Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Sailing downwind faster than the wind: resolved?

  1. Nov 16, 2008 #1
    I've been following this on another forum and am planning to build a non-propeller design to sidestep the sailing references. I accept that the vehicles in the videos linked in the other two threads presented on this forum are doing what it looks like they are doing without any trickery. I also feel that the treadmill test is a valid way to test and refine a design. However, it appeared that there was some disagreement about that before the previous thread here was locked.

    My understanding is that general tone of the disagreement is what caused the thread to be locked, not the topic or actual disagreement, so I'm hoping that this thread won't be automatically shut down. I've read and agreed to the conditions of this forum and will abide by the guidelines.

    I feel I have a pretty good understanding of what is happening but would like to make sure that I haven't overlooked something that could affect the outcome of my experiments with a non-propeller design. I want to observe and measure the various forces and interactions of this under controlled conditions to optimize a small device before attempting to scale this up.

    Is the treadmill test a valid substitute for an outdoor test? If not, why not?
  2. jcsd
  3. Nov 16, 2008 #2


    User Avatar
    Homework Helper

    One way to explain why these carts work is to note that the power input is equal to the force at the driving wheels times the forwards speed of the car relative to the ground. After losses, the power output is equal to the thrust times the relative air flow through the prop, which is much slower than the ground speed. Via gearing, prop diameter, and prop pitch, the torque at the wheels is multiplied so that the prop generates more thrust than the force from the wheels, but at a much lower speed, so that power output remains well below power input. As long as the difference between wind speed and ground speed is large enough, (and perhaps not too large), the cart can go downwind faster than the wind, depending on the ratio of power output versus power input (efficiency factor), and the ratio of air flow speed through the prop versus ground speed.
  4. Nov 16, 2008 #3


    User Avatar

    Staff: Mentor

    Yes, I think so. One thing to note: on the treadmill, you start with the wheels moving. With the videos of these devices, they need a push to get them to work, which is the same thing.
  5. Nov 16, 2008 #4
    I don't think it's required to get the cart started but is more dependent on the wind speed and the surface area of the prop. The Jack Goodman cart did start to roll on its own after the brake was released, but once the propeller is spinning it does seem to help noticeably.

    I did notice that the smallest cart that was tested on the street in the wind caught enough of a gust that the wheels and prop were turning backwards briefly but that might have more to do with the traction of the drive wheels and the strength of the gust.

    By the way, a nice long extended version of the the treadmill is available at some airports, enough to go from a stop relative to the moving surface without running out of room too quickly. I don't know if the speed of those walkways is sufficient though.
  6. Nov 17, 2008 #5
    Yes, this has been resolved. In order to get there you have to tack downwind. If the best "velocity made good" (VMG) is greater than wind speed, at a given wind, then you "get there before the wind." Otherwise, no. Also, even having faster VMG you cannot make it if you go straight downwind (wind direction 180 degrees). You find best VMG by finding the point where the tangent to the sailboat-speed Vs. real-wind-angle graph is parallel to the horizontal axis of the graph (prove this). A positive example is with the 18ft Skiff class that has best VMG of ~12kn at 10kn wind. Data are from such polar diagram for skiff in the book "The symmetry of sailing" 1996. I believe that by now other fast classes can make it too. It is done easily by ice and land sailboats.
    Last edited: Nov 17, 2008
  7. Nov 17, 2008 #6

    Our smallest device works in winds down to 2.7mph. I believe most moving walkways are above that, but not by much.

    I'd love to find a moving walkway that had a surface suitable to our small light wheels. Most of those walkways are a myriad of slots.

  8. Nov 19, 2008 #7
    With regular sails or with other means the result is the same (my previous comment):

    At 180 degrees wind (wind from behind), when reaching wind speed, the relative wind at the vehicle is 0. This is a fact that cannot be changed. With wind 0 no lift can be generated on the vehicle: either by sail, or wing, or wind turbine. If passing wind speed by some external means, e.g., push, immediately the relative wind is from front, and will apply a stopping force that will reduce speed, and so forth.

    The only way to pass wind speed in the direction of the wind (180) is to tack (zigzag) downwind. If the vehicle has high lift and low drag, it is possible that the best VMG (velocity component in 180) is larger than wind speed (my previous comment), and the vehicle "gets to target before the wind." In this case, when zigzagging, always a side wind component, even small, must exist, i.e., wind direction is <180.

    This applies to all videos with wind turbines that I have seen, with, or without treadmills. Neither magic nor unfamiliar physics.
  9. Nov 19, 2008 #8

    I agree with this 100%. Going Directly Downwind, Faster than the wind, (180 Deg) is not possible. This has been my stand all along. The best data I have seen was for an iceboat in a 35 mph wind doing 34.9 mph DDW. That is very impressive, but only proves my point that it is not possible. The treadmills are very confusing, in that merely advancing against the tread has convinced some people that this is equivalent to outrunning the wind. I was confused about this also. In retrospect, advancing against the tread by employing another medium, such as air, is no more amazing than a right-angled drill; just another way of redirecting force! I hope this issue is finally resolved.
  10. Nov 19, 2008 #9
    Schroder, of course a traditionally equipped ice-boat can only do 34.9mph when going DDW in a 35mph wind. There's no lift involved when going DDW - it's always going to be Wind Speed minus Overall Drag with a traditional rig.

    Now, let the ice-boat zig zag and their VMG (velocity made good, or the downwind component of their path) can exceed the speed of the wind dramatically

    In land yachting, VMGs of 3x to 4x of true wind speed are common.


    Unfortunately, your point that "it is not possible", doesn't hold up to much scrutiny.

    Perhaps they are, but here on a physics forum, where the principles of 'the equivalency of inertial frames of reference' should rule, I'm surprised to find you stlll clinging to the notion that a treadmill running at 10mph relative to the air is somehow different than a street with 10mph of of relative air passing over it.

    And apparently, you still are.

    The device in the videos is powered *only* by the relative motion between air and a rolling surface.

    If the device advances on the treadmill running in a still air room, it is now running DDWFTTW (directly downwind faster than the wind).

    If the device is hovering on the treadmill (that is no forward or rearward motion relative to the belt), it is running DDWTSSATW (directly downwind the same speed as the wind).

    I'm not concerned about how "amazing" you do or don't find it. I'm am concerned when you say that it's not what it claims, which is DDWFTTW.

    Again, a treadmill running at 10mph relative to the air is exactly the same as a street with 10mph of of relative air passing over it. I'm sticking with Galileo on this one.

    I wish it was, but alas DDWFTTW is such a maddening brainteaser to some that it wouldn't be over if they were run over by a vehicle doing it.

    Russ ... could you help our gentlemen friend Schroder here to a lesson on inertial frames of reference? He's not going to listen to me.


    Last edited by a moderator: Sep 25, 2014
  11. Nov 19, 2008 #10
    Actually, you could sale downwind, 180 degrees, faster than the wind--just not for very long. :smile: I have a computer model with variable L/Ds and L_sale/L_keel options. I was very suprised to find that it could sail downwind, and when setup, would perpetually sale in circles with a net downwind drift each cycle.
  12. Nov 19, 2008 #11
    To discuss this in unambiguous manner, we need to define some terms. Running in circles, is not running directly downwind, but is in fact, a form of tacking. There is no argument that by means of tacking or zigzagging, it is possible to outrun the wind. That is well established in yachting and especially in iceboating. So there is no need to argue that point any more. The question here is about a wind powered vehicle proceeding directly on a downwind course, outrunning the wind on a steady basis. There is no evidence that has ever been accomplished. I have cited the best case of an iceboat doing 34.9 mph DDW in a 35 mph wind and Thin Air Design has agreed with that. However, TAD seems to think that because his propeller driven cart can advance on a treadmill, that is equivalent to moving DDWFTTW. That is where the issue is. I can easily design and build a cart which is driven by a treadmill that will advance against that treadmill. If I place two wheels on the tread, which extract power from the tread, and two wheels on the stationary floor, which are driven by the wheels on the tread, it will advance. As I said, that is about as amazing as a right-angled drill. It is a simple redirection of force. Using a propeller instead of wheels is just a variation on this idea. To extrapolate from this example, that a cart can outrun the wind, while going DDW is a fallacy. Until Thin Air can Prove this assertion, which he has never done, this argument can go on forever. My final statement is: Prove it! Do it in the wind, get it certified and then the issue is over.
  13. Nov 19, 2008 #12
    Excellent. Glad to hear that. It's amazing how many folks will *not* concede that point even after 90+ years of it being physically demonstrated on ice and a couple decades on water.

    For the record, I will state my two claims:

    A: I have built and can demonstrate on demand a vehicle which will go directly downwind, faster than the wind, powered only by the wind, steady state.

    B: Based on the principle of the equivalency of intertial frames of reference (IFOR), a treadmill in a still air room is a technically perfect environment to prove or disprove claim "A".

    One can extract energy from the difference in speed between the air and the ground. One can use Galiliean relativity to prove that physics of the cart at fixed position on the treadmill are identical to the physics of the cart moving at wind speed on level ground. From that it follows that steady advancement against the rotation of the rolling surface constitutes a perfect and valid demonstration of claim "A".

    And yet I have a device that regularly accomplishes it sitting right here on my desk and have demonstrated it countless times now in front of all sorts of folk.

    That *is* where the issue is, and fortunately for me I have time tested principles of IFOR behind me. 10mph of relative airflow is just that "relative". I can create that relative movement anywhere I wish -- in the back of a truck, on the back of truck, on the highway, in a gymnasium, inside a space station, on the equator -- ad nauseam. The cart won't care and there isn't a single scientific test available to tell those apart.

    Excellent, you have now devised a mechanism which will successfully climb a treadmill using relative motion between to solid surfaces. Congrats. Next time someone asks me if that can be done I will tell them "yes", Schroder devised such a system quite some time ago. Unfortunately, to go DDWFTTW, we need a device which exploits the relative motion between a solid and a gas, not two solids.

    I didn't "extrapolate", I did it.

    I did do it in the wind. 10mph of it to be exact.

    Last edited by a moderator: Nov 19, 2008
  14. Nov 19, 2008 #13
    I also didn't understand why earlier thread on this issue was locked - the discussion climate was rather polite, I thought.

    Regarding this downwind device, I also was a little duped and doubting at first - but after thinking a while I realized it must work. But I think it is not necessary compare with sailing, tacking and so on. Although it could be accomplished using vectorized aerodynamics.

    The theory behind is rather simple: Assume vehicle velocity = V1, wind velocity = V2 and "propeller" backwards projected pushing velocity V3 at force F. Received power Pr =
    F x V1 , picked up by rotating wheels. Consumed power Pc = F x (V3 - V2) at driving "propeller". Net power received is Pr - Pc = F x (V1 +V2-V3) . All velocities related to +Z direction. Net power > 0 as long as V3 < V1 + V2. Or relative velocity
    V3-V2 < V1.

    Some error may have occured - but the core is that consumed power is lower than
    received power because the propeller acts att lower relative velocity toward wind than the wheels against ground - at the same but opposite directed force F.
  15. Nov 19, 2008 #14
    It appears that the treadmill test is in question again. That was the original question that I asked to start this thread.

    Is the treadmill test a valid substitute for an outdoor test? If not, why not? Please state your view and explain your reasoning. I am not addressing DDWFTTW.

    For the record, I feel that the treadmill test is valid because of the replication of a moving ground/air interface as experienced in a wind. The difference in speed and direction between the two is identical. The energy available is a result of the relative motion between the ground and the air.

    A test of this would be to interchange the observer's perspective. Increasing the scale of the treadmill test in a still room would allow the observer to ride on the treadmill surface and measure the velocity of the devices mentioned as well as the air flow relative to the observer.

    What would the riding observer get for measurements? If the treadmill were running at 10 mph and the air in the room was still, the observer would measure a wind of 10 mph and a ground speed of zero when the observer is sitting on the treadmill surface. If the cart were to move at the same speed as the walls of the room, it would appear to be moving at the same speed as the wind since it would be stationary relative to the air. The cart's speed would be measured as 10 mph forward relative to the viewer on the treadmill surface.

    Let's freeze this for a moment and add a second observer outside the room watching through the window. What would the second observer see when we unfreeze the scene? They'd see the first observer moving backwards at 10 mph and the cart holding station in front of them. If the cart starts to move relative to the second observer, that movement is either added or subtracted from the speed that the first observer would be measuring the cart's progress at.

    Have I presented this correctly?
    Last edited: Nov 19, 2008
  16. Nov 19, 2008 #15

    Thanks for keeping the thread on track mender.

    Like you I am interested to hear peoples arguments for and against the application of IFOR in this case.


    PS: M_Grandin, I'm also happy to discuss DDWFTTW if you wish to PM me. My above comments were not meant to insult your thoughtful post.
  17. Nov 19, 2008 #16
    mender, I certainly don't wish to patronize anyone but at the same time I don't wish to leave a stone unturned.

    If you are not sure you understand the basics of IFOR -- the issue at the heart of the treadmill matter, I am happy to give you some good examples and explanations.

    Your thread and your call.

  18. Nov 19, 2008 #17
    To me it is obviously the same thing - if you understand the "theory" behind the machine.
    Imagine for instance the thread-mill moving at extremely high velocity. The slightest force
    from propeller transferred to apparatus would generate a corresponding enormous power generated from apparatus wheeels. Power = Force x Velocity. While that slight force
    would claim very small power extracted from wheel generator to rotate the propeller giving that slight force.

    If vehicle is hold still the wheels are rotating att speed of thread-mill. Already a fraction of max power obtained that way may be sufficient running the propeller holding the vehicle still or accelerating passing the velocity of ambient wind = faster than downwind .
    Last edited: Nov 19, 2008
  19. Nov 19, 2008 #18
    I agree. The force that is available is from the relative movement between the surfaces. A method of harnessing and redirecting that force is all that is needed to provide movement of a device. It's a matter of gearing and total drag vs force as to how fast the device moves and in which direction.

    Shroder provided an explanation of a similar interface. By using the relative motion between the treadmill and the stationary frame of the treadmill or the ground beside the treadmill, energy can be extracted and through the appropriate gearing cause the device to advance on the treadmill. The forces involved can be measured and the experiment is repeatable by anyone with the same equipment under the same conditions.

    I accept this solid to solid or ground/ground interface as a valid test. If the device moved forward on the treadmill, it would be moving forward faster than both the treadmill surface and the treadmill frame or the supporting ground around it. It is an exchange of force in one direction for movement in another.

    An intermediary step between ground/ground and ground/air would be ground/water. If a trough of water were to be placed around the treadmill and paddlewheels substituted for the wheels in shroder's device, the device would have a less direct link but would still move forward as long as the total drag from the device was less than the force generated.

    To me, having a less obvious and more tenuous medium to work with does not negate the principles involved. Nor does interchanging which surface is moving relative to the observer, since the energy is extracted from the relative movement of the media. The ground/air interface works just like the ground/ground and ground/water but with less obvious interaction visible.
    Last edited: Nov 19, 2008
  20. Nov 19, 2008 #19

    You have presented it perfectly.

    Well done.

    Again perfectly put. Thanks.

  21. Nov 19, 2008 #20
    Amazing reactions.

    All experiments I have seen on videos do not prove the 180 possibility, to my opinion.
    It is impossible in such conditions to keep the wind at 180 all the time. Even small fluctuations generate deviation from 180, which makes it equivalent to tacking at <180.

    Even a fully controlled wind tunnel experiment with a treadmill, with laminar stable flow in the exact direction of treadmill and vehicle wheels, will be hard to convince me that motion above wind speed is generated without wind direction (possibly only minor) fluctuations around treadmill and vehicle direction.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook