Sailing to Hawaii Continued


A lot of people have contacted me for more pics, so I thought I would throw up some more. I will give some minimal content with them. If you have any questions, fire away and I will answer in the comments.

The first iteration of controller and Pixhawk. At the time I thought The Pi would be too big and draw too much power. Power was ok, but definitely too big.


When I got the hull from Radtek, I originally planned to use three 50 watt solar panels since I could not find a 150 watt panel that was narrow enough to fit on top. The width of the hull is 21". The length is 92"


I also planned to use the push pull system for the rudder. I quickly found out that even a heavy duty servo could not tolerate a week pushing water back and forth in my torture test bucket.

This caused me to rethink and redesign the rudder system. I eventually went with a thru-hull rudder system. This is a typical powerboat brass rudder port mounted on top of a 1/4" carbon fiber plate. On the bottom of the hull is a 1/4" ABS plastic plate. With weight ever increasing, I went with a kayak rudder and an 1" diameter aluminum rudder shaft.


After weeks breadboarding everything and testing, the final control system looked something like this. As you can see from the picture I moved from the servo rudder control to a linear servo. This one has 135 lbs of thrust and handled the bucket challenge with no problems.



The servo installed in the hull. I was very fortunate that the servo lined up almost perfectly with the rudder aperture. This is the only pic I could find of the servo so the candy obscures part of it.


Once all the systems were figured out, the challenge became how to fit them into a small enough water tight container. This caused a lot of grief.  This is a first attempt.


Finally I found the perfect waterproof container for the electronics


At the back are the waterproof ports for most of the electronics. The exposed sma connector attaches to the Satellite antenna. In all I drilled about 15 holes in the case for cabling.

3689632368?profile=originalExpecting that the ocean will batter the crap out of this, I mounted a carbon fiber plate to the the hull and bolted the electronics to the hull via vibration mounts. These mounts will keep the electronics about an inch off the hull. Hopefully the vibration will be significantly reduced.


Juggling family and a full time job meant that everything was done from 8pm till the wee hours of the morning, hence this taking a year (and counting). Soldering success at 3am. I had blown through 4 boards before this one by letting the solder touch and failing to clear it properly before plugging the controller board into a power source. 


I went with a 100AH battery after figuring in the power requirements. Just running the electronics 24/7 added up quickly not to mention powering the motor and the rudder. Since this platform is designed to drift with the current and monitor the ocean/air once it reaches a certain point, there will not be a constant high load demand on the battery. Getting to its destination is another power story entirely. lol...

I wanted to be able to stay on-station for many days even in the event of a lack of sunshine. A 100ah LIFePO4 battery weighs about 26lbs. You can see the battery installed in the hull in my previous post.


Once the battery was in, I then had to wire up to the charge controller and also wire up a secondary chrage port so the battery could stay charged while inside. At one point I had wires coming out of all parts of the boat. The usb ports for talking to the pixhawk and other devices are in the shot as well

That's all for now. If you have any questions, let me know.

I should have some more photos up at

E-mail me when people leave their comments –

You need to be a member of diydrones to add comments!

Join diydrones


  • If the birds are sitting on the solar panels and crapping all over them, you might want to consider some little spike strips to stop the birds from landing there. It's pretty common on expensive commercial buildings, some outdoor eating areas, etc.  You don't have to put spikes everywhere but may want to consider them strategically to prevent panels from getting covered.

  • @PB

    We used the serial JPEG camera from Adafruit. The image is read directly to the APM in chunks and sent over the satellite link. No need for a linux computer onboard.


  • @rusty

    Very cool interface.

    Shoot me a PM if you would like to see the web interface I have setup.

  • @rusty


    I have had the trolling motor running for 2 hours each night for the past 6 weeks and the motor stays in 24x7. I started out using the trolling motor to simulate water action to test the solar panels but along the way decided it would be a good test for the trolling motor as well. So far everything has checked out ok with both.


    Yeah, that is what I heard from others who have undertaken this. I remember seeing a picture of a liquid robotics drone that had its solar panel completely covered with bird crap.


    That is fantastic about jpeg pictures. That was one of my goals that I was not able to achieve. What setup are you using (camera/processor, etc.?) Originally I was going to use a Pi/pi cam setup but I kept pushing that back until the other programming got done. After that I hurried to get the boat launched before the final window (late august). But now that it is winter I should look into it again.

  • @PB

    Cool. Sounds like you are getting similar power and speed as us. Have you tested the trolling motor for long periods of time in saltwater?

    We used the RockBlock naked as well. Since it sends messages in 50 byte chunks, we spent a lot of time optimizing our message format to fit a ton of information into 50 bytes. We have a few longer message formats, like one to send jpeg pictures, that are broken into smaller messages and reconstructed once received. Our web interface is at and it's all open source.

    We'd also like to integrate the AIS data because we were constantly looking at AIS websites to look for traffic.

    A couple other things to watch out for:

    - Our surfboard came back with beak marks from seabirds who tried to peck it to death.

    - Somebody will probably find it. We were out for a total of 4 days and two boaters found it and called us. We printed our phone number on top and a "do not disturb" warning. I imagine more may have found it and never called.

    Good luck!


  • @Rusty

    Power breakdown as follows:

    25% 4 amps

    50% 9 amps

    75% 14 amps

    100% 22 amps

    At 25% the motor was using about 4 amps which gets me about about .5 to 2 knots on the intex in SF bay depending on current, wind and sea condition. So I figured Edgar could do around 2 to 4 knots knots going with the current following the clockwise route toward the open ocean at 25% thrust. Higher thrust is reserved for bad weather, high seas, etc.

    Originally I planned on running the motor during the day, and leave it off at night. Somewhere along the line, some conversation or piece of information caused me to rethink that plan and instead run at night. Somewhere along the line Running only at night jumped from being one possibility to Dogma. Go Figure.


    RockBlock naked with dual GPS/Iridium antenna.

    All the programming and The web interface was built from February to June 2014 - so it is solid and stable, but I am thinking on having the web interface rebuilt for aesthetics.Through the website I can poll the telemetry or set it up to automatically send slices of information at predetermined intervals. So I can receive weather information every 10 minutes, GPS updates every hour/ battery condition twice a day, etc. One of the things I am integrating on the website is an AIS overlay that uses AIS satellite tracking so I can do off-board collision avoidance.


  • @PB

    Thanks for all the answers! You've clearly put in a lot of thought and done a good job.

    I'm surprised that you are only planning to be under power for 4 hours a day. What continuous power draw do you expect while running? What speed? Since drag through the water is proportional to the velocity squared, I think it is generally better to travel slow and steady versus short bursts of higher speeds.

    During our first two tests we used a power management scheme that would basically keep the battery topped off and use all available solar power. The speed of the motors followed a sine curve based on the sun's position during the day. We only needed a small battery since it was only there to power the electronics at night and act as a buffer for the solar power. I imagine that you could eliminate your "prohibitively expensive" battery if you traveled during the day instead of night.

    All of the power management was done via PID controller that constantly adjusted power to maximize power output. That also allowed it to work really well on both sunny and cloudy days without our help.

    Also, what satellite radio are you using? Have you built the web interface for control yet?



  • @Rusty

    I actually am looking at your thrusters for another project I have. Congrats on a successful Kickstarter Campaign!

    Now that rover does skid steer, I can think about dropping the rudder on the next build!


    Edgar (this craft needs a name, so I am naming it after my father who served aboard an aircraft carrier)  has a trolling motor that hangs down about 36" from the bottom of the boat. I estimate the weight (right now) at about 100lbs, I am working on getting some of the "fat" out, if possible. The motor at the end of the shaft weighs 16lbs. The hull is 21" wide and has a convex deck. The section that supports the weather station and nav light is filled with aeromarine floation foam. The Hull as well is completely filled with flotation foam that will be poured around components.


    One of the benefits of getting a trolling motor is that the propellers are "designed" to cut through the seaweed and muck from close to shore fishing. I did some empirical tests with my intex mariner 4 and was amazed how it chopped through the crud. 


    One of the reasons I went for a jumbo and prohibitively expensive 100ah LIFePO4 battery. The plan is to have the engine and rudder silent during the day as the solar panel will charge the battery. At night the motor and rudder will activate via web commands. Right now I am figuring on three to four hours under power. All electrical systems (except ruder and engine) will remain on 24x7. I managed to get the amp draw down to around .5 Amps during the day and .75 Amps at night. The Nav light is LED and comes on automagically all dusk. Amazingly it only consumes about .25 Amps has 360 degree coverage and has two mile visibility. LOVE LEDs.

    In the event that the battery goes below 10.25 Volts the system is designed to shutdown. Once the battery is at 13 volts the system will reboot and continue. 

    Staying on course:

    Since the motors/rudder will only be assisting in propulsion and steering, I have charted a course along the current. I did a couple of simulations where i tried to make a straight line for Hawaii and it never worked. Once I plotted a course to follow the currents about 50 miles out from the California coast and then a right turn near the tip of baja it appeared I got more a less a free ride across the pacific and approached Hawaii from the south. Once I got close to Hawaii I would deorbit out of the current.

  • @PacificBots

    This looks great! I have a couple questions.

    Have you done the "bucket torture test" with your trolling motor? If so, how'd it go?

    How do you plan to manage power? Will you run the motor at night or drift?

    What happens if it flips over? I'd be particularly concerned about this because the hull is pretty narrow.

    We've been on and off for a long time on our SolarSurfer, which is very similar in size, solar power, and goal. We've done a few tests in the ocean and run into a few issues that you might want to consider. We did not have enough power to run at night and our progress was basically erased each day by the drift at night. We added a lot more battery power and solar power but haven't had a chance to test again, plus the days are very short right now.

    Our last test was supposed to run for a few weeks but was cut short by seaweed snagged in on of the thrusters. Do you have any plans to mitigate seaweed entanglement?

    I wish you the best and will be following the project!



  • I had pretty much the same idea to launch a vessel to travel from Cape Town to travel to the Prince Edward Islands (Marion Island) and back but dropped the idea due to cost and knowing how treacherous the Southern Ocean can be

    I admire your ambition and wish you the best of luck

This reply was deleted.