Blowing machine

[Terry McGee]

Aaarghhhh, I've been sucked in by the power of the written number! The Digital Manometer reads CM(H20), not MM(H20), and I've transcribed them literally. I've corrected the figures below. Note that also changes the Resistance figure.

And I've added an Average Resistance figure, just for fun.

Code: Select all

Old Generation		Blade to foot:	271mm		
					
Holes	Note	Cents	Flow	mm(H2O)	Res.
xxx xxx	D5	0	7.7	16	0.519
xxx xxo	E5	0	8.3	18	0.511
xxx xoo	F#5	-20	9	23	0.533
xxx ooo	G5	0	9.6	26	0.531
xxo ooo	A5	-15	10.3	30	0.532
xoo ooo	B5	0	10.8	33	0.532
ooo ooo	C#6	0	12.5	44	0.531
xxx xxx	D6	0	13.3	50	0.532
xxx xxo	E6	0	15.5	71	0.544
xxx xoo	F#6	-25	16.4	76	0.532
xxx ooo	G6	-10	18.3	91.5	0.523
xxo ooo	A6	-15	22	135	0.528
xoo ooo	B6	0	24.2	153.5	0.512
ooo ooo	C#7	-10	26.3	198	0.535
oxx xxx	D7	0	25.4	169	0.512
					
Average resistance = 			0.53		

How does that look now?

[trill]

OK, trill, here you go: . . .

Thank you sir !

I haven't processed it yet, but I'll post when I do.

In the mean time, I ran some cases with my new flowmeter.

Here are some notes on the setup:

1. Medical oxygen tank with dial-a-flow regulator (set to 6 lpm).
2. Air compressor for nebulized medicine (to clear oxygen + fill with air).
3. Floating-ball flowmeter, for use by nurses, 0-8lpm, +/- 0.25.
4. 100' (33m) of oxygen hose.

With the oxy hose filled with air, I ran up the flow to 6lpm on regulator. The reading settles out at bout 5 seconds.

At about 8 seconds, the slug of air is cleared, and pure oxygen hits the meter. The ball rises noticeably, about .3 lpm to my eye.

Here's a video: https://youtu.be/Dt5F1nEyaSc

I also ran 2, 3, and 8lpm. The increment-due-to-density in reading appears to be proportional to flow.

I haven't done image scaling yet to refine the measurement. TBD.

There it is. I'm convinced the difference between oxygen + air has a measurable effect on indicated flow-rate.

[Terry McGee]

I reckon I can get to less than 1mm readability on the analog manometer, but it is tricky. You are aware of the top of the meniscus, the bottom of the meniscus and a bit of a blur in the middle. I think precision in the comparison will be limited by
......

IIRC and I hope I am not teaching my grandmother.... but can you not float a small object on the liquid? Said small object can have a clear line marked on it. There will be a constant offset from meniscus to the line, and a slight distortion (liquid needs to support weight of object). Re-calibration may be required for absolute measurements.

It crossed my mind, DrPhill. But I couldn't think of anything that would float lightly on water, and have a predictable horizontal axis to mark.

If it turns out that we only need the u-tube manometer for high pressure uses, we can get by with the current arrangement. But if for some reason we rule out the Digital Manometer, and need to use the analog one down low, it would be good to improve its readability. Let's see what pans out....

[Terry McGee]

OK, trill, here you go: . . .

Thank you sir !

I haven't processed it yet, but I'll post when I do.

In the mean time, I ran some cases with my new flowmeter.

Here are some notes on the setup:

1. Medical oxygen tank with dial-a-flow regulator (set to 6 lpm).
2. Air compressor for nebulized medicine.
3. Floating-ball flowmeter, for use by nurses, 0-8lpm, +/- 0.25.
4. 100' (33m) of oxygen hose.

With the oxy hose filled with air, I ran up the flow to 6lpm on regulator. The reading settles out at bout 6 seconds.

At about 8 seconds, the slug of air is cleared, and pure oxygen hits the meter. The ball rises noticeably, about .3 lpm to my eye.

Here's a video: https://youtu.be/Dt5F1nEyaSc

I also ran 2, 3, and 8lpm. The increment in reading appears to be proportional to flow. I haven't done image scaling yet to refine the measurement. TBD.

There it is.

Now, have we determined that my flowmeters were designed for oxygen, or do we simply not know? There's nothing marked on the gauge about what gas it's calibrated for. And when I look up their site, I find useful stuff like:

Nitrogen flow meter
1.Remark
1).Application:widely used for air or oxygen.
2).Body Material:Acrylic
3).Feature:beautiful look,low weight,unbreakable and long life span.

Sigh!

Now, your "Air compressor for nebulized medicine." I guess that hasn't got the flow rates needed for our task? Or am I wrong?

[trill]

. . . your "Air compressor for nebulized medicine." I guess that hasn't got the flow rates needed for our task? . . .

Oh, no. Only on/off. Seems to deliver about 8lpm.

[trill]

. . . Now, have we determined that my flowmeters were designed for oxygen, or do we simply not know? . . .

Well, . . . when you put it that way . . . I guess we don't.

FWIW: the only times I've seen such meters have all been in medical settings. Only for oxygen.

Also, seems to me, oxygen metering would need to be closer than air. Costs a lot more. Here, you need a prescription for medical oxygen.

Hmmm. . .

[stringbed]

. . . adding a varying degree of concavity along the length of the windway’s ceiling. It began at its midpoint and was worked toward the inner edge of the chamfer . . .

Is this what it looks like ?

Not quite. I'm not sure if you intended to indicate chamfer but if you did, it should be more prominent. The arching is worked from the midpoint of the windway. The focus is on how it meets the chamfer but it also extends asymmetrically in the other direction.

[david_h]

I find useful stuff like:

Nitrogen flow meter
<snip>
1).Application:widely used for air or oxygen.
<snip>
Sigh!

I think that is telling us that it is not claiming to be accurate enough for purposes such as Tunborough's modelling. There are lots of applications, including probably medical ones,where flow is adjusted depending on what is happening downstream. There may be a recommended initial setting, a need to keep the flow the same, to make a record for the person on the next shift, to know you have turned it up or down a bit, that the spare machine will should give similar values, that a replacement meter will be near enough the same, etc. I was told somewhere - maybe at school - that the scale divisions on a measuring device give an idea of its accuracy/ precision/repeatability and that estimating between the divisions can give misleading illusion of precision. I still do it though - seems a shame to throw away information. But I check the pressure on all four tyres with the same gauge.

[Terry McGee]

Confronted several times now, as I have been, with incontrovertible evidence that I can't be trusted to take down any data that involves doubling readings or moving decimal points, I've come up with a new data recording format that should help. It's the same data, just reformatted:

Code: Select all

Old Generation		Blade to foot:	271mm				
							
Holes	Note	Cents	Flow1	Flow2	L/Min	CM(H2O)	Res.
xxx xxx	D5	0	7.7		7.7	1.6	0.519
xxx xxo	E5	0	8.3		8.3	1.8	0.511
xxx xoo	F#5	-20	9		9	2.3	0.533
xxx ooo	G5	0	9.6		9.6	2.6	0.531
xxo ooo	A5	-15	10.3		10.3	3	0.532
xoo ooo	B5	0	10.8		10.8	3.3	0.532
ooo ooo	C#6	0	12.5		12.5	4.4	0.531
xxx xxx	D6	0	13.3		13.3	5	0.532
xxx xxo	E6	0	15.5		15.5	7.1	0.544
xxx xoo	F#6	-25	16.4		16.4	7.6	0.532
xxx ooo	G6	-10	18.3		18.3	9.15	0.523
xxo ooo	A6	-15	11	11	22	13.5	0.528
xoo ooo	B6	0	12	12.2	24.2	15.35	0.512
ooo ooo	C#7	-10	13	13.3	26.3	19.8	0.535
oxx xxx	D7	0	12.6	12.8	25.4	16.9	0.512
							
Average resistance = 			0.53				

You'll see I'm separately recording the measurements on the two flow meters (Flow1 and Flow2), added together in the L/Min column. And accepting data from the Digital Manometer in CM (H2O) rather than MM. I've tweaked the Resistance formula to expect that.

But now the big question. Are we getting anywhere?

[trill]

. . . But now the big question. Are we getting anywhere?

I think so. Most definitely !

A question though:

From the previous old-Gen series:

. . .
And here are the Resistance test results:
Old Generation
MM H20 SQ RT A/P Flow
113 226 15.0 30
53 106 10.3 20
12 24 4.9 10
0 0 0.0 0

Resistance factor = SQRT(P)/Flow 0.501109879279097

Today's data shows:

1) a shift-of mean (.53 v. .50)

2) Remarkably stable RF.

Has anything changed in your setup besides replacing the u-tube with the digital gauge ?

[trill]

. . . to a T-junction to feed the Manometer and through to feed the item-under-test. . .

Terry, that tube between the manometer + the item under test:

1) how long is it ?

2) what is the inner diameter ?

3) are you using the same tap for the digital unit as the u-tube ?

[trill]

. . . I think that is telling us that it is not claiming to be accurate enough for purposes such as Tunborough's modelling. . . .

For the price of a new flowmeter (whose design is known: air or oxygen), we can eliminate that source of uncertainty.

Still unaddressed is the effect of source-pressure on the floating-ball flowmeter, as pointed out by david_h.

New floating ball: https://www.amazon.com/Flowmeter-2-20LP ... B07B6HQDDW
(have to find one whos design-gas is *known*)

New digital: https://www.amazon.com/cjc-Digital-Test ... B0BG78ZDZH

The usual trades: cost, convenience, accuracy.

I, for one, would be willing to contribute towards the purchase of a digital unit.

Note also the refund policy . . . Terry sure is fast !

[david_h]

But now the big question. Are we getting anywhere?

Yes I think so. The pressure measurement seems to have been sorted out. The manometer is a really good 'first principles' check on the digital meter. And from the player experience and whistle comparison perspective that, as stringbed has pointed out, may be enough.

Flow measurement probably isn't good enough for modelling and whistle head design. I think something like an orifice plate, where theory might indicate a square root relationship of pressure to flow throughout the range, should show up any irregularities in flow meter scale and would remove the risk of distraction by spurious wanderings of plotted curves. Should we just trust a digital meter?

As trill points out that last set of measurements have a very consistent resistance. A straight line fit of Flow v sqrt(P) goes almost through the origin and with or without a zero-zero point inserted gives a resistance of 0.52.

I have in front of me an assemblage that was last on the table together about 15 years ago: Feadog Mk1, Soodlums Mellow D, nickel Generation with line down the head, another blue Generation head with no line (off a tabor pipe), a brass generation tube and what I think was its original no-line head, Shaw conical D (w squashed windway). When they were last together I decided to give up and buy a boutique whistle so at least one person who knew what they were doing thought it was OK.

I have had another quick go with s8tuner.exe. Wow are they different! I'll try out my idea of recording to measure the pitch, break points and hysteresis but a set of measurements will have to wait until I have a good long time in the house on my own.

[Terry McGee]

. . . to a T-junction to feed the Manometer and through to feed the item-under-test. . .

Terry, that tube between the manometer + the item under test:

1) how long is it ?

2) what is the inner diameter ?

3) are you using the same tap for the digital unit as the u-tube ?

It's a irrigation tube T junction, with the pressure takeoff being the stem of the T. And yes, both manometers come off the same place.

The bore of the T-junction is 9.5mm, and from the centre of the top of the T to either side is 32.75mm.

From that I have a short length of nominally 1/2 irrigation tube to join it to the whistle. Hmmm, perhaps too short a piece. It currently overhangs 18.35mm from the end of the T-joiner, and that means the fairly pointy Generation beak runs right up to the joiner. That might introduce some error as the rectangular slot of the windway entry is a poor match for the circular bore of the joiner. I should probably cut a longer piece of tube so that no whistle beak can get too close to the joiner. I'll do that and check to see if it makes any noticeable difference at the high end. What do we think, allow say a 20mm minimum gap from windway entry to T-joiner outlet? That will mean bigger gaps for blunter beaks, and for calibrators. Or do we need to get really picky and standardise the gaps?

[Terry McGee]

And on the orifice plate or better flow meter front, I'm probably up for them but I'd want to be pretty confident that the effort and costs are worth it. Wouldn't want to run into another stone wall after clearing these!

I did wonder about getting a 10L/Min bead style flowmeter, if we could guarantee it's calibrated for air. That would give us better readability in the low range, and we could compare its idea of 10L/Min with halfway up the 20L/Min gauges I already have to see if they agree. A 10L and a 20L in parallel would give us a top range of 30L/Min, which I imagine is enough for practical purposes. And there's still the option of 2 by 20L if we need more.

If we went to the Digital flowmeter, I guess we'd need the 100L/Min model.