Tunborough wrote: ↑Wed Feb 01, 2023 7:36 am . . .you might want to check this by (
on one whistle * ) measuring air flow and pressure over a wide range, and plot air flow and the square root of air pressure to see if you get a straight line that goes through zero.
OK, a bit of progress to report here. I've done this test on the Feadog Mk 1 at four flow rates (0, 10, 20, 30, 40 L/Min), and graphed the square root of air pressure as Tunborough had suggested. Sure enough, I get a straight line that goes through zero. Well, straight within the limitations of the measurement technology, but no question.
I did the first test with all the finger holes uncovered (OOO OOO). Then repeated it with all the holes covered (XXX XXX). No significant changes. So did the same test again having removed the fingering tube from the head. No significant changes. Tried covering the window, pressure went up very slightly, flow remained the same.
So then I repeated the test with the thin-windway Killarney. Stupidly applied the 40L/min flow which was still set up from the previous test. Far too much flow for the Killarney, so the dramatically increased back-pressure shot the blue manometer liquid out the top again. Fortunately didn't make it to the ceiling this time. Moral of the story, always reduce flow to zero between models! Refilling the manometer and proceeding more carefully, I found I could only get this one up to 30L/Min. A taller U-tube Manometer could push it further, but I think this is plenty, so I've now made that the limit on the testing.
So then repeated the test with the large windway Tweaked Mellow D. It required a little more pressure than the old Feadog to achieve the same flow. Or, if you'd prefer to look at it the other way around, the same flow generates a little more back pressure. But nothing compared to the Killarney.
Looking at the back pressure of all three, at 30L/Min flow rate (so we can include the Killarney) we find they are:
Killarney: 340mm H20
Tweaked Mellow D: 170mm H20
Feadog Mk 1: 134mm H20
Regard these as interim results for the time being. Making sure all these funny-shaped beaks can form an airtight seal with the tubing feeding it is quite a challenge, and of course any leakage here is fatal to accuracy. I'd like to come up with a better attachment system. It would then be wise to at least spot check the findings.
I did also try what happens if you don't have any whistle connected. I thought it might give me a figure for the resistance of the tubing below the point where the Manometer is connected. But I got negative results - the level dropped a few mm. M. Bernoulli messing with me, I suspect.
We can say safely that the pressure needed by a whistle to set up a certain flow (and thus airspeed) is a windway thing, and unrelated to things going on from the window downwards. Presumably, we could assign each model of whistle with a value for its resistance. Measured in MM H20/L/Min or some similar unit? And presumably we could check that value against predictions we might make by measuring the windway length, height, taper, etc of the whistle.
It also probably means we can leave off taking pressure readings while investigating regime boundaries etc, as they can be calculated from the flow readings, using the resistance value mentioned above. That would speed up the regime readings.
One source of minor error I noticed is the tendency for the last of the liquid in the manometer to take a while to come back down the tube and rejoin the body of liquid after a high reading. But it's pretty minor.