also have a couple of whistles that I have made also using the Sindt style construction.

I was taken aback by the "Comparison : Kerry Optima / Nightingale" discussion that some cylindrical whistles can be sharper in the second octave than the first. I wasn't aware that was possible. I know, I should get out more. If most tend flatter but some are sharper, it would suggest we could make some that are spot on. Any thoughts on what's going on here? viewtopic.php?f=1&t=114359

[Moof]

I have two low Ds that tend flat in the upper octave (Howard and Goldie), and two that tend a bit sharp (MK Pro and Goldfinch three-parter).

As pancelticpiper said in the other thread, one of the advantages of setting the upper octave pitch a bit sharp is a less stark volume contrast and a less steep breath pressure curve. As a novice, I also found it a bit easier to learn to play them in tune. You can just acclimatise to a slightly increased air requirement on lower notes, but it becomes noticeable when the higher ones need less [relative] push.

As for how it's done, I've no idea. The whistles have small differences in relative tone hole size and spacing, but that's as much as is visually obvious. There must be a reason the porridge is usually either a bit too hot or a bit too cold, but ...

[Terry McGee]

Ah, thanks for that, Moof. I think it confirms I'm still suffering from Flute-maker's Myopia. In the flute world, we assume any discrepancy in tuning is acoustic in nature, and requires attention such as bore slopes, hole placement and sizing, undercutting etc. So when confronted by say sharp or flat octaves in the whistle, I'm pre-programmed to thinking acoustics. Ignoring fluid dynamics. We flutemakers are excused worrying about blowing, we leave that to the customer to supply.

So, if I'm now shrugging off the Myopia, it's not that suddenly we can have cylindrical whistles that seem sharp or flat in the 2nd octave, it's a question of where we place the blowing resistance in relation to the octave jump. I think I'll need to play with that to get a feel for it (next time I make a whistle head).

But I think it also confirms that I'm right to be thinking in a tapered whistle future. With an appropriately tapered whistle, I should be able to exert control over both acoustics and fluid dynamics. Yes, you can see this is all about power!

The expression "an appropriately tapered whistle" has a chilling effect though, doesn't it....

[Moof]

The expression "an appropriately tapered whistle" has a chilling effect though, doesn't it....

But I think it also confirms that I'm right to be thinking in a tapered whistle future.

I'm sure that's right. While it's not clear why there's only a choice of sharp or flat, you'd think the most likely answer is playability. Not much point developing a whistle where only a limited percentage of the population can cover the tone holes, or where the degree of stretch makes it hard to play anything at more than 70bpm.