August 5, 2008

Snippets, 8/4/2008

July was a bit of a slow month as I was wrapping things up at Google, but we now have some key initiatives underway:

  • Community
    • Our conference is coming up on October 10th, a little over two months away, and we are going to start actively organizing it, recruiting attendees, etc.  If you are interested in attending and/or helping out, please email
    • Publicity – At least two pieces are coming up in major global media outlets, and my Bureaucrash interview is up.
  • Engineering – Consulting work has started and we’ve identified a potential design that meets our requirements.  In a few months we should have rough cost estimates and an understanding of its performance characteristics.  For patent-related reasons, we will probably not be able to release details for awhile, sadly.  (We will be obtaining patents for defensive purposes and licensing them to anyone who wants to use them).
  • Research – Nothing.
  • Administrative – We are working hard on our non-profit application to the IRS, which is important and fairly involved.


19 Comments on “Snippets, 8/4/2008

August 5, 2008 at 6:31 am

Is your potential design a spar?   Or do you not even want to say  the general type of design?

I ask because my spar model seems to lean in the wind and if you are using this general type I would be more inclined to investigate this.

I was also thinking of measuring how much force is needed to move my spar model through the water. Because of currents and wind, and needing deep water for my spar model, it has not been as easy to measure this as I had hoped.    If we want to do migration, the power needed to move a seastead through the water is an important thing to understand.  We could tell from measuring a model.  I have some ideas on how to measure it, but again I would be more inclined to bother if it was relevant.

  – Vince


Wayne Gramlich
August 5, 2008 at 7:43 am

My preferred design for small inexpensive seasteads is a mini truss-spar. This has a short stubby cylinder (say 20′ across by 50′ deep) with a truss that extends another 40′ – 50′, with ballast on the end. The cylinder is ballasted to about the mid-point 25′ above water and 25′ below water.

The closer the center of buoyancy is to the surface, the less wind loading there is, but the more wave interaction there is. The deeper the center of buoyancy is, the lower wave interaction there is, but the higher the wind loading is. The less the spar sticks up into the wind, the less the wind can push it over, but the easier it is for waves to cap the spar. It is all a bunch of trade-offs.

Note that the winds in the open ocean can reach “hurricane level” for scale models very easily. If a seastead tilts during a hurricane, but does not sink, that is probably a reasonable trade-off. A 20/30 degree tilt during a hurricane would be unpleasant, but not as unpleasant as being on a ship that gets sunk by said hurricane.

The goal of this design is to cost less per square foot than a new/used boat. If it is cheaper per square foot to purchase a new/used boat, that is what any prudent person would do. While it would be nice if the mini truss spar seastead is more comfortable than a new/used boat, it is not a requirement. If the mini truss-spar is safer, that would be nice too.

Wayne Gramlich
August 5, 2008 at 7:46 am

Vertical spar towing is apparently a solved problem. A vertical spar with smooth sides will apparently oscillate back and forth as it towed. To prevent this behavior, a spiral surface is added to the spar to break up the side to side forces. Do not ask me why it works because I do not know.

August 5, 2008 at 1:06 pm

>Vertical spar towing is apparently a solved problem.

I was not even thinking of oscillations, just the force needed to tow it.  So if we wanted to migrate at 1 MPH, how much force is needed?  My worry is that if we use diesel it will just cost too much to move.

I think the towing and leaning is more a problem for a 200+ foot long spar, and less an issue for the one you describe. I think a big kite would be reasonable for moving it along at 1 MPH. 

For the short spar I think the question is stability.  From earlier conversations you seem to think the truss will work much better than the hanging ballast that I tried with my short cylinder (paint bucket).   I don’t really think so, so I will probably try an experiment or two.




August 5, 2008 at 2:44 pm

>Note that the winds in the open ocean can reach "hurricane level" for scale models very easily.

Yes.  But my model just had a small cup on top, maybe 2 inches by 3 inches.   If there were a model of a house on there it would have been more like 1 foot by 1 foot.  This is a much bigger area for the wind to push on.   So slower winds and a much bigger surface area could still have a leaning problem.   Again, this was for the long skinny spar.

Wayne Gramlich
August 5, 2008 at 7:55 pm

The truss should have a significantly greater righting torque. I’ve worked through the math once before and posted it. I’m busy today, so I don’t have the time to do it again. I’ll try to draw some diagrams and post them later tonight along with the math again.

August 5, 2008 at 8:02 pm

 I don’t want to be too specific, but I will say that I don’t think a single spar is a good design for a large seastead, although I agree with Wayne that it’s good for a small one.

August 5, 2008 at 8:48 pm

As before, I agree that the maximum torque is greater with a truss but don’t think my models got to where it makes a difference.  To be clear, I am not arguing with your math and you don’t need to work it out again.   I think as long as you are in the range where the ropes are all in tension that pulling on ropes is the same as pulling on beams.   This means the weight is always under the cylinder.   But if you have a 2  foot truss and a 1 foot wide cylinder, you have to tip a lot before you get past the edge and the truss wins.  

I will at least shoot some underwater video and see if the ropes go slack at all.  


Wayne Gramlich
August 6, 2008 at 8:12 am

The truss wins because it gets to multiply by the length of the truss as opposed to the cable system where it is the radius of the cylinder. Let me try the math again:

Let L be the distance from the center of buoyancy to the ballast mass M. Let r be the radius of the cylinder. For the truss, the righting torque is:

T(a) = M × g × L × sin(a)

For the wire ballast, there is the critical angle c, which is when the mass is directly under the cylinder edge. I am assuming that the cable is attached to the cylinder horizontal to the center of buoyancy. For this situation:

L = sqrt(L2 + r 2) × cos(c)


r = sqrt(L2 + r2) × sin(c)

The second equation can be rewritten as:

sin(c) = L / sqrt(L2 + r2)

which will be reused below.

For the cable guy system, the torque for angles between 0 and c is:

T(a) = (a/c) × M × g × r

For angles greater than C, the maximum torque is:

Tmax = M × g × r

At the same critical angle for the truss, using the equation for sin(c) above:

T (c) = M × g × L &timps; sin(c)
= M × g × L × (L / sqrt(L2 + r 2))

If L = 4r (what you used in your example):

Tmax = M × g × (4r) x (4r / sqrt(4r2 + r2))
= M × g × 4r × (4r / sqrt(16r2 + r2))
= M × g × 4r × (4r / sqrt(17r2))
= M × g × 4r × (4/sqrt(17))
= M × g × 4r × .97
~= M × g × 4r

Thus, the truss has 4 times the maximum torque of the cable system when L = 4r. After the cable system exceeds a tilt of c, no additional torque is applied to the system. For the truss system, the longer L is, the greater the torque multiplier. For the guy wire system, longer L, reduces the critical angle c.

Honestly, by the math, the truss system completely beats the guy wire system.

August 6, 2008 at 8:53 am

The ropes don’t go slack:

The ropes are always in tension, so replacing them with rods that were always in tension would not change things in this test.   You can even imagine that the straight lines in the video are rods and nothing appears to violate the laws of physics.

Another interesting way to make waves for models came up today.   If a powerboat goes by the wake can be some good waves for testing a 1:25 scale model, depending on the boat, speed, and distance from the model.  Using a powerboat to generate test waves could be much cheaper than wavetank rental, so it might be a useful trick to remember.

I put the tension circle in the water again and it took todays waves like they were nothing.  This short cylinder has a lot of motion and the central post on the tension circle is very stable.  Since we can buy 4 foot diameter  steel pipes and weld them together, I think the tension circle would not cost much more money and is much more stable.

Looking at the two videos it just does not feel like the two models are in the same waves, so I think I will shoot another video with these two models side by side in the water taking the exact same waves at the same time.

The weights came near the bottom and probably at least touched the sea grass.  Might have damped out some oscillations but even if it did hit, the strings never went slack.



August 6, 2008 at 8:55 am

Before we get to the criticle angle "c" for the hanging ballast, so angles between 0 and c, I think the truss and the cables have the same torque.  Do you agree?   If not, do you think replacing the straight part of the ropes in my latest video with rods would make any difference?  If so, why?  Nothing would push on the rods and a pull on a rod does the same thing as a pull on a rope.

My claim is that in my experiments we never go past "c" where any string goes slack.  The fact that a truss can do more torque after this angle never comes into play in my experiments.

>The truss wins because it gets to multiply by the length of the truss as opposed to the cable system where it is the radius of the cylinder.

That is only true when we are at 90 degrees.   When we are at small angles I think things are the same.

Really have to take into account the dynamic motion of the weights and the cylinder, because it is as if "down" for the hanging ballast is not really always toward the center of the Earth.

August 6, 2008 at 5:44 pm

>Thus, the truss has 4 times the maximum torque of the cable system when L = 4r.

I agree that with L four times as long as r  that the maximum torque for the truss is 4 times as much as for the cables.  But this would happen when the model was flat on its side with the ballast/truss coming out of the water.  The reason it can get 4 times the torque is that it can get the ballast 4 times as far to the side.

My claim is that in my experiments I never get the ballast over to the side so much, so the maximum is not relevant.   As long as the angle is less than c, then both cables and truss move the ballast just as far to the side, so the torque must be the same.  The angle c is where one of the cables would go slack, and in my experiments it never goes that far.    So for my experiments the cables and the truss would do just as well (which is not very good).

August 7, 2008 at 1:12 am

I made a page with a bunch of videos from my "Short Cylinder" experiments.  I think my best result was with a second bucket underwater.  This held the 20 lbs and provided drag against movement in the water (up/down/sideways) and the extra inertia of the water (no lid but still gets inertia).


Wayne Gramlich
August 7, 2008 at 4:33 am

I agree, the double bucket experiment looked pretty good. Assume a 1:50 scale, those waves were pretty substantial – 35′-45′. I suspect that people would be a bit uncomfortable, but in no danger of dying if that model were scaled up by a factor of 50.

For a truss system, I would add a heave plate to reduce up/down motion. I have no idea what a heave plate would do to a cable system.

I remain pretty convinced that a short cylinder with attached ballast is going to cheaper than a boat and probably more comfortable in heavy seas. It looks safer to boot.

(I appreciate the patience it takes to just sit there in the water for 10 minutes and just shoot video.)

August 7, 2008 at 5:30 am

The double bucket had about 4 feet of rope between the buckets.  Each bucket is about 1.5 feet long, so the total length for the system is about 7 feet.  For sure the bucket reduces heave (both because of increased drag and increased inertia).  If you ever had heave plates big enough that the ballast could not drop fast enough to keep the ropes from going slack, there could be trouble.  I think that is much bigger than the area of the bucket.

After watching the underwater video of the hanging ballast and thinking about it more, I now think that longer ropes can provide better motion for the hanging ballast models.  The longer the rope, the less the ballast will actually move sideways and the more the pull is still vertical even if the model moves sideways some during a wave. 

The longer a pendulum is the longer the period of oscillation.  Sometimes these models got swinging back and forth a bit.  Longer ropes should make longer period oscillations for a hanging ballast. 

So I am thinking it is probably worth trying this again with the ballast like 20 feet below.   I suspect it will be  even more stable than the double bucket results.

I think of these as 1:25 scale models.  If you scale by 50 then this is around 75 feet high and nearly 50 feet in diameter.   That would not be a single family dwelling.  But yes, in normal seas that would be stable.

Note that since I slow by a factor of 4, a 10 minute video on youtube only takes me 2.5 minutes to shoot.  Though getting setup and cleaned up add many times that.

Wayne Gramlich
August 7, 2008 at 6:23 am

I couldn’t tell how bit the bucket is (no reference object). I think 1:25 is a more accurate scale.

August 7, 2008 at 1:59 pm

So I can slide some 2 foot sections of PVC pipe (have some 3/4 inch around) over the 4 ropes.  So I think an experiment to test out  rods vs ropes is not very hard and so will do it.   I also think there could have been a small amount of stretch in the ropes, or they might not have all been exactly the right length, so this might just help.  Also, the extra sideways drag of the pipe vs just a very small rope could make it work better. 

I will hang the ballast off this and try different lengths to see how it affects things (like to 20 feet down).


August 7, 2008 at 4:25 pm


>I couldn’t tell how bit the bucket is (no reference object).

It is a 5 gallon paint bucket and I just measured one.  The bucket is 10.5 inches at the bottom, 11 inches diameter at the top, and the lid is 12 inches across.  The bucket is 14 inches high.  

Scaling 11 inch diameter by 14 inches at 1:25 scale factor gets to 23 feet diameter and 29 feet high.  If the inside was 22 feet across then each floor would be 11*11*pi=380 sq-feet.  If there are 3 floors inside that is 1140 sq-feet.    This is really huge compared to the average live-aboard boat and even more space than some multi-million dollar yachts.

It would need to scale by 43 to get to the 50 foot high size you had talked about, so your scaling by 50 was not far off.  This would make it into the range of a nice house.  Scaling by 43 also means the model was doing alright in some big waves.  Also, the video should be slowed down by 6.5 and it is only slowed down by 4, so really the motion would be slower than in the video.   So I think you are right that the motion of a 50 foot high cylinder could be OK.

Does anyone know some software for slowing a video down by something other than a factor of 2?




August 8, 2008 at 7:49 pm

Thank you, Patri, for this update.  It’s good to hear that we are getting more media coverage, and that the basics for this plan is progressing nicely.

If there is anything I can do to help here, please let me know.  Thanks.

Comments are closed.