Auto-Flight
Proas are uber-cool. They offer the most bang for the buck, they have least material stress of all multihulls, they have a super-cool backstory—proas have been the fastest sailing craft on Earth for 500 years. Who wouldn’t want one?
Some of the first Europeans to witness proas saw a bit of the “secret sauce” that makes these boats so fast and so fun—“flying” proas. Let’s face it, flying a hull is a blast. The wetted surface drag is reduced by a third. The wracking stresses and wave drag associated with two hulls trying to react to out of sync waves; it all drops away in that surreal few moments—dare we dream minutes?—of flying. AOXOMOXOA dude!
The risk remains though. Flying a hull is flying without a safety net. There’s only a tiny addition in sailforce between flying and capsizing, and for most of us, that’s simply a leap too far. Flying a hull is a little like lap dancing; it looks exciting, it’s something you’d maybe like to experience one day…… but not today.
What if we could automatically limit the flying hull’s altitude to some set height above the water, and never higher—or lower? What if we had a magic box that monitored hull flight all the time, yet let’s us actively sail the boat, when, where and as hard as we like. Our box would automatically ease the main sheet when we fly too close to the edge, yet follow our lead, bringing that sheet right back in as the hull comes down. This box would be Ginger Rogers to our Fred Astaire; we guide and she makes us look fabulous. Slow, slow, quick quick quick.
Yeah, we’re talking about computer control, actively measuring ride height then reacting in split seconds to control sideforce from the rig to sustain and control flight. We’ll need wave followers, we’ll need sensors and CPUs. We’ll need actuators and power supplies. We’ll want both kinds of RAM—computer memory and hydraulics. The system will be expensive to prototype and always susceptible to salt water shorting, but boy oh boy, won’t it be fun?
OTOH, do we really need all this stuff?
Could there perhaps be a purely mechanical “computer” capable of getting this job done? Most autopilots today are electronic computing works of genius, but once upon a time—and for hundreds of years—sheets were lead to tillers and the feedback loop/programmer’s wishes couple which defines “computer” steered our boats across the world’s oceans. Can we maybe pull this off for flying the ama? Go here for the rest of the story
Yeah, we’re talking about computer control, actively measuring ride height then reacting in split seconds to control sideforce from the rig to sustain and control flight. We’ll need wave followers, we’ll need sensors and CPUs. We’ll need actuators and power supplies. We’ll want both kinds of RAM—computer memory and hydraulics. The system will be expensive to prototype and always susceptible to salt water shorting, but boy oh boy, won’t it be fun?
OTOH, do we really need all this stuff?
To be Devils advocate for a moment, computer control is in the middle of a major transition that’s being integrated in more and more capable R/C actuators. The expense can easily be minimal and a challenge to the right computer crowd can produce some mind blowing solutions. Check some of the Arduino stuff going on.
Still salt water is salt water.
That being said, mechanical is where it’s at. Once (if) my wingsail stuff is sorted out, suspect there’ll be some trials based on your insight.
Thanks,
Skip
You could just have a skimming / planing / dragging wand linked to a mainsheet….similar to moth and laser foil controls.
Or as in my proa, if you are looking to have a transversally pivoting centre board foil, just link the mainsheet to the foil. The lateral force pulling the board and boat out to windward is balanced by mainsheet tension…..as the ama lifts, sideforce reduces, mainsheet eases and the foil pivots towards neutral lift position or beyond into downforce territory until there is enough foil back in the water to pull it back out to weather again, bringing the mainsheet in with it. If the ama suddenly lifts and pops the foil, no sideforce, so the mainsheet dumps. Timing and quantum are pretty well connected and instantaneous…....
I was planning to use the pivot control as a form of “mainsheet"anyhow and one of the problems was how you could “dump” the foil automatically….
Lots to think about ..........I will post a solution if I can make it work.
Thanks Dave.
Before I read Skip’s reply, my first thought was that this is an easy system to set up for a tail controlled wingsail.
One of my models had a similar system, except that the rig was a lifting rig and the boat heeled to windward. It used a sensor paddle to adjust the AOA of the kite sail. http://www.users.on.net/~malcolmandjane/Vproa/Vproa4A.html
Mal.
Dave,
Great idea, I don’t see why one could not make it work!!
The tricky part in my view is twofold:
- do you have enough power/torque/stroke, as you rightfully noted. And I would think you would.
- do you have the feedback system properly sized (for isntance, you do not want a seesaw effect), and more importantly, does it need to be adjusted for different conditions (windforce, speed, closehauled vs. broadreach, etc)? Only trial and error would tell.
I looooong time ago, I used to co-owned a Muscadet, design by a French Naval Architecte, Philippe Harlé. See below.
http://1001boats.blogspot.sg/2011/11/muscadet-french-peoples-boat.html
It was really as described the French version of the “VW Beetle of the seas”.
On it, I had a “Regulateur d’allure Navik”. See video here, especially at 2:35 minutes and 4:15 minutes.
http://www.youtube.com/watch?v=P43SaUUa_tY
That thing worked like a charm.
But as you can see the mechanism is not utterly simple….
You had a wind vane, pivoting on an almost horizontal axis that you would orient into the apparent wind. If the boat changes course, the wind push the windvane on one side, which with a mechanism of pushrods, acted on the fletner of the immersed blade. That fletner would rotate the main spade over a vertical axis, which would generate a significant side force, and therefore swing sideways over its main horizontal axis of rotation. That would turn a quadrant with 2 lines attached to the sides of the tiller.
Pfffff!
As you can imagine, the length of each “lever arm”; and the push rods, the size of the air vane, the counterweight on the airvane, etc, had to be sized right for the system to work. And it was sized right, and it worked. But I wonder how much trial and error had been through the design process. (in very light air, or when going downwind, we used to attach clothe lines clips at the top of the airvane to make the system more sensitive to the lightest wind changes; in those conditions the counterweight was too heavy).
Granted, you system can be much simpler; you do not care about the whole airvane, fletner mechanism. and so on, but you will still have to play with many parameters, some by design and some during operations; size of the blade, angle of attack, length of arm below the pivot point, length of arm to the quadrant above the pivot point, attachement position of the sheet on the boom, etc.
Ideally, you do not want to end up with something like: “I have 4 different positions for the quadrant to adjust the lever arm, depending on wind speed”.
It would be a fun project to work on.
It’s got to work.
Cheers,
Laurent
Yes I think it is playable!
If one approaches the principle of international moth, we can get the same result.
You could just have a skimming / planing / dragging wand linked to a mainsheet….similar to moth and laser foil controls.
That is precisely what this device is, Rob. I began with your version, then replaced the inefficient drag-powered wand with a more efficient lift-powered one and rotated the pivoting axis 90 degrees. This change much increases the power of the system because it uses lift rather than drag. Might note that the lift generated by the gadget compliments the main daggerboard lift, so again is efficient.
Or as in my proa, if you are looking to have a transversally pivoting centre board foil, just link the mainsheet to the foil. The lateral force pulling the board and boat out to windward is balanced by mainsheet tension…..as the ama lifts, sideforce reduces, mainsheet eases and the foil pivots towards neutral lift position or beyond into downforce territory until there is enough foil back in the water to pull it back out to weather again, bringing the mainsheet in with it. If the ama suddenly lifts and pops the foil, no sideforce, so the mainsheet dumps. Timing and quantum are pretty well connected and instantaneous…....
See later in the article, under “Alternative Arrangements” where I talk about just this alternative. Also another even more powerful—and simpler. 😉
I was planning to use the pivot control as a form of “mainsheet"anyhow and one of the problems was how you could “dump” the foil automatically….
Please note that the gadget includes a backup system as well; the mainsheet is double-ended; the skipper controls one end and the gadget the other. With the gadget fully “on” the skipper has complete control of whether or not it is actually controlling the flight, how high it flies and how much power is provided (Laurent’s veneer control). Also the pilot can dump/shut off the system in a split second, at any time. and in a fail-safe direction.
Dave.
To be Devils advocate for a moment, computer control is in the middle of a major transition that’s being integrated in more and more capable R/C actuators. The expense can easily be minimal and a challenge to the right computer crowd can produce some mind blowing solutions. Check some of the Arduino stuff going on.
You’ll still need sensors and actuators—and a power source. Even at the simplicity of Arduinos, the cost and complexity of a computer-controlled gadget will be several orders of magnitude higher than this one.
Computer control has a major advantage over us in that we can use an accelerometer to anticipate control input sooner. By sensing only position with the simple gadget, we can infer velocity (angular velocity) which is a second order function, but we can neither directly measure nor infer acceleration—a third order function—accurately. Modern all-electronic autopilots use accelerometers (often offered as an add-on) in addition to both position and velocity sensors, and it is spooky how they initiate rudder movement even before the boat begins to yaw. OTOH, accelerometer-less autopilots are far more common (cost issues) and work perfectly well with a much more demanding design brief than the autoflight gadget.
OTOH (three hands?) the gadget literally could not be simpler (I tried!) It is designed to be built, should one wish, of nothing more than a salvaged oar or paddle carried within a simple pipe sleeve, pivotally mounted on one aka. The entire device can be built in a single afternoon with $10 worth of hardware store stuff and a bit of rope. The device can be added, simply and cheaply, to any existing boat. Though it will work, it’s probably not for monohulls and will not live peacefully with hydrofoil bats, but for anything from a Coleman canoe & PVC pipe ama, to Hobbie cats, to 100’ ocean racing trimarans, the cost of entry is very near zero, labor included 😉
It is my strong desire that this device be built by as many people as who are willing, put on any handy multihull boat (including models, Hobbiensteins, or what-have-you, and tested. It it’s hopelessly simple-minded it will be obvious. If it’s effective, it will be as apparent. I am frankly a bit shocked that these don’t already exist; the potential gains for offshore racing—the gadget doesn’t require a proa, it will work fine on cat or tri—are very significant.
Before I read Skip’s reply, my first thought was that this is an easy system to set up for a tail controlled wingsail
Absolutely. The gadget isn’t novel in any sense and will work for many boats; the principals are widely used in autopilot control as Laurent points out, also hydrofoiil flight altitude systems, as Ricco and Mal point out. The novelty is in the simplicity; this isn’t “naive” simplicity but rather (I hope to demonstrate!) “more profound” simplicity, per Albert Schweitzer. Please see this thread for more on the philosophy of “enlightened” simplicity in design.
I’ll comment on Laurent’s concerns later today. Many share them (not me!) and the solution to—or willful ignorance of—these may allow further simplification of systems from ama flying to autopilots to hydrofoil flight. Meaty stuff. Please add your comments, pro or con. Thanks!
See later in the article, under “Alternative Arrangements” where I talk about just this alternative. Also another even more powerful—and simpler. ....
....Please note that the gadget includes a backup system as well; the mainsheet is double-ended; the skipper controls one end and the gadget the other. With the gadget fully “on” the skipper has complete control of whether or not it is actually controlling the flight, how high it flies and how much power is provided (Laurent’s veneer control). Also the pilot can dump/shut off the system in a split second, at any time. and in a fail-safe direction..
Noted…. Obviously I missed the bottom of the read more article. My eyes were probably glazed over at the thought of complicated electronics…...Not my bag and always a problem in a small boat on the sea…and my mind occupied with my own “Eureka” moment….
I mentioned moth / laser wands because they are pretty common proven technology for use even if only for starters.
Thanks again…Hopefully it will solve one of my big outstanding conundrums…......
Rob
Inserted hi-rez versions of Michael’s amazing graphics here and there above—go find ‘em! —And click on them to see the full size versions.
Dave
Wow! Absolutely amazing! I think this device is an absolute work of genius—and there is no way in hell that my boat will go without one! 😊
Brilliant work Dave, keep it up!
Marco
P.S. - Why would I need one board for each direction? Couldn’t I use a bidirectional foil? Obviously if the foil ventilates then the ama will crash, but if I’m willing to accept that increased risk with a bidirectional board, then the position of the board could just as well be at the midplane right?
On another note, with your system the daggerboard has to go in the vaka somehow… With a concept like ‘Delirium’ that isn’t an issue at all, with a Russell Brown style proa like my concept for ‘Firstborne’ on the other hand, it’d be a bit trickier.
I’m thinking that something along the lines of the rudderboards which Cheers had could be the solution: the rudder only extends along the upper half of the rudderboard. The rear board is fully down, in use as a rudder, and the front board is only down half way for instance, so the control surface of the front board, is still inside the trunk. You’d only need the two rudderboards then, but what I don’t like about solution is that you would reduce the total available steering moment, and you’d be decreasing the average aspect ratio of the boards which are producing your sideforce, particularly if the alternative would have been a high aspect ratio daggerboard…
Cheers,
Marco
Dave,
Love the ellipsoid ama, but can’t help wondering, why not finish the job off???...... If you turned it int a (fat) inverted aerofoil with a half ellipsoid leading edge, a flat topside and an arched underside, cant the ama at an appropriate angle to get the best AoA at (say) 10 degrees heel, then you have an ama which becomes another auto self righting mechanism when you get enough air under it…..J.S. Taylor mode. You also have a lot more reserve buoyancy for being caught aback.
If both side rudders were used all the time (pivoting through 180 degrees during a shunt) and their gantries were strong enough, you could Laser T foil and wand them (so that the wands swing around with the rudders), and you then you would have a fully flying proa using proven existing technology…...Tripod support using 2 rudder foils and ama / centreboard foil should be a pretty stable platform.
Rob
P.S. - Why would I need one board for each direction? Couldn’t I use a bidirectional foil? Obviously if the foil ventilates then the ama will crash, but if I’m willing to accept that increased risk with a bidirectional board, then the position of the board could just as well be at the midplane right?
Looking at the AutoFlight renderings, I suspect that the sheet control foil shown is too small to do the job. I think that you would need to use the full extent of the lateral resistance area to derive enough force to get enough sheet travel. I agree that a single bi-directional board at midships might be the simplest solution.
Mal.