Wing Sails on Proas!
There are interesting things happening in the electronic control world which make it easy to experiment with control systems e.g. http://www.arduino.cc , which makes it tempting to investigate the possibilities.
Amen! I’ve done a couple of small arduino based projects recently and the power, versatility and simplicity of the things boggle this old mind (at less than $10/pop). But still can’t hold a candle to the Worsley cam in terms of simplicity, reliability and bang for the buck. Before the cam popped up I had been cogitating on an arduino based proa self steering gear using some of the current RC stuff which has also improved greatly over time.
Cheers,
Skip
when close-hauled, very little wing deflection is wanted for best VMG (hardwings typically have their max L/D at 4-6 degrees AoA), while at a beam reach apparent (broad reach, true), the wing wants to be operating at max Cl, which is closer to 22-25 degrees (for a 1-element hardwing)
If the “working range” or your wing is going to be sort of +/- 16-20 degrees, this is going to be difficult to maintain at optimal with a non-automated system. Hand-holding the cam or hand-setting the tail’s deflection will thus lead to even less optimal average settings of the wing, over time.
Thank you Dave. That really helped me understand the relationship between heading and the way the cam is acting between being in irons and wind abeam!
Here’s a question; is the assumption that the apparent wind is never aft of the beam? Because if not, I haven’t yet figured out the second half of the story—wind abeam to wind astern. Why does the angle decrease there?
Thanks for your patience!
Here’s another thought. It seems like the cam needs to be pretty rigidly located. it probably slides on plain bearing skids (UHMW strips?). But how do you get super rigid adjustment of the cam?
Perhaps this is a job for a non-backdrivable gear train. I know that a worm gear on a regular gear, like a guitar tuner, is non back drivable. Would a worm gear on a rack be? Then you have your locking for free, and beautifully precise adjustment of the throttle from a loop of line you can take wherever you want on the boat without cleating it—pull one side to throttle up and the other to throttle down. The loop would be linked to the worm gear by two oppositely wrapped furler drums, for lack of a better metaphor. Now all of the slop that can come from line stretch is gone.
No?
chris
I suspect (and hope, because it would be very cool) that the tailplane is fixed and that the trimtab is enough to control it. This may be possible if the entire wing is well balanced. We’ll have to wait for an answer from George Seyfang to know for sure.
If the wing and the tail are both fixed relative to each other - does it mean that the trim tab has to provide the whole force to set the angle-of-attack of the combined wing+tail ?
Yes. This may seem like a tall ask, but if the trim tab (actually, flap) is given a large deflection, it can cause the tail to generate some significant side force. It can be deflected by up to 90 degees, in which case it would act like a gurney flap.
I expect we’ll all know soon enough. 😉 Putting the trim tab at high AoA would be draggy, no? We’re far beyond it’s max L/D. Then again, who knows what Richard was thinking—or what the design brief calls for.
Bill, you’re correct, the stalk couldn’t pivot at the wing trailing edge—at least not very far. If the pin allowed something like 5-15 degrees of “hinge,” and then hits a stop, what you’d have is a binary (on or off) actuator in the trim tab, with the against-its-stop tailplane to do the heavy lifting. Before you say it, this sounds unlikely when I look at it in the light of day, too. I don’t have another likely scenario, though. Mal’s often right (though not always!) and I agree a trim tab on a fixed tailplane’d be pretty cool, but I don’t see the horsepower to get the job done. It’s OK if I’m wrong though.
Dave Culp
Here’s another thought. It seems like the cam needs to be pretty rigidly located. it probably slides on plain bearing skids (UHMW strips?). But how do you get super rigid adjustment of the cam?
Perhaps this is a job for a non-backdrivable gear train. I know that a worm gear on a regular gear, like a guitar tuner, is non back drivable. Would a worm gear on a rack be? Then you have your locking for free, and beautifully precise adjustment of the throttle from a loop of line you can take wherever you want on the boat without cleating it—pull one side to throttle up and the other to throttle down. The loop would be linked to the worm gear by two oppositely wrapped furler drums, for lack of a better metaphor. Now all of the slop that can come from line stretch is gone.
No?
chris
There seems to be 2 assumptions here which are not quite correct.
Firstly, its important to note that there is not any great force on the cam, since the main wing is pivoted
at or very near its clr the force is nowhere near as great as the force of sheeting in an unbalanced sail. It will happily stay where it is just by the friction in the system if a cable and tube is used.
Secondly, the idea of the system is that you set the throttle and do not need to fiddle with it very much anyway, its automatic, you don’t have to follow wind-changes by hand. It’s just set and forget.
The nature of the cam is different than that which is used in (for example) a four-stroke engine, in the engine the cam has only one follower which is held against the cam by a spring, therefore wear and friction occur from the pressure of the spring. The cam in this case is unusual in that, apart from being perfectly circular, also has two followers one on each side, and therefore does not need a spring to keep the followers in contact. If the cam is perfectly circular and the two followers are perfectly parallel there need only be gentle “kissing” contact between the cam and the followers, and thus wear and friction can be almost non-existent. As mentioned before if designed correctly there is very little force on the cam anyway. This should make for a very long lasting and foolproof system. Accuracy in manufacture is very important to achieve this. Of course the only way you could have and “adjustable” cam in this way is if the cam is perfectly circular. I have not seen any other field where such and “adjustable” cam is used, would be happy to know if there are any other examples of this idea anywhere else.
Peter W
Secondly, the idea of the system is that you set the throttle and do not need to fiddle with it very much anyway, it’s automatic, you don’t have to follow wind-changes by hand. It’s just set and forget.
Peter W
I can see that the system will deal automatically with changes in wind direction but I don’t see how it deals with changes in wind strength. From what I can deduce about proa-people, a fair few of them will want to sail with the ama actually, or close to, flying. I guess it’s simply a case of throttling for the gusts rather than, as now, sheeting for them.
Here’s a question; is the assumption that the apparent wind is never aft of the beam? Because if not, I haven’t yet figured out the second half of the story—wind abeam to wind astern. Why does the angle decrease there?
chris
I think it may simply be that the cam is circular - as it has to be if the wing can swing through 360 degrees. So you get the same fall-off. Nearly the same fall-off in fact because of the asymmetry in the wing-angle v. tail-angle curve. The problem, if it is a problem, will be more difficult to design out for a proa which will be sailing backwards as much as forwards… :o)
And presumably the ability to quickly develop full-power backwards will make shunting a pretty snappy proposition.
Somebody mentioned “user-friendly” as one of Peter’s potential design goals. I won’t put words in his mouth, but perhaps he was thinking instead that when close-hauled, very little wing deflection is wanted for best VMG (hardwings typically have their max L/D at 4-6 degrees AoA), while at a beam reach apparent (broad reach, true), the wing wants to be operating at max Cl, which is closer to 22-25 degrees (for a 1-element hardwing)
Dave Culp
While a solid wing may have it’s best L/D at a low AoA, the lift coefficient at that AoA is also low, say about 0.3. What this means is tat you need about 4 times as much area for the same thrust as you do if operating close to a CL max. of say 1.2. So for the same thrust at L/D max. the wing could weigh up to 4 times as much and the centre of effort is much higher. So even though the wing efficiency is higher, the net boat performance may be significantly lower. I’m not saying that this renders the system inoperable, but one does need to be aware of the compromises.
Mal.
Given the concerns about heeling moment and the considerable ‘no-go’ volume swept out by wing + tail + counterweight, is there a good case to be made for adopting a ‘Harry’ type of proa with the wing(s) carried by a very slender vaka and the skipper and crew safely out of the way in the ama ? The resulting separation of crew and wings would allow the wing to be brought down much closer to the deck of the vaka (similar to the SailDrone). Attention to a wing at sea could be given by crew kneeling, sitting or lying on a tramp.
Efficiency of wings a low angles of attack.
My page http://www.sailwings.net/marktwo.html shows tests on the first single-axis wingsail system back in 2000.
I used a simple “pulley and string” system for automatic control, which had exactly the same effect as the later cam system, in that the wing angle AOA was max when wind 90 degrees to boat direction and zero when exactly head to wind. The only difference was that with the “pulley and string” does not allow 360 degrees rotation so you can see from the video that the rig was constrained and had to flip over in a gybe (jibe). The cam, which came later does allow 360 degrees rotation.
At first I thought that the diminishing angle of attack as the boat goes to windward might cause poor performance on these courses, but on testing found that the opposite was the case. The windward performance was very good indeed. In fact it was so good that even when head to wind the boat did not drift downwind at all and just “hovered” stationary. This behaviour is, I think due to the fact that the boat had extremely low drag and as can be seen, tiny variations in wind direction were enough to give forward thrust to counteract the wind. Maybe this is the only boat that has ever shown this degree of efficiency? A boat that cannot be blown downwind when head to wind? Has this ever been achieved before? Any amount of theoretical reasoning would not predict this result. So I just think it is always worthwhile to test things in real life rather than spend hours theorising. Not to say that theory doesn’t have its place but in this, the theory was put forward to explain the observable results rather than the other way round.
Concerning the effect of the cam system on downwind courses, yes, of course the angle of attack diminishes as you go more and more downwind. This is perhaps not perfect, but some compromise ise often required. Tacking downwind is an option, and this was, I believe what Walker advocated on his wingsail systems.
Peter Worsley
My preferred option would be a schooner rig with as much separation as possoble between the rigs. This allows plenty of room between the rigs for crew or a raised cabin. If the rigs can be kept well seperated they won’t interfere with each other too much, so the effective overall aspect ratio will be closer to that of an individual rig. The schooner rig then lowers the COE without too much loss of efficiency. I would go for a similar arrangement to the Saildrone rig, which has the boom and tail well above the deck.
Mal.
Tail control on Saildrone
George Seyfang, one of the design consultants on Saildrone, has told me that the very small trim tab was all that was required because they made the rig only marginally stable to avoid too much power consumption from the tail adjuster. In other words the rig could weathercock - but only just.
They did tests in a wind tunnel to verify if the small tab was sufficient, and apparently it was.
Peter Worsley
Thanks for researching that Peter, it is interesring information.
Mal.
Concerning the effect of the cam system on downwind courses, yes, of course the angle of attack diminishes as you go more and more downwind. This is perhaps not perfect, but some compromise ise often required. Tacking downwind is an option, and this was, I believe what Walker advocated on his wingsail systems.
Peter Worsley
I think if one looks at the details, this isn’t much of a handicap. It is rare for a powerful and slippery multihull to sail deeper than 90-100 degrees apparent. If the boat needs to sail deeper than this, then neither a wing—nor a main + solent jib are appropriate. This is spinnaker (or kite!) territory, pure and simple. That Peter’s mechanism delivers maximum deflection at 100 degrees apparent, as shown here by others, is actually near optimal for a hardwing.
Back to maximum deflection, a 2-element wingsail is no good at a deflected angle much above 30 degrees—the really highly deflected, super high lift wings one sees are 3 element, even 3 1/2 elements (parts are 3 elements, parts are 4). The 2-element wing is going to be—somewhat—better than the soft sail close to wind, and hugely better at any wind angle above about 25-30 degrees. But it will more or less suck at really high AoA (30+ degrees) just as an equivalent soft sail + jib will. An unstayed wiing with Peter’s mechanism isn’t limited by the wing’s deflection, however, since it can swing right around and hang off the front of the boat—still generating upwards of 2-2.5 lift coefficient. Plus it will never need to look at AoA anywhere near 90-100 degrees, so Peter’s controls are more than adequate.
Mal offered a great suggestion—that a non-linear response would be nice from Peter’s mechanism. Be nice if the first few degrees of tailplane deflection resulted in more than equivalent flap deflection—followed by many degrees of the opposite for mid-range AoA. I’ve seen this done with elliptic cams, but that’s not on for Peter’s setup—the round cam with dual followers and no spring tension is just too sweet to abandon. But how else to get the elliptical-style performance?
Dave
Mal offered a great suggestion—that a non-linear response would be nice from Peter’s mechanism. Be nice if the first few degrees of tailplane deflection resulted in more than equivalent flap deflection—followed by many degrees of the opposite for mid-range AoA. I’ve seen this done with elliptic cams, but that’s not on for Peter’s setup—the round cam with dual followers and no spring tension is just too sweet to abandon. But how else to get the elliptical-style performance?
Dave
Rather than drive the flap by a linkage from the tail - why not drive it from the main cam follower ?
If the flap were spring-biased to its ‘feathered condition’ then a rotary cam on the axis of, and driven by, the main cam follower could be used to actuate the flap in the desired non-linear manner. If no spring tension is an essential feature then a dual cam, akin to that in a Ducati Desmo motorcycle engine, could be used.
Peter H
Mal offered a great suggestion—that a non-linear response would be nice from Peter’s mechanism. Be nice if the first few degrees of tailplane deflection resulted in more than equivalent flap deflection—followed by many degrees of the opposite for mid-range AoA. I’ve seen this done with elliptic cams, but that’s not on for Peter’s setup—the round cam with dual followers and no spring tension is just too sweet to abandon. But how else to get the elliptical-style performance?
Dave
Would an over-centre mechanism offering only full flap (plus or minus) be useful ? It could be triggered by the cam follower as it swings from one side of the wing centre-line to the other in changing from forward to reverse.
Peter H
I think the varible throttle control is important, if not vital. I wouldn’t like to do away with it!
I’ve been thinking about an sdditional link which uses a pin and slider, which might give a more elliptical curve. I haven’t drawn it up yet so I’m not sure if it will work.