Old Debate: Canopies with no input, *will* fly downwind?

[kallend 1,679]

I don't believe the laws of physics have changed since I wrote this six years ago.

The atmosphere is NOT a uniformly moving block of air. Canopies (and other flying machines) can be turned by wind shears.

...

The only sure way to survive a canopy collision is not to have one.

[billvon 2,476]

>I have to agree that in a perfect world the canopy doesn't care about
>wind direction. We all know we don't live there don't we?

Well, in a perfect world, the 45 degree rule wouldn't work. We can all agree we don't live in a perfect world - and the 45 degree rule doesn't work here, either.

[Andrewwhyte 1]

Quote

I don't believe the laws of physics have changed since I wrote this six years ago.

The atmosphere is NOT a uniformly moving block of air. Canopies (and other flying machines) can be turned by wind shears.

John are you being intentionally cryptic? Your learned tidbit of information merely adds uncertainty to the question. Given that canopies can be turned by wind shears, is there good reason to postulate that there is a bias to turning downwind?

[kallend 1,679]

Quote

Quote

I don't believe the laws of physics have changed since I wrote this six years ago.

The atmosphere is NOT a uniformly moving block of air. Canopies (and other flying machines) can be turned by wind shears.

John are you being intentionally cryptic? Your learned tidbit of information merely adds uncertainty to the question. Given that canopies can be turned by wind shears, is there good reason to postulate that there is a bias to turning downwind?

I've written on this topic several times previously, for example, HERE.

...

The only sure way to survive a canopy collision is not to have one.

[topdocker 0]

I can just see Buster doing several jumps on Mythbusters now. Static line, of course!

I got a few old canopies that I'd be willing to donate. A remote control to line up into the wind and a old harness- we'd be set!

top

Jump more, post less!

[mdrejhon 8]

I have decided to split off a much-better worded post to this more-relevant thread:

Quote

Quote

Hi Phree,
It has many years ago but the old memory says that it stopped on/about the windline.

This would not happen if wind was 100% laminar flow, no turbulence, with no wind layers.

However, even in fractional mph differences have a tendancy to shift some uncontrolled light-weight aircraft (gliders, parachutes, paper airplanes, something with sufficiently low momentum that can't resist the micro turbulence) gradually towards the downwind direction where they tend to approximately linger. Also if you fly through turbulence, while keeping your brakes stowed, there tends to be bigger turns off your present heading if you're flying upwind than if you're flying downwind. (I even notice tihs -- remember I still fly a Sabre 170 after 500 jumps so I notice turbulence more than most of you skydivers do.)

Someone explained there is turns in one direction when wind speed changes in one direction, and a turn in opposite direction when wind speed changes in opposite direction. It should cancel out!? NAH! In reality, higher-speed air typically has more instability/turbulence than slower-speed air. As a result, transition from slowspeed to unstabe highspeed air will cause a turn of a different degree, than the identical mirror reverse transition from the same unstable highspeed air to the same slowspeed air.
Rinse and repeat several times. Eventually, the difference leads to a downwind heading.

Still argue it shouldn't matter? HAH! There's something else too: Even if the micoturbulences are exactly the same for both cases, remember planeforms aren't the same shape from the front as from the rear. Sudden pressure changes in front of the wing behave differently from sudden pressure changes behind the wing. Pushes and pulls from changes in pressure (wind boundary layers) will have different effects from behind than from the front. Wind tunnel tests prove that. Only a Frisbee can fly forward and backwards with the same effect. Wings cannot. Blow a paper airplane from behind (sudden gust/turbulence from behind), the airplane behaves differently than if you blow the paper airplane from front (sudden gust/turbulence from front).

THEREFORE, due to the reasons above, wings turn a different angle when transitioning from a highspeed to slowspeed wind, THAN when transitioning from slowspeed to highspeed. And you know, the atmosphere is full of wind layers and turbulence, the atmosphere is not laminar flow.

One says they are an airplane pilot and don't notice? It's not a meaningful factor: Don't worry -- metal bird pilots typically don't notice this because their plane has a lot of mass and inertia. Our wing needs to be low wingload, like paper airplanes and parachutes, and are more easily affected by the above. Keep reading.

Eventually, all the forces average out to a preferred downwind direction for many kinds of planeforms/wingprofiles as it is the flightline of maximum stability for many planeforms in turbulent air.

Just toss a paper airplane into slight breezes, the paper airplane has a tendancy to turn downwind as it hits the transition layer of the breeze. Same damn thing with parachutes (though it takes time for the forces to build up as the turns are tinier with micro-tubulence and slow wind speed changes in wind layers, etc). We're never parachuting in laminar flow air (zero eddies, zero microturbulence, identical wind speed for the whole altitude, that never happens for the whole skydive), so our parachute will definitely turn as a result of the wind speed changes. Turns are noticed. But turn where? Well, apparently, it seems to be downwind on average.

Pilots flying metal birds won't notice, as the wingloading of these aren't usually often enough to turn dramatically. (And any dramatic uncommanded turns are often followed by a startled pilot correcting heading. Dramatic uncommanded turns in certain metal birds also tend to dive the airplane a little, which just increases the panic factor of the pilot.) Takes much longer for the forces to build. Most pilots aren't willing to let planes glide themselves uncommanded long enough with no autopilot keeping compass heading. Therefore, scientific data from airplane pilots are less valid than parachute dummy drops and paper airplane tests, in the perspective of the "unconscious skydiver" uncommanded parachute flight.

Comment about overshoot tendancies: During the turns caused by going through transition layers, it is presumed that some parachutes/planeforms will 'hunt' heading more than others as it goes through random transition layers. Especially if the weather is very very stable and/or there isn't much altitude, so the downwind heading with tend to 'hunt' a little off (i.e. 20 degrees left off downwind, 30 degrees right off downwind), overshooting back and fourth past the downwind heading. Yes, the planeform's apparently tendancy to randomly hunt for a heading might sometimes massively outweigh its tendancy to point downwind. Yes, the planeform can have a built-in left turn or right turn that overcomes the weak desire to turn downwind. Yes that happens. Yes, there's been at least one unconscious skydiver that didn't land downwind, because of all the above factors. But in all the cases, there's always a preferred heading of maximum stability (point of maximum resistance to turning when going through transition layers) that goes into the equation, and that's almost always downwind.

To maximize likelihood of downwind heading, you would appear to need: (1) low intertia, bigness, like a big parachute, (2) maximum time under parachute wihch means sufficient altitude (3) more wind-speed boundaries will accelerate the tendancy (4) totally uncommanded with no body shifts or arm/leg movements to cause mico-harness-turns (5) stable planeform that automatically returns to level flight that (6) nearly no built-in turn. No planeform is perfect, but built-in turn tendancy should be easily overcome by the turning forced by the wind-speed boundaries. Most well-trimmed well-maintained not-too-old student parachute seem to fall in this category. (7) more immune to other shifts like moving body, engine vibrations, that might distort other turn tendancies. (8) and you have a wing that does have a clear noticable have a tendancy to turn when it hits turbulence;
NOTE: You can get away with just having 'most' of the above factors, but the fewer bullets that match the criteria, the downwind tendancy gets weaker, and the wing's desire to 'drift' or 'hunt' for a heading becomes greater.

So it makes sense that many wings tend to turn downwind -- especialy confirmed to happen in dummy drops (this thread), confirmed to happen with unconscious skydivers (numerous other threads), and confirmed to happen with paper airplanes in a drafty/breezy room (try it yourself). And yes, I've gotten 90%+ marks in Physics, Math and Chemistry classes back in my day, and it still makes sense.

It'd be nice to see proper scientific testing on this, though.

[mdrejhon 8]

And quoting Jerry, who said that all his test dummy drops, all turned downwind:

Quote

Quote

Hi Mark,

Interesting reading ( and, yes I read it all if somewhat long ).

***It'd be nice to see proper scientific testing on this, though.

What is your definition of 'scientific testing?'

Just curious really; so please do not take this in a negative sense.

JerryBaumchen

Tough one; defining scientific testing: Do we use the outdoors as a testing lab? Or do we reproduce the variable winds in a wind tunnel? For the former, we cannot control the winds. For the latter, we won't be able to produce all kinds of the same turbulence conditions and wind transition layers in a scientific wind tunnel.

Thus, I think dummy drop tests probably would have to be one of the most scientific, but a lot more data should be collected for future dummy drops (by any parachute manufacturer), and recording all the data.
-> Altitude, Canopy, Canopy size, Wingloading, etc (determine co-relation of these variables versus tendancy of preferred heading)

Weatherstation-style telemetry should be recorded from the dummy (compass heading, GPS position, dynamic G-forces caused by turbulence. The technology is simple nowadays -- even iPhones have GPS/accelerometer/compass capabilities already!)
-> Determine coorelation of heading changes that correspond with sudden changes in speed or turbulence (increased random G forces detected by accelerometer), etc.
-> Post-analysis can determine a bias towards a preferred heading. GPS tracking determines wind direction, and computer compass determines heading.
-> Determine whether more turbulence/more variable winds, causes more frequent tendancy towards preferred heading (i.e. downwind)
-> Determine how much 'heading hunting' occurs after the heading points downwind: Is there less heading changes during a downwind-pointing heading, than upwind-pointing heading? Telemetry analsys and a graph would show a pattern like a beacon, once enough data was collected in many test drops...

There's more data to record (for a proper scientific plan), but the above examples will likely be sufficient "to get started" making conclusions...

Test drop. Record data. Rinse and repeat. Enough times. Various sets of drops in different kinds of wind conditions.
(How many? That is up for debate. 10? 100? 500 times? Enough dummy drops to get ironclad data that's not coincidence.)

Graphs can then, thus be generated out of all the record data. The data would be scatter plotted or many different lines from many different test drops.
-> One example graph is parachute altitude (as it drifts down) along horizontal axis, and parachute heading "degrees off wind axis" along vertical axis ... This would in theory show a pattern whether it stabilized to a downwind heading as altitude decreased, showing all the data converging into compact data along the "0 degree" level.
-> Another example graph is.... Along one axis is "degrees off wind axis at landing" .... Along the other axis is data such as altitude or canopy size or variability level of winds that day, amount of turbulence (detect turbulence by using accelerometer data analysis). Scatter-plot the recorded data. Bam. See convincing pattern that co-relates with a specific variable (if any).
-> Many different kinds of graphs could be experimented with to gain easy-to-visualize insight to the recorded data.

For generating the graphs, an Excel macro or quick-and-dirty one-day .NET app could even be used to do this, as a programmer time cost-saving measure.

In *theory*, if the precision of the iPhone sensors are sufficient enough (to be determined) then a custom iPhone app, with only 5 to 10 days of good programmer time ($2500 to $5000 at a generous $500 per diem programmer time, and just use a quick-and-dirty Excel macro to generate the graphs), and a sacrifical iPhone can be embedded into the dummy, in an off-the-shelf waterproof Otterbox case surrounded by additional shock-absorbing foam. The iPhone must be more-or-less horizontal for the digital compass in the iPhone to work properly, and to maximize the quality of the accelerometer data (for post-analysis to detect turbulence conditions, with accurately-enough distinctive G-force telemetry from swinging harness versus wind changes versus turbulence.). The iPhone could even be used as a backup method to assist in recovery of a dummy becoming lost, because an iPhone can be remotely GPS-tracked too, either through additional off-the-shelf software, or as a feature built-in into the custom private iPhone app. (the special private "black box recorder" iPhone app would not be put into the app store, for competitive reasons) Or upon the parachute manufacturer's option, part of the algorithm could be open sourced to gain maximum trust in the scientific technique and to make sure data wasn't being fudged by anyone....

Ideally, instead of an iPhone, a more precision lab computer instrument would be better, but such telemetry recorders that support recording in-depth GPS/compass/accelerometer data might be expensive. Whereas an iPhone is an expendable $500 instrument and the custom software can be reloaded onto it. Parachute manufacturers doing dummy drops might have a limited budget to "slipstream" or "embed" scientific tests like these, into their regular normal parachute testing regimen. (hint, hint Aerodyne, PD, etc -- if you are reading this -- please feel free to steal my ideas from this thread) Heck, I'd do it (both the iPhone and ground analysis software) if I got two custom-made (Both with my Rainbow-circle-logo on underside) parachutes for free out of it -- or bribe a different skydiver to program this telemetry recorder app for a mass market GPS/compass/accelerometer equipped pocket device such as iPhone.

Note: Down-headingness is expected to happen more often with stable squares that have quick recovery characteristics (i.e. reserves), than ellipticals. So down-heading tendancy likely varies from canopy to canopy. Therefore, I recommend my enhanced telemetry-recorder idea be implemented next time one of the parachute manufacturers are testing a new reserve, since we are more interested in down-headingness in situations where it's likely to happen (i.e. reserves) since if a skydiver is ever unconscious under parachute, it's likely happening under a reserve.

Once a parachute manufacturer does this, how useful is the data? Hard to say, but it could lead to a study of survivability of downwind landings, and recommended reserve wingloading for survivability of an unconscious skydiver, etc. (especially if the dummy is equipped with G-force shock meters, as in a crash test dummy). Or optimization of brake stow position for maximum downwind survivability (tradeoff horizontal versus vertical speed etc). Or if co-relation of downheadingness is proven by the testing, then more accurate searching patterns for tracking a lost skydiver if we now know we should begin the searching downwind of the drop. Or other unexpected 'useful' uses of this data might occur.

[mdrejhon 8]

Added Note to above post: Wind direction measurements would need to be known too (i.e. meterological data). But for improved accuracy, another sensor may need to be added to the dummy to determine forward airspeed (pitot type sensor). This would be subtracted from the GPS track so that the leftover GPS track determine the direction and magnitude of wind. A new wind vector could in theory be be calculated every few feet of altitude. This would be much more accurate than the day's meterological data for wind direction. If no airspeed sensor is available to help improve accuracy of these calculations, then known glideslope data for the parachute could also be used instead (unless brake stops and planeforms was being changed from jump to jump), to deduce wind direction from a GPS track, while using meterological data as verification for these calculations. It can be done mathematically, and we might only need rough accuracy (+/- 30 degree wind direction granularity). Ideally, I'd prefer to include an airspeed sensor to maximize accuracy of wind direction calculations from GPS data.

[DanG 1]

Source?

I don't read anything there that explains the supposed causes for downwind turns. It makes a lot of suppostions and declarations, but doesn't actually explain anything.

- Dan G

[mdrejhon 8]

That is a legitimate question, and is which is why scientific testing is needed to really prove those out.

There is numerous anecdotes from dummy drop tests by parachute manufacturers and riggers, including two postings made in the last few days in a different thread. (There are dozens others that come up time and again, including in the Incidents forum, too.)

.

JerryBaumchen:

Quote

Hi John,

Quote

but to this date I've been shown none.

Every square reserve that I have ever drop tested with a dummy has turned down-wind; bar none.
JerryBaumchen

.

captain1976:

Quote

Quote

Quote

Quote

the square will assume down wind flight

Myth.

Fact. There are probably exceptions, but since I hang out, fly the airplanes and help at a parachute testing facility, I would put the figures in the high 90% range for an unassisted square with or without brakes to assume down wind flight.

______________________________________

Although Mythbusters isn't that scientific, it could make one kick-ass Mythbusters special.

One Quick Way to Disprove Naysayers
If one doesn't care about anecdotes, just throw a paper airplane crosswind through the air stream of a fan. Paper airplane nose tilts towards downwind as the leading edge of the plane hits the faster moving cross wind. This even still happens halfway across a room from the fan if you're flying a very light slow-moving paper airplane with large wings.

Note, This is Just ONE test case
Yes, there are more kinds of wind-speed boundaries in the atmosphere, but you've still just instantly personally disproved the naysayers by the above experiment. But, this single test case is very repeatable and reliable, and is ALL we need to prove that, at least in ONE cases, there EXISTS a way for a wind to point a wing towards the downwind direction.

Some wind-speed boundaries, are similiar, at a grander scale, with smaller turns to downwind resulting (that eventually builds up). There's all kinds. Horizontal pressure bounaries. Vertical pressure boundaries. Eddies. Vortices. You name it.

But we've definitely proven, personally, without a doubt, that at least when a wingform goes through at least one kind of wind speed boundary, there's a definite, reliable, repeatable downwind shift. (try it yourself! All you need is a fan and a paper airplane.)

More advanced paper airplane test:
Another, experimental test (using the real outdoors rather than a fan), is throwing paper airplanes crosswind in an intermittent breeze. Repeat throw many times. The noses of paper airplanes tend to rotate towards downwind more often than upwind, when you throw paper airplanes crosswind in an intermittent breeze. Count each occurance (X times turned upwind, X times turned downwind, X times no change). Consistent winds won't do much, so you need intermittency or turbulence during the flight of the paper airplane, like a gusty day, or throwing behind an obstacle that injects turbulence into the wind. This is easier with a 10 minute parachute flight than a 5 second paper airplane flight, since winds don't always vary much over a 5 second period, so bear this in mind when testing. Also easier with lighter paper airplanes with big wings will also, be more affected, than sleek fast-flying paper airplanes, so test with lighter ones that otherwise glide in a straight line in a dead-still room. (good to calibrate the paper airplane's flight path reliability first, and throw crosswind in both directions (left-of-upwind and right-of-upwind) to eliminate chance that a built-in turn in the paper airplane, is distorting the experimental data.

_______________________________

In closing, this is a legitimate question that requires more scientific tests, and is which is why scientific testing is needed to really prove these out. This may be a useful topic for a university student to take upon, or a parachute manufacturer's dummy drop test, to include additional sensor tests into their dummy to record data.

[JerryBaumchen 1,081]

Hi Dan,

Quote

I don't read anything there that explains the supposed causes for downwind turns.

Here is my $0.02 on it; and it is just my thoughts.

If you look at a 28 ft C-9 canopy from the front of a jumper, it is symmetrical. If you look at a C-9 from the side, it is symmetrical.

If you look at a modern canopy from the front, from underneath, & from behind, it is symmetrical. But if you look at a modern canopy from the side, it is definitely not symmetrical.

My 'thinking' is that as the air passes over, around the canopy, etc, it creates different elements of drag on the canopy at every finite location. These elements of drag variation, in turn, cause the canopy ( as with everything in our universe ) to seek the path of least resistance. It is my 'thinking' that these different elements of drag across the air-foil cause it to turn towards the path of least resistance.

But then I may be completely wrong & have no idea what is going on.

You asked & I have tried to answer.

JerryBaumchen