Progress so far
I bought the boat hull on eBay in June 2013. Since then I have been adding
supports for decking, fixings for the battery box and mast, and filling as
much as possible with two part waterproof foam. I've also rigged the sails
my wife made for the boat (thanks due to Mandy here).
On Sunday 15th June 2014, the boat made its maiden voyage at the Cambridge
Model Boat Club pond, on the outskirts of Cambridge. This was only a test of
the sailing ability of the boat, as the control was using a standard radio
control system, with me controlling the rudder from the bank. Overall this
was reasonably successful, but it did highlight that the rudder is too
small. Steering was not as responsive as needed, and it had difficulty
tacking, and gybing. Sailing along on a reach was OK though.
23 June 2014
I have now enlarged the rudder, but not tried it out yet...
26 Jan 2016
As I haven't updated this for a long time, I have forgotten a fair amount of
what I've been doing. So here is a somewhat incomplete history.
The larger rudder did help, but it was still difficult to gybe. Or tack, but
that isn't surprising as the boat is so slow. The mainsail is on a fixed
length sheet (rope), so the boom moves out to about 15 degrees from the boat
centre line, then is held on the sheet. This position was chosen to give
reasonable performance into wind. The boat would point pretty well (within
45degrees to wind) and sailing in light winds was OK.
However, due to the problem with gybing, and what I thought was too much
heel most of the time I decided to use a smaller mainsail. This is the same
height as the picture of the boat in the garden, but shorter along the boom.
This did improve the ability to gybe without apparently sacrificing any
(We're now probably up to january 2015)
About this time I became concerned about how low in the water the boat was
floating. The boat hull is not waterproof, and it is filled with expanded
polystyrene (EPS). The theory being that there is no way I could keep it dry
for many months, so I may as well bow to the inevitable and instead ensure
the boat floated even when wet. My first attempt (before the EPS) was the
two part foam as mentioned in the Boat tab. I gave up with this as I
couldn't reliably mix this to produce long life foam. Sometimes it worked
fine (and I have some samples from september 2014 that are still exactly the
same size as when poured), but other times is started collapsing within two
weeks. Instead I cut a sheet of EPS into many pieces with a hot-wire, to
fill the hull. Despite being as careful as I could be, there was still about
a litre of water in the hull. This was contributing about a kilogram of
weight (two pounds). I decided I wanted the boat to float higher, so I
replaced all the EPS, being even more careful and sanding the cut pieces to
fit better. This reduced the water somewhat, but only down to about 0.75kg,
so probably not worth the effort. I also replaced the battery for a
smaller capacity one (1.2Ah). This all helped, but the boat is still really
too heavy, and floats a bit low. This is also a problem because it doesn't
need to heel much before the deck starts to get submerged, which decreases
the righting moment, meaning it tips too easily. Without changing hull
though, I can't see what to do, so I'll stick with it.
I was also progressing with the navigation system, to the point where I
could try out sailing round a track on the pond. I wrote some program to
self-generate a route, based on the point the boat was at when the software
re-boots. Then I made the software reset if the boat was upside-down for 10
seconds, so simply by turning the boat upside-down on the bank, then
pointing it at the lake whist it booted up, it generated its own GPS course
to follow. This is only for testing. Trying it out by walking round the
lawn, it seemed to work.
So off the the pond to try out its self-navigation. (Up to now, most sailing
had been under radio control, or using the RC to generate a compass heading
to see how well the boat could follow that, which it appeared to OK.)
Navigation round the course worked pretty well, and the boat did actually
come back to the back! Success. I had to stop this sail when it got cloudy,
and as the battery had gone flat, the boat was in danger of shutting down in
the middle. Feeling pretty pleased that the boat had been autonomous, at
least for a few hundred metres anyway, I went home.
So far most of my sailing had been in lightish winds. Force 3 or less.
However, even in these light winds, I had noticed that gusts tended to knock
the boat completely over, so the mast was virtually in the water. There are
several issues at work here. One is that the boat is rather too heavy for
its size, and so floats very low in the water. This means that well before
45degrees the deck is already starting to go underwater. This greatly
reduces the righting moment, so the boat heels further. A second problem is
the fixed sheets. When a gust hits, the sail is basically side on to the
wind, so giving a very large heeling force.
I then checked the pilot charts I had previously found on the internet to
see what wind strengths were likely in the atlantic. It seems that Force 4
is about average, so having a boat that only sails in up to Force 2 wasn't
going to be much use! Something needed to be done.
What was needed was a self adjusting mainsheet.
At this time I finally managed to work out how to make a simple mechanism
that would keep a rigid aerofoil at a constant angle into the wind, using a
weather vane and a cam. (I'd been thinking about this for months.) I then
discovered that Ivor Bittle had a very similar system which he had patented,
and prototyped on a catamaran. At least he showed it worked. So I
constructed my system and replaced the sail with it. It was useless.
Basically it would barely sail on a reach, and couldn't make any upwind
progress at all. The problem seemed to be that the system just swung back
and forth around the mast so the wings rarely faced the right way. Rather
disappointed, I abandoned this scheme.
Springs seemed to be a possible method of self-adjusting the sails. I found
some suitable springs and stared putting them around the boom until it sort
of worked. It was more complicated than expected because the boom pivoted a
few cm away from the mast, and the outhaul caused forces in the opposite
direction to what I wanted. So I had to negate this force with on e spring,
then add the ones I wanted. It looked dreadful and I never sailed it.
The breakthrough was to buy a pLastic bearing large enough to go around the
mast and fix the boom to this. Now both the sail and boom rotated around the
mast, meaning outhall wasn't a problem. What looked like it should work,
(and I proved this using Excel) was a spring from the end of the boom to a
point slightly behind the mast. This gave pretty close to what I wanted,
which was for the mainsail to let out when it was in a strong wind, but
sheet in when in a light wind, or going more into wind. However, it meant
the boom jammed when in a strong wind going downwind as the spring hit the
mast. To overcome this I replace the fixed point behind the mast with a
large bearing, so the bearing was eccentrically mounted around the mast.
This gives the same effect as fixing the spring behind the mast, but allows
360 degree rotation. As long as there isn't a jib, so I removed this.
A trip to the pond with this single sail configuration showed the system
worked pretty well. Some experimentation with the spring was required, and
then it self adjusted pretty well, and coped much better with gusts. A trip
to the pond in stronger winds (probably up to force 4) showed much better
strong wind and gust behaviour. It was also much better downwind than the
fixed sheet as the sail could open up, giving much better downwind speed.
Close hauled was worse however, as there was virtually no force on the boom
when close to the centre line. I thought this was a price worth paying.
Another notable point from the pilot charts (which gives the percentage of
the time for each direction of the wind and its average strength) is that
the wind direction is actually very variable, and doesn't seem to blow from
even the prevailing direction for much of the time. So close hauling isn't a
total necessity. As long as you can avoid being blown downwind too much, you
can make progress when the wind is favourable.
The first version of the eccentric bearing system used a stainless metal
bearing. Looking up the metal actually used it became clear that the bearing
was going to corrode in salt water. I have a 3D printer that has hardly been
used, so decided to try and make a plastic bearing using this. I bought some
4mm diameter acetal balls, and printed innner and outer races. The cage
wasn't so successful, and I had to modify a simple cylinder with a soldering
iron. With a bit of fettling, and squeezing the outer race to a bit of an
oval, I managed to assemble it. This produced a pretty decent bearing once
the cage was pressed on. So I quickly bodged up a bit of brass to fix the
spring to (the other end is attached to the end of the boom), and went back
to the pond. Again, this worked pretty well, but it looked like more
eccentricity was needed as the sail sheeted out too far, too easily.
I then 3D printed a larger bearing using 6mm diameter glass balls (sold for
a BB gun), and fixed this up similarly.
A few days after Christmas a severe southerly storm blew up, so I took the
boat off to Grafham Water, about 20miles from Cambridge. The sail was making
a din flapping severely even in the car park. It was about force 6 or 7.
Much stronger than I had been out in before. Fortunately the car park was on
the downwind shore, so i figured the boat would get blown back if it didn't
work, or broke. The waves breaking on the shore meant my wellies weren't
high enough to launch off the beach without getting filled up, so I walked a
bit around to a sort of breakwater where I could throw the boat in to deep
enough water. After this unceromonious launch, I scrambled back up to get
the transmitter. The boat was spending most of its time with the mast
horizontal, but was making a little progress across the wind. (For a
reservoir, there was a decent swell, about one foot in the deep water which
then broke in the shallows.) Then I found that the rudder had minimal to no
effect. Whatever I did the boat slowly carried on across the wind. This was
a problem as its direction would have sent it off across the lake, to the
other side about a mile away. This would have taken a couple of hours, and
probably ended up on an inaccessible part of the bank. So I was very
relieved when I managed to get it to gybe and head back. It got washed up on
the beach, close enough to shore for a rescue without filling my wellies.
To make the most of this trip I adjusted the spring and had another go, with
pretty much the same result, only just managing to gybe and get the boat
back to shore. It looked again as if the rudder was too small (or
ineffectual for some other reason), or the sail was too big. Thinking I was
pushing my luck already, and not wanting to have to retrieve the boat from
the other side of the reservoir I went home.
After the problems at Grafham, I thought that perhaps an aerofoil instead of
a sail would be better. It should have a better lift to drag ratio (I
thought), so wouldn't heel as much in a strong wind, and get more speed.
These two things should both help steering.
I made an aerofoil based on NACA 0021, (which has a thickness 21% of its
length) from a sheet of plastic bent round some formers. Mounted vertically
on the mast, with a plastic bearing at the top and bottom allowed it to
pivot around the mast. This looked just like an aerofoil should and I
thought it should work pretty well. (The aerofoil is upright like an
aeroplane's rudder, not like the wings.) Using the same spring and eccentric
bearing system that had worked well on the sail, I set off to the pond.
Note that I had to change ponds after Christmas as the first one I used had
been taken over by a fishing club, who put a big fence all round. I
remembered a farmer had a small reservoir nearby, and went and asked if I
could use it. He was very obliging and allowed me to sail there. The
reservoir is actually much better than the larger pond as it is raised up
above the surrounding fields, and only has trees along one edge. This gives
a much better wind, with much less turbulence.
The new aerofoil sail didn't seem to be terribly efficient. Despite being in
a strong wind (at least force 4) the boat wasn't exactly zipping along. And
it was heeling at least as much as the sail. The heeling causes a couple of
problems, first the boat heads up into wind, making gybing very difficult,
or often impossible until a lull. Second, the boom and end of the 'foil drag
in the water, meaning that the 'foil catches the wind side-on, making it
heel more. (This effect will be familiar to anyone who has tried to learn to
sail a Laser in a reasonable wind. Or is it just me?)
Back at home I raised the 'foil up the mast, and shortened the boom, to give
more clearance when the boat heeled a bit.
At the pond the next day, I threw the boat in from the shallow water, and
was somewhat surprised and alarmed when the boat just capsized and lay
there. Fortunately the wind blew it back to shore, I lowered the 'foil and
had another go, but basically unless the boom was right down by the deck, it
capsized. This explained some of the heeling problem, as the 'foil is too
I bought some lighter plastic and lightened the internal ribs, managing to
make an aerofoil that was half the weight of the first one. My wife also
made a smaller sail: shorter in height, and boom length, as the mid size
sail was too prone to being blown over in stronger winds (ie anything above
about force 3). So I now had two options: a lightweight 'foil, and a storm
Another trip in about force 4 (on a freezing day) showed that the sail
performed better than the 'foil in every respect. With the sail, the boat
went faster, could gybe and tack at least as well, and the sail is much
lighter than the 'foil. So I'll probably abandon the 'foil and stick with
A long time passes...
In the intervening years, I gave up on the foil sail and concentrated on a
more standard sail. Did lots of software and trials at the Wakeboard Hub
between Cambridge and Ely. (Many thanks to the owners for letting me use
First Launch, 9th August 2017
The first launch was from Chesil Beach (near Weymouth on the south coast of
UK). Everything went well for a week or so when the boat stopped
transmitting Iridium messages. A few days later I was contacted by a group
on a fishing trip to the channel islands. They rescued the boat, and
returned it to Weymouth from where I collected it.
The problem was that the copper tracks on the back of the solar panels had
corroded through, and the battery had gone flat. I probably could have
tested this a bit more thoroughly!
After this attempt there was a major rebuild of everything that needed to be
waterproof (control box, battery and servo box, and solar panels). Because
this added a significant mass, I had to add a layer of polystyrene above the
original deck level to ensure the boat self-righted if turned turtle. I also
slightly lightened the keel as it was floating very low in the water.
Due to all the changes, I missed launching in 2018.
Second Launch, 1st July 2019
Again, launched from near Weymouth. However, within a day a fisherman
had rescued the boat and returned it to Torbay. He contacted me and I asked
him to put it back into the sea as I thought it should still be able to get
back on route from there. Unfortunately the rudder had become damaged and it
bobbed around just off the coast of Torquay for a week until rescued by a
ferry owner taking tourists back and forth between Brixham and Torquay. So
another trip down to Devon this time to collect it.
Third launch, 13th May 2020
After a delay caused by Covid lockdown, I took the boat down to Weymouth to
be launched off Chesil Beach. Perfect conditions with a gentle northerly
wind and sunshine, meaning no significant waves breaking on the beach. A
slight hiccup as the Chesil Beach cafe car park was closed, but fortunately
another car park almost opposite was open (this being used by local
kite-surfers). So a slightly longer walk to the sea than I'd hoped, but not
After checking everything still looked OK, I turned the boat upside-down to
restart the software. This causes a new GPS read to be performed, and sends
back an Iridium message, hence giving a feedback that everything is probably
OK. After a few minutes the email from Iridium arrived successfully, and it
was time for the launch.
Paddling into the sea, carrying the boat I waited for a gap between the very
small waves which were just about breaking, and pushed it out to sea. Off it
went looking like it knew where to go. Slowly at first as it was in the lee
of the beach, then a bit faster as it got further out and into slightly more
wind. And it just kept going, and going...
Every few hours a new Iridium message was received showing the boat was
heading as expected. However, this didn't last long as occasionally it
reported a completely wrong heading, meaning it was sailing back the way it
had come, or some other apparently random direction. This was disappointing
as hours of testing (in the garden) hadn't shown any such error.
I then thought I could send an override message to force the boat back to
shore and collect it. However, the override system didn't work, and sending
messages had no effect. I tried contacting some trawlers in the area, but
not surprisingly they weren't interested in stopping fishing to search for a
toy boat. So I just watched the progress it was making, which was a bit like
two steps forward, one step back, but overall for a month or so progress was
broadly in the right direction. Up until a Force 8 storm with 15m waves in
the bay of Biscay. At that point it had sailed about 500miles from the
launch point and was further west than any of mainland Europe, so not a bad
effort. But not a big proportion of the route across the Atlantic.
I then stopped paying for the Iridium service, as the boat wasn't going
anywhere useful, just being blown by wind, and current. I did restart the
Iridium service about a month later, and the boat was still obviously not
under control of its destiny. After a few weeks the battery voltage started
going down, and finally it stopped transmitting, the final message being:
This means 14 August 2008, last GPS reading at 14:52:43 (GMT). Position
45.8degrees N, 6.01degrees W. Other numbers are wind direction, desired
heading, battery voltage (1/10 volts), hours of running since last restart,
and temperature (centigrade). That was the last I thought I'd hear of it.
7 Jan 2021
was put up by the lighthouse keepers of the Cordouan
We found Phil's boat yesterday, buried in the sandbanks of
It participated in
a competition to design and build an autonomous sailboat less than 4
meters long, able to cross the Atlantic Ocean without human interaction.
We contacted its owner.
This was a huge surprise to me as I had expected the boat to have been
smashed to bits on the NW French coast. Looking at a map, it seems it got
blown to one of the safest places to be washed up, and rescued by the
French Coast Guard. Amazing.
There are also photos there, this one being the best:
So apart from the missing mast, it looks in remarkably good condition. I
can't tell if it still has a rudder, but without a mast a rudder isn't much