Possible Hardware for Future Model

I am thinking of having the next size model be about twice this one (so about 8 times the displacement).   If it is 8 feet long it could hold 2 of the 100 watt panels by the same people that make the 50 watt panel used in the first size or 4 of the 50 watt.  Either way the total for solar would be about 8 lbs.

I would like to have 4 motors, one on each float, so that if one breaks we can keep going.    With 200 watts it would even be able to pull a kayak.  This means testing can be riding in a kayak that is pulled along by the robot.  Should be fun.

In this larger model I would put the batteries down inside the floats.  This lowers the center of gravity and also makes for a high rotational inertia as the weight is then widely spaced.  With 8 times the weight capacity we could use more and larger batteries and so make it through the night.

The first version just uses relays so the motor is either on or off.   With an electronic speed controller (ESC) we could use whatever power level we wanted.  There is a board called IOIO which can talk to an Android phone and an ESC.   There are other boards to go from bluetooth to ESC.   There are some ESCs that can talk bluetooth.

There is an underwater T100 thruster by Blue Robotics who is also making an autonomous solar robot boat.   You can get these with ESC built into the thruster.   These can go up to 140 watts but at about 12 watts it seems to give about 1 lbs thrust.  So 4 of these could be very nice.    We would normally run all 4 thrusters at low power but if one goes out then we run the other one on that side at higher power.

The BlueESC is the T100 model with the builtin ESC.  This also has the ability to be controlled with I2C, which is a protocol where one controller talks over  a bus to several devices.   With this a Bluetooth to I2C board could be an easy way to go.  However, a bus is not so good as far as keeping independent failure modes.  Saltwater getting into the wires by one thruster could disable all the thrusters.

For good independent failure modes I like having 4 different 50 watt solar panels going to the 4 different floats.  Inside the foam for each float there could be a charge controller, battery, bluetooth relay, and motor.   So the phone would be talking bluetooth seperately to each corner of the boat.   Could have a diode protected power line from each of the 4 batteries going to the power adaptor for the phone.   The diode would make sure that a short in one battery could not drain power from the others.

Could even have a backup phone that only took over if the first phone stopped working.

There are phones starting to come out with Ubuntu on them.  This Ubuntu phone has compass and GPS for $192 US.   This would let us write fast native code.    It could be interesting for a future model. 

I think having a 3D printed part to connect the drive shaft and the motor and help mount the motor would be good.   With shapeways anyone could just get one printed even if they did not have their own printer. 

Voltage Sensor

I would like to be able to have the robot go on multi-day trips.   But for this it really needs to know how much power it has.  I would like a small bluetooth voltage sensor.  Seeed used to make one for $39 but it seems it is discontinued.  There is a $120 Mooshimeter by moosh.im that could work.

Field Testing Navigation

We adjusted for magnetic north, and it now points very well. We had to use two phones together.   The phone with a compass does not have GPS and the phone with GPS does not have a compass. We followed the arrows as we went around the field.

Autopilot Demo

In the video below I demo the manual mode and switching to autopilot.
The arrows point the direction we want to go.
If were pointed about the right direction both propellers spin.
If were pointed too far left or right then only one propeller will spin.

Testing the motor and propeller

We hooked our 12 volt battery to the motor and propeller.

We then put it in a 5 gallon bucket of water, and it had enough power to get the both of us wet.
The shaft for the propeller was too wobbly, it needs something to stabilize it.

Future Data Options

Our current phone can use data roaming near many of the islands in the Caribbean.  So we could send the boat to other islands where it will still be able to communicate with us.   At some point we will do this and have it send back status updates and pictures of the island.

Data Roaming is rather expensive with our regular phone service so this Skyroam service is interesting.  It gives you lots of data for only $10/day.  However, it does not seem to cover the Caribbean yet.

Longer term we want to have some sort of satellite link so we could send the boat even further than the Caribbean and still keep in contact with it.

Elon Musk and Richard Branson are both working on cheap satellite Internet services.  However, these seem to be 5+ years out.

There is an Iridium Go that is a satellite link that makes a wifi hotspot  Today the data rate is slow but Iridium will be upgrading to Iridium Next soon and this device will work with the new faster service.   With this our Android phone could send us data while crossing the Atlantic without us having to change any of our code.   The slow data rate would even be ok, we just would not be sending many pictures.    The Iridium Go device gets power from a USB port so that is easy for us.    At the moment this is our first choice for communications when we are ready to send the boat far from land.   It seems Iridium next has been delayed till 2017+.

 Globalstar has data service for $40/month for 9000 seconds and there is a Globalstar hotspot device for $158.   Not sure how much data this really is but something made it seem like 9000 bytes.

Orbcomm has a data service and devices.   Seems to be flat monthly rate but can't find products on Amazon or other easy retail location so far.

Some Spot Connect but seems only one way and costly.

Quadmaran vs Sailboat

In the videos below both the quadmaran and the sailboat are each about 10 lbs.  In real boats the cost is roughly proportional to the weight of the boat.  So in some sense this is a fair comparison.  However, the sailboat gets stability from moving through the water and from the wind against the sail and in this test it does not get those. 

We are interested in slow moving solar powered boats and just want to show that the sailboat shape is not ideal design for our purpose.  The 4 floats on the quadmaran are widely spaced and the sailboat is much smaller with much of the floatation in the center.    So the sailboat rocks much more in this slow no wind against the sail test.   Note that the quadmaran has a lot more area for solar panels as well.

These videos are slowed down by a factor of 4 which would be right if these are models of boats 16 times larger.  So imagine the 4 foot quadmaran model is 64 feet long and 32 feet wide.

The wind for this scale is hurricane force, the waves on this scale are large for the Caribbean.    

Testing chase kayak

Teryn with paddle and Corey and Amoni in kayak

We will use the kayak to chase the robot boat around in case it has any trouble.  We had not used it in awhile so wanted to check that it was still OK.   This is Sandy Hill Bay and just 5 minutes from our house.  There are calmer bays so this will probably not be our first test location but it will probably be our most common test location.

Teryn and Ethan

Big enough bay for some good robot boat testing

Awesome Science!

We want to use a magnet and an electromagnet to connect two boats together.
we still don't have a phone with a compass so we made one but please tell us if you know of a cheap android phone with a compass, Or better yet does anyone have a old phone with a compass they could donate. To test if it has a compass install a free compass app.


we need to know if it is safe to have a magnet near our phone compass. so we built several compasses to test how far away a magnet has to be so the compass still detects the Earth and not the magnet. We made a compass using a magnet and string, and with a magnet floating on water. Our tests say that for the magnets we are using if we are more than one foot away it dose not affect the compass. Since our modal will be 4 feet long we should have no problem.

App Inventor for Big Projects

We are starting to get enough code that finding a block you need to change is painful in App Inventor.   We looked at maybe switching to Android Studio but my boys think that does not look nearly as friendly as App Inventor and they are probably right.  

Teryn was doing the autopilot and Ethan doing the manual mode and bluetooth code and when it came time to combine them we had to create all the blocks from one project inside the other project by hand (still not finished).  This was/is sort of annoying.

Another issue was Teryn had a full screen of user interface for the autopilot and Ethan had a full screen of user interface for the manual mode and I thought we could just have two screen in App Inventor.  We want both the autopilot and manual mode to call the same routines for talking to the bluetooth relay so that you can do a manual check of the hardware before turning on autopilot and be sure the connection to the relay and low level routines are all ok.  But different screens in App Inventor can not share blocks!   This was unexpected and seemed like a big problem.  The solution was to have two screens worth of user interface in one screen but only setting one to "visible" at a time.  This has solved what seemed like a big enough problem to think about leaving App Inventor. 

There was a project to convert App Inventor to Java but it is no longer active.

There is a system to convert between App Inventor and a text language that seems interesting.


It would be nice if App Inventor had the ability to make a bunch of different workspaces and we could name these.  Then a pulldown menu that let you go to whichever one you wanted.  Then we could keep a large project organized by putting blocks for different parts in different workspaces.  To App Inventor this can be viewed as just one large collection of blocks, so I don't think it would be that hard for it to do.  With this we could easily handle the complexity of our project.   App Inventor is open source so theoretically I could add this or pay someone to add it.


I can use Java (though have not used Android Studio) and like it, so we might switch to Android Studio at some point but for now we are staying with App Inventor.  We are expecting a new android phone with a compass to get here Tuesday and will probably have the software ready by then.

Update Feb 14, 2015

As we used App Inventor more and more it got easier even as our code grew.   One trick for finding things is to do control and "f" then enter what you are trying to find and click "highlight all".  Then do control and "-" to zoom out.   Once you find it then control and "+" to zoom in.   The first couple weeks we did not even notice that you can do local variables, which do help to reduce the complexity.   Also, after you do "organize by category" a bunch of times you get a feel for where it places the different parts of the code.   We have all the main parts of the code working now and the size of the code is just not an issue any more.   We are very happy with App Inventor and will stick with it.

Blu Advance phone does NOT have compass

Several sourced claimed that the Blu Advance phone has a compass but this does not seem to be true.    Teryn has a navigation app but it really needs a compass.   Have a used phone coming from ebay should be good.

Testing Water Brake Version 0.2

We have decided to have "Wet Wednesdays" (the day after "Taco Tuesdays") and go to the beach every Wednesday.   So testing of hardware in the water will probably be on Wednesdays.

  Today we tested a new design for a water brake and it works really well.

Cut a section out of a 5 gallon paint bucket and added some hinges so that if the boat goes backward it goes down into the water and slows the boat alot but if it is moving forward it mostly rotates out of the water and adds only a little bit of drag.

Going forward water brake mostly up

Put some red foam cushion any shock

Using some rubber bands it seems about 1/2 lb of force will make the boat go forward  at about 1.5 MPH.  With about 1 lb force against the water brake the boat moves very slowly.   Drifting in maybe a 10 MPH wind the sea brake seemed to keep it down to about 1/8th MPH.  Drifts far faster without brake but forgot to measure that.  :-)   Very happy with these results.

If the robot dies for any reason this water brake should let it wait patiently for us to come rescue it.   Without this, on one side of the island it could get blown into the rocks and on the other it could get blown out to sea, faster than we might get to it in a kayak.  Having an automatic water brake should reduce the risk of loss during development. 

Testing Bluetooth Relay

We Got The Relay In The Mail.  It was cool to get my app working with the relay and the lights.   Also tested the motor too but it is not in these videos.  Quite glad it works!