Recently, [AlphaPhoenix] weighed an airplane. Normally, that wouldn’t be much of an accomplishment. Except in this case, the airplane happened to be in flight at the time. In fact we’re not sure what is more remarkable, as he not only weighed real actual airplanes but a paper airplane too!
To test the concept, a large scale is made from foamcore and four load cells which feed into an Arduino which in turn is connected to a laptop for a visualization. After a brief test with a toy car, [AlphaPhoenix] goes on to weigh a paper airplane as it flies over the scale. What we learn from the demonstration is that any weight from a flying object is eventually transferred to the ground via the air.
In the second part of the video a new, smaller, type of scale is created and taken to the airport where airplanes flying overhead are weighed over the course of three days. This new apparatus is basically a pressure sensor enclosed in a nominally air-tight box, essentially a fancy type of barometer. Measurements are taken, assumptions are made, and figures are arrived at. Unfortunately the calculated results are off by more than one order of magnitude, but that doesn’t stop this experiment from having been very cool!
If you’re interested in weighing things for fun or profit be sure to check out Hackaday Prize 2022: Arduino-Powered Weighing Scale Has A Real Analog Display or Reverse Engineering A Bathroom Scale For Automated Weight Tracking.
Isn’t random gust of wind gonna change the air pressure more than any airplane flying by possibly could?
Yes
I mean, at some point there’s a line of accuracy where you can’t really say you’re “weighing” something any more. I looked at the picture in the article and said, “Yeah, about 150 tons”. I’m off by a factor of two or so, but I’ve done better than this “weighing” apparatus.
Yeah, this really isn’t weighing anything. It’s just click bait.
Would John mind off his payment for this article was under by an order of magnitude?
The video is not about weighing. It is about showing momentum conservation when producing lift. I think this article misrepresents that a bit.
Passive way of seeing if someone’s sneaking a plane across borders.
Or seeing which plane is carrying large quantities of gold out of the country,
The true professionals doing that would fill the unused space in the plane with helium or even hydrogen to fool the weight detectors (/s🙂)
It’s G-ERTI, isn’t it?
Off by more than one order of magnitude?
You’d do better pointing your camera at the plane and asking an AI assistant what plane it is and what it weighs.
Don’t bother with an AI. NEVER bother with an AI.
Point FlightRadar at the plane, it will give you info on what plane it is, then look it up on Wikipedia.
The name is WonkoTheSane for a reason. Listen to this sage advice re: AI, otherwise evntually risk a T-800 knocking at your door . . . . .
Hey, any T-800 is any time welcome to my place. I either get a IRL, working AI assistant, i get killed by a T-800 or in one a billion chance, i kill the T-800 and get to hang it on the wall. Any way, what’d be cooler than that?
Wait, so we do have anti gravity?
Despite accuracy issues I’m impressed with the cleverness of this.
I’ve seen some pretty impressive video’s from AphaPhoenix but for this one I would have given him more credit if he realized this can never work and if he did not make the video at all.
I mean it did work at the small scale and it was pretty clear that he knew he wouldn’t get the full weight of the plane at full scale. He still managed to demonstrate the concept, explain the physics, made a barometer out of a strain gauge, and made a quite entertaining video to boot! Full credit from me :)
It did work in the sense that you can get individual readings off the sensor, and fiddle with the setup until it shows you the correct number of grams that you already know is going to be the answer, but then that’s just fitting the test to the theory.
To actually proof (prove) it, you’d have to take differently sized, heavier and lighter, faster and slower, higher and lower, paper airplanes and show that the results generalize before you can say you’ve “weighed” the airplane based on the pressure of the downwash.
For one thing, lift is about conservation of momentum – it doesn’t say where that momentum is ultimately going. With the wing tip vortices for example, the spinning mass of air has momentum that is conserved even though it’s going in a circle. In other words, you don’t absolutely have to push off the ground via the air to get lift – lift should work perfectly fine even if there was no ground whatsoever – so this is where the theory is wrong,
AlphaPhoenix is a physics channel. Explaining the physics is the point. This is not a video about ‘weighing an airplane’.
Of course it’s not about “weighing an airplane”, which makes the title misleading (As David in TX also mentioned below). And I simply boycott video’s with misleading titles (or clickbait). So I have not seen the video, and won’t see it either. It’s so annoying when even the makers of the video’s themselves can’t be bothered to put a fitting title on it.
How is it clever if it doesn’t actually work?
it doesn’t successfully weigh the plane but it does successfully detect the pressure wave. i think that’s pretty cool. i think hackaday chose to highlight the fail rather than the win shrug
Detected the pressure wave and made a measurement. Job accomplished. Theory is a bit off, but otherwise good starting point
But it did work. Are you criticising detecting an effect or measuring an effect? Accurately measuring an effect is best but simply detecting the effect is a pretty good experiment.
It’s overselling the concept and fitting the test to the theory, which is a bit bad as far as science is concerned.
Because it’s a real science channel (not pop sci), he explained the hypothesis, did the experiments at different scales, showed the results, and proved it wasn’t simply an artifact?
Of course, but then nobody questioned the fact that airplanes emit pressure waves that can be detected at ground level. The problem is in claiming that these pressure waves are a direct proxy for the mass of an airplane or the lift that it’s generating.
Indeed, you don’t need to “push off the ground” to get lift by accelerating a fluid such as air. Lift works perfectly fine in an infinite column of air with no ground beneath. The momentum imparted by the wing, which is what it is actually pushing on, is conserved in the circular return motion of the air (angular momentum), so not all of the conserved momentum has to be linear momentum that reaches the ground with the pressure wave, and no pressure wave necessarily reaches the ground. The amount that does depends on the shape of the plane, the altitude, winds, air density, temperature gradients, all sorts of effects that throw multiple variable fudge factors in the attempt to “calibrate” the sensor to measure the mass of the airplane.
All that means is, the “theory” works if you tune the setup to work under particular conditions, when you already know the mass of the plane in advance so you can tune the system to measure it.
“Lift works perfectly fine in an infinite column of air with no ground beneath.”
no it doesn’t. there’s no infinite column and all finite columns will disperse unless confined. the ground doesn’t have to carry the force in a straightforward fashion but the confinement has to eventually carry the force of all pressure waves or the material disperses, no exceptions.
It’s a bit backwards as far as science goes. Usually the starting point is showing that an effect exists (measurement), then forming a hypothesis that explains the effect and quantifies the expectations – including alternatives that might also explain it – then formulating a prediction based on the most likely hypothesis, then repeating the experiment to find whether the prediction works.
Here we got the hypothesis first, then an experiment that shows there is an effect that fits the hypothesis under certain conditions, but here’s the catch: the effect that is measured is not necessarily the same thing or relevant to the hypothesis. This is fitting the test to the theory.
When you fit the test to the theory, you overstep and ignore the alternative hypotheses and other sources of error that might cause the same effect. You’re liable to accept any result that looks like the one you’re expecting as confirmation of the hypothesis, when in reality you’re modifying the circumstances to cause the expected effect yourself.
A classical example is the weighing of the electron: back in the day when people attempted the experiment, they got the mass wrong. Subsequent scientists who repeated the experiment got it right, but because their results didn’t agree with the previous ones they thought their setups were faulty and they started tweaking them until they got the same wrong results. The right thing to do would have been simply to publish the disagreeing results, but then you’d become open to criticism for being sloppy with your experiment.
Or in the case of a youtube video, if your experiment doesn’t meet the expectations, you don’t really have a video to publish. Nobody wants to see you throw a bunch of paper airplanes and go “welp, I got nothing.” That’s why a youtube channel rarely features real science.
i love that “The right thing to do would have been simply to publish the disagreeing results, but then you’d become open to criticism for being sloppy with your experiment.” is truly the recommended path. almost all scientific journal articles i’ve read have a section where they explain the slop. rather than waiting for the accusations, they’re straight up about it. in medical research it’s especially brazen, “n=10, self-selected walk-in clinic patients.” it’s still valuable information and they (ideally) don’t hide the weakness
It reveals a different methodology which might be refined in the future, and it’s unexpected and thus might spur some inspiration in others
Super interesting video, misleading video title.
What about detecting small meteor strikes in the upper air via the multi array of pressure sensors? Seems this could do both count strikes and “weigh” aircraft.
The failure here isn’t in the accuracy of the result. The failure here is the assumption that there is a definable relationship between the measured pressure wave and the weight of the plane. The undefinable influencing variables are many, with unsupportable assumptions necessary to come up with a result at all. An incident sensor is not a measurement … it’s like pretending “yes-or-no” translates to “how much”.
Then again, I suppose that an argument could be made that, as a general condition, you don’t really know how undefinable the variables are until you do the experiment and find out how far off you are.
Momentum conservation says these two things must not only have a definable relationship, but be identical. The scale was a sucky pressure sensor, but it WAS accurate at small scale with the paper airplane
Is the weight you actually measured a relatively constant proportion of the true weight? I’m wondering if instead of trying to extrapolate the rest of the data to get a prediction from the measurement, could you use some of these measurements to calibrate or build a model? I know that’s kind of beside the point of showing the conservation, but I wonder how accurate it would be.
A generic model would have to have height as an input, but assuming all the planes pass at around the same height, it could be treated as a constant . Or you could do large scale tests with the paper airplane to build a model for how the measurement changes with height, then extrapolate that model to real scale.
The faulty assumption is that all momentum must be linear momentum directed towards the ground and not, say, angular momentum conserved in the vortices that form after the wing has passed, or simply dissipated into random eddies and heating the air along the way to the ground.
In other words, you don’t know how much of the momentum should show up in the pressure wave that you’re measuring.
For more fudge factors, consider that there can be temperature gradients or even inversion layers near the ground that result in differences in air density and the speed of sound. These are effectively impedance mismatches that will cause part of the pressure wave energy to be reflected off before it reaches the sensor.
A mirage is formed when there’s a layer of air 1-2 meters off the ground that is considerably hotter than above, and the density difference is enough to bend light coming in at a shallow angle. Just this is enough to throw off your measurement, so you might be better off placing the sensor on the top of a 30 ft pole, preferably well above the treeline to keep it out of the turbulent wind layer next to the ground.
Because you created special conditions for it. You may ask, is the small scale setup in any way relevant to the large scale test when you don’t have plexiglass walls a mile high around the airplane that you’re measuring? Does the behavior of air currents change when you scale up from a paper plane to a jumbo jet and change the time scale by 100x?
For a point of reference, see for example the experiment of pouring down a laminar stream of water from a tall tower. There is a point where the water will just disperse into a mist from air resistance as it approaches terminal velocity. The small scale test done from some centimeters to meters would suggest that it just keeps going down in a narrow stream no matter what the height, so you might falsely assume that it’s the same for 100 meters or 1000 meters.
To have a definable relationship you need to be able to define and quantify the variables, and that most certainly is not the case here. Wind, density gradients, wave reflections off terrain or nearby structures … please explain how you would quantify any of those. Even the profile of the wave coming off the underside of the plane isn’t going to be known since the lift isn’t uniform and the non-uniformity is going to create interactions. Even the turbulence from the engine thrust will have an influence on the downward wave from the lift.
You clearly haven’t thought this through very well.
Good !
Now improve the thingy to compute what is the airspeed velocity of an unladen swallow.
And eventually carrying a coconut…
Somewhat related to this, there are youtube video’s of whole buildings collapsing because of the downwash of an overflying chinook.
IMHO, ballpark figures using ordinary high school physics topic. Not magic. Basic science and proper planning/thinking, even though it is quite good ballpark.
As a side note, you can weight your car without expensive equipment, cargo, passengers, and all. All you need is the timed accelerometer measurements matched with the current speed. That will give you the mass, which, as you know, multiplied by G, will give you the weight. You already have accelerometer and timer in your cell phone, and all you need is the speed readings (or you can go fancy and use the GPS or cell-towers triangulated coordinates to calculate such). Yes, you can weight your car with your cell phone. Ballpark figure, not absolute reading (rolling resistance, maybe some air resistance, too), but quite good ballpark figure, enough to use as the basis for calculating the distance that can be covered on one tank/charge of gas/battery.
The mass of your car dominates the drag equation at low speeds, and the aerodynamic coefficient dominates at high speeds, and the number of stops and accelerations is also unknown, so you can’t extrapolate from mass to fuel economy without knowing a bunch of other stuff which are frankly going to be more important than the weight of your car.
Yep, tho, number of stops and accelerations, too, can be extrapolated/triangulated from GPS or cell tower signals.
IMHO, fuel economy is almost always ballparked; if I am to design something properly, for ICE engine it should factor in things like ambient temperature/pressure/humidity (important), engine temperature, cold, hot, etc; readings from both oxygen sensors (engine and exhaust), fuel grade (this one is a doozie, just HOW one averages the fuel tanks’ contents’ octane number?), CCs, etc etc.
Not impossible, just reasonably sophisticated to be just outsourced to some kind of Arduino reading sensors, more like “start with the Carnot Cycle and work backwards to the actual/particular engine”.
Ah, the mass of your car has zero bearing if it is at rest OR moving with constant speed. Mass only shows up as a thing once the speed changes, which is how it can be measured.
Aerodynamic coefficient is usually measured by the certifying agency, so can be looked up in the public documentation. I don’t know what speeds they measure that at, but I would wager “highway speeds”, which is good enough basis to start with/from.
It may be instructive to actually try doing this calculation to see what extra information you need, and which you have not provided here. Hint: F=Ma.
It’s interesting to compare the “weight” pressure you might expect, to the actual acoustic radiation pressure you get from a low-flying airplane passing overhead.
With the help of some spherical cows, I get about 10 watts of acoustic power hitting that 0.1 square-meter force sensor sensor. The radiation force is simply the power divided by the speed of sound, 10 W divided by 330 m/s = 30 mN. (yes, the units work out: a Watt is just a Joule/second is a Newton-meter per second. Thanks, SI units!).
So, depending on whether you believe the speed-of-sound momentum transfer mechanism explains all the lift or not, the acoustic radiation pressure from a jet is quite a bit less force, but still substantial.
(For the unfamiliar: Yes, acoustic radiation pressure is a real thing, just like light photon pressure. It’s a gold standard method of measuring power output from acoustic transducers. It’s also a million times easier to measure than light radiation pressure, for obvious reasons.)