Blowing machine

[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
- readability of the analog manometer (about 1mm, but remember it gets multiplied by two)
- possible remnant stickability of the liquid
- the 1mm minimum increment on the digital instrument
- any non-linearity in the digital instrument

And once we bring the flow back in
- the readability and precision of the analog flow meters.

The notion of moving the flow meter to after the item_under_test is interesting, although I'm not sure what it gains us. And yes, very tricky to achieve with a whistle!

[Terry McGee]

Tunborough, on turbulence: "It hasn't been established ... That was idle speculation on my part. It doesn't sound relevant, so I won't attach any more significance to it."

But a good concept for us to keep our antennae up about. I am aware, listening to my various whistles as I play, that there are pretty big differences in tonality and clarity that I'd like to get to the bottom of. Bit by bit, we are alerting ourselves to the differences in these whistles. Hopefully some pennies will drop as we proceed....

[david_h]

The notion of moving the flow meter to after the item_under_test is interesting, although I'm not sure what it gains us.

The pressure of the air in the flow meter - and hence its density - would then stay almost the same throughout.

The issue is related to what trill has pointed out - what gas is the flow meter calibrated for? My hunch would be that the physics of the flow round the bead that lifts it is described by equations involving density - probably similar to the ones Tunborough is modelling at the same applied pressure and flow rate.

Ah yes Tunborough, my mistake about the water density. I had much earlier had a quick look and decided it was not worth bringing up in the context of workshop temperature changes. Haven't done the same regarding temperature in the flow meter.

[trill]

. . . 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 . . .

Stringbed,

I'm trying to understand. Is this what it looks like ?

[trill]

. . . flow round the bead that lifts it is described by equations involving density . . .

Indeed it is: Fdrag = 1/2*Cdrag*rho*v^2*Sref

I used that relation in trying to estimate the effects of density (air vs. oxygen). My current guess is that using air (instead of oxygen) in the flowmeter will yield an "indicated" flow rate ~10% lower than the actual. I ordered a flowmeter to check.

Test results pending.

[david_h]

So drag varies directly with density. Gas density varies directly with pressure. If Terry' workshop were at average atmospheric pressure (a little more than 10m of water) 0 to 360mm of water is roughly 3.5% variation in density. That may be all that would be gained by moving the flow meter, though that variation in scale spacing indicates something complicated is going on in the flow meter. I guess flow changes as the space the air is going into gets shorter.

Question to Tunborough. If a really simple test sample was made - hole in a thin plate maybe - is the modelling then simple enough to calibrate the flow meter? Or at least get a handle on any idiosyncrasies.

[trill]

. . . If a really simple test sample was made - hole in a thin plate maybe . . .

Excellent idea !

That would take care of density uncertainties due to both pressure + chemistry !

. . . though that variation in scale spacing indicates something complicated is going on in the flow meter. . .

My guess is that the mfr simply tailors the curvature of the flowmeter channel to give greater resolution for flows ranges of interest. Just a guess.

[david_h]

If one had two simple samples, maybe slightly different size holes, it might show if odd things were related to flow rate rather than the thing under investigation.

[trill]

. . . modular flow guides such work pretty much by definition . . .

Question, if I may: what does the term “modular flow” mean ?

[stringbed]

Question, if I may: what does the term “modular flow” mean ?

It means that I'm a poor proofreader. It should read "laminar." I've corrected the original post.

[DrPhill]

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.

[Tunborough]

Question to Tunborough. If a really simple test sample was made - hole in a thin plate maybe - is the modelling then simple enough to calibrate the flow meter? Or at least get a handle on any idiosyncrasies.

You would be talking about an orifice plate. It would have to be installed in a long stretch of pipe, with pressure taps at specific distances above and below the plate. Once we have a discharge coefficient, we can calculate flow rate from the pressure differential across the plate. (The formula they give for discharge coefficient depends on Reynolds number, which depends on flow rate, so I've no idea how you calculate that, but I think there's a reasonable approximation for our purposes.)

[david_h]

Looking at the equation I think the critical thing is that we are back to Bernoulli's equation. If a hole is a simple enough thing that the flow regime doesn't change over the range of pressures we are interested then a plate in Terry's sample holder should give flow proportional to sqrt(pressure). Done with two hole sizes should give an indication if departures from a theoretical curve shape are related to measured flow rates.

That would allow a correction for unevenness in the flow metre scale (but not a full calibration) so the the shape of a Flow - Pressure curve would be more related to the thing
under test.

There is chance everything may then get simpler to model and attention could get beyond the windway. Or, of course, that it may suggest a change of flow regime in the windway so harder to model but more interesting.

[Terry McGee]

OK, trill, here you go:

Code: Select all

Old Generation		Blade to foot:	271mm		
					
Holes	Note	Cents	Flow	mm(H2O)	Res.
xxx xxx	D5	0	7.7	1.6	0.164
xxx xxo	E5	0	8.3	1.8	0.162
xxx xoo	F#5	-20	9	2.3	0.169
xxx ooo	G5	0	9.6	2.6	0.168
xxo ooo	A5	-15	10.3	3	0.168
xoo ooo	B5	0	10.8	3.3	0.168
ooo ooo	C#6	0	12.5	4.4	0.168
xxx xxx	D6	0	13.3	5	0.168
xxx xxo	E6	0	15.5	7.1	0.172
xxx xoo	F#6	-25	16.4	7.6	0.168
xxx ooo	G6	-10	18.3	9.15	0.165
xxo ooo	A6	-15	22	13.5	0.167
xoo ooo	B6	0	24.2	15.35	0.162
ooo ooo	C#7	-10	26.3	19.8	0.169
oxx xxx	D7	0	25.4	16.9	0.162

Notes:
Excuse the truncated column headings - needed to keep in line with figures when transferred to this forum
I put in the fingerings so I could show what worked for D7 (I didn't seem to like oxx ooo). Would also be handy for c-naturals if added.
The cents column shows I couldn't make it up to 0 in some cases. Arguably should have shortened the Blade-to-foot length. Regard the ones that made it to 0 as very close to breaking! The whistle would probably benefit from touching up the tuning, but who's going to bother with such an old whistle!
Flow as shown on 20L/Min flow meter up to G6, adding second meter in parallel from then on.
Pressure as shown on Digital Manometer. It is much easier to read!
Resistance as calculated by my dodgy simplistic method (until we agree on something better)! Left to 3 decimal places though of course data doesn't support that!

[Tunborough]

Something's off with those numbers, Terry. In the first batch of measurements on that whistle, it took 27 mm H2O to get 10 L/min of flow, which was a bit less than I expected. This time it took only 3 mm H2O to get that much flow, which is much less than I would expect. I realize the first batch was on the manometer without dish soap, but that doesn't explain that much difference.