Terry McGee wrote: ↑Mon Mar 13, 2023 7:16 pm . . . It's certainly offering greater accuracy and resolution at STP. But it would be a bitter blow to pay all that money and find out then that it's highly susceptible to pressure.
I think I finally understand your concern.
I'll look into it.
Terry McGee wrote: ↑Tue Mar 14, 2023 12:07 am
Good man!
My understanding of these MEMS flowmeters is as follows. They are "mass flow" meters. They are Thermal/Calorimetric. They detect a temperature change due to fluid flow. That temp change is governed by mass-density. If the pressure goes up, the mass-density goes up. What is key is that the flow *displayed* is in SLPM "Standard Liters per Minute".
So for our (your!) test bench, the pressure blown into the whistle is the same as what a player blows. That pressure is just above ambient atmospheric. A few %. The air being blown is at a higher pressure, and so is slightly denser. But, by the time it leaves the whistle, it's at "Standard" atmospheric. That "Standard" is what the meters display.
And do the Digikey people offer something in that range?
Although before upgrading, we'd need to be sure that it was going to make a difference. I'm getting the impression that it's not so much the flow meters that are confusing us, but more the complexity of the physics involved. For example, when I added 35mm or so of tubing on to the end of a calibrator that previously went to air, I expected to see a rise in pressure across it. Resistance of calibrator + tubing is more than resistance of calibrator, so expect more pressure. But we saw a drop. I can see why I went into electronics rather than pneumatics. Sigh, and then I go and become a flute maker....
I am looking at the recorder data and wondering what the characteristics are of the departure from the 'flow proportional to square root of pressure' relationship that Bernoulli and losses in the windway would predict.
Ignoring the zero points and the 3 Pa point for the bass recorder it looks like a straight line fit of Q v sqrt(P) for all the recorders with just a hint in the lower pressure readings that the points might be starting to curve towards the origin. Fitting lines to the straight part we have (with the source units of pascals and 0.1 l/sec)
So presented that way there is a consistent trend across the instrument pitch. We can't simply suggest that their flow meter has a zero offset.
Looking at it the other way the intercepts on the sqrt(P) axis for the sopranino, soprano, alto, tenor and bass are equivalent of 11, 8, 8, 7 and 7 Pa. For the sopranino that is 0.5% of the maximum pressure measured, so less than that of the full scale of the pressure measurement instrument.
What flow mechanism could do that? Does it show in Terry's whistle data? I wonder if the resistance is in fact constant up the instrument scale. Even if it is not is the slope of the straight line portion (expressed as a resistance) a reasonable way to compare 'how instruments blow'?
Terry McGee wrote: ↑Wed Mar 22, 2023 11:18 pm . . . so expect more pressure. But we saw a drop. . .
Fair question, if the flow was exactly the same with+without the 35mm extension.
But, how do you know the flows were exactly the same ?
Of course, they were *not* exactly the same. But, to what degree were they different ? A half-diameter of the dancing ball ? A quarter-diameter ?
This is an instance of the "discrepancies" I mentioned above. In this case, the "model" was your expectation of a "rise" in pressure (which I think is correct). The "measurement" is the observed "drop" in pressure. So the question is: how much error in the flow measurement would lead to that discrepancy ?
Stated another way: ask a modeler you trust "how much difference in flow (lpm) would bring the measurement in to alignment with expectation". That answer will be a clue about how "accurate" you want the flowmeter to be. Stated yet another way: that answer will be a clue about how much error is introduced by the current flowmeters.
Terry McGee wrote: ↑Wed Mar 22, 2023 11:18 pm . . . so expect more pressure. But we saw a drop. . .
Fair question, if the flow was exactly the same with+without the 35mm extension.
But, how do you know the flows were exactly the same ?
Of course, they were *not* exactly the same. But, to what degree were they different ? A half-diameter of the dancing ball ? A quarter-diameter ?
This is an instance of the "discrepancies" I mentioned above. In this case, the "model" was your expectation of a "rise" in pressure (which I think is correct). The "measurement" is the observed "drop" in pressure. So the question is: how much error in the flow measurement would lead to that discrepancy ?
Stated another way: ask a modeler you trust "how much difference in flow (lpm) would bring the measurement in to alignment with expectation". That answer will be a clue about how "accurate" you want the flowmeter to be. Stated yet another way: that answer will be a clue about how much error is introduced by the current flowmeters.
I don't see any way to blame that drop in pressure on the flowmeter. The results are too consistent across different flows, pressures, and calibrators. Yes, we are seeing anomalies when we try to switch among the 5 L/min meter, the 20 L/min meter, and two meters in parallel, but for now I'm not ready to blame failures in the model on the tools.
Tunborough wrote: ↑Thu Mar 23, 2023 9:46 am
I don't see any way to blame that drop in pressure on the flowmeter.
Neither do I. If the pressure is seen to drop it doesn't matter whether any change on the flowmeter was noticeable or not. If Terry tweaked the regulator to bring the pressure back to what it was without the tube we may then notice the flow go up a little.
There is physics related to what happens after an orifice, and the example of venturi tubes, that suggests a tube may reduce energy loss from the system which would be seen as a drop in pressure. It seems to be more physics of a jet rather than a windway. Trouble is - they inseparable.
So, is it the case that, in the past, everyone has been relying on pressure-at-the-beak measurements, because they imitate what we do by mouth, and are easy to take. But that measured pressure cannot really be taken as the pressure drop across the length of the windway, as you'd need to subtract the pressure inside the window area. And that pressure is a boiling maelstrom of higher pressure coming out of the windway, and lower pressure in the backwater at the bottom of the head? And you'd have to measure "some way down the tube" for those forces to have sorted themselves out, but we're not really interested in pressures "some way down the tube"?
So, where do we go from here? I've been holding off doing any other measurements (eg your "same head but with two lengths of tubing attached", Tunborough) in the hope we could reach a point of confidence in the rig and the measuring equipment. Anything useful I can do at this stage?
And, remembering that my interest in all of this stems from finding me a whistle that suits me better than my Highly Tweaked Mellow D, I keep wondering about tapered whistles. I remember owning a Clarke C whistle way back in the day, but I don't see it lying around anywhere. And of course I hated the inefficient fipple arrangement! Any thoughts anyone on how best I could get some more up-to-date experience with tapered whistles without incurring too much expense? I've noticed Clarke now do a Sweetone that is in D and tuneable (both essential for me!). The saccharine-sweet name is a put-off, remembering I want a whistle with some strength. Any thoughts?
Alternatively, I'd be happy to knock something up if anyone has any dimensions that would make a good start. I did wonder about making a whistle head to suit my tapered piccolo bore, but I imagine the degree of taper might be too much?
Terry McGee wrote: ↑Wed Mar 22, 2023 11:18 pm . . . For example, when I added 35mm or so of tubing on to the end of a calibrator that previously went to air, I expected to see a rise in pressure across it. . .
Favor to ask: could you point me to the particular case (which calibrator, which tube) ? I can only find 100mm+ cases.
I confess, there have been so many cases, I have trouble keeping them straight.
Terry McGee wrote: ↑Thu Mar 23, 2023 5:59 pm . . . I've noticed Clarke now do a Sweetone that is in D . . .
I have one. Sounds ok to me. Yes, it's tapered. Slightly different feel to grip.
However, one of my *most* favorite Low-Ds is tapered/conical/profiled. It's the "Onyx" by Walt Sweet. Closest to a flute that I've found. The low octave absolutely purrs.
Terry McGee wrote: ↑Thu Mar 23, 2023 5:21 pm . . . as you'd need to subtract the pressure inside the window area. And that pressure is a boiling maelstrom of higher pressure coming out of the windway
It's also true that an ant walking East at the equator will slow the rotation of the earth . . .
(It's also true that the exponent will be mighty negative. . .).
Terry McGee wrote: ↑Wed Mar 22, 2023 11:18 pm . . . For example, when I added 35mm or so of tubing on to the end of a calibrator that previously went to air, I expected to see a rise in pressure across it. . .
Favor to ask: could you point me to the particular case (which calibrator, which tube) ? I can only find 100mm+ cases.
I confess, there have been so many cases, I have trouble keeping them straight.
Terry McGee wrote: ↑Thu Mar 23, 2023 5:59 pm . . . I've noticed Clarke now do a Sweetone that is in D . . .
I have one. Sounds ok to me. Yes, it's tapered. Slightly different feel to grip.
However, one of my *most* favorite Low-Ds is tapered/conical/profiled. It's the "Onyx" by Walt Sweet. Closest to a flute that I've found. The low octave absolutely purrs.
And I've noticed that Clarke's also offer a "Celtic". Anyone know if that's different to the "Sweetone" other than the paint job?