The final final end of the LoCARB project
After 107 days of ocean faring, LoCARB has finally failed to send updates. The boat began sending a flurry of updates with details which indicated a rapid drop of voltage before it went silent. I believe the power system had finally shorted out. At the time of this post, LoCARB had spent 107 days out on the ocean, traveled 1327 miles from launch, traversed a total of 1729 miles of ocean (including all the wonky zig zags and backwards motion), braved and survived bad weather with winds up to 24 kts (also made it farther than the solo-kayaker, and solo wing guy), and ended up 1476 miles from Hawaii! What a run for a boat built by a stay at home dad! 🙂 Learn more about what happened before LoCARB went silent.
Position of LoCARB at the time of this post
Some reflections about a project like this (for anyone whos interested…).
When I began this project (which started February of 2019), I literally had no deep hands on experience with RC, 3D printing, CAD design, coding, micro-controllers, fiber-glassing, boat construction, solar systems, and programming. Uhhh…yeah, basically all the skills you would need in tackling a project of this scope. The project was incredibly intimidating because I literally knew nothing.
As the project progressed, it became less about the destination and although very cliché, more about the journey. The skills I was forced to learn were worth every dollar put into this project ($1900 for v1 + $700 v2. Is this still considered low cost? Ha!). The newfound skills provided a springboard for me to jump into building a she-shed from scratch in my backyard, a really fancy automated rabbit cage, and my recently completed “just-for-the-hell-of-it” project, an active spoiler for my corolla. Jumping into any other project now doesn’t seem as daunting as it was before completing this boat project.
Good is good enough
Among the things I set out to learn is if one can build a cheap and low cost autonomous boat that is robust and capable. As with all projects, one comes up with an ideal concept that maximizes every aspect of its parts. “How fast, how efficient, how durable, how amazing, could I make this” is the name of the game. But in the end, I realized…sometimes good enough is good enough.
I set out to design a magnetic coupling (like Damon from the Seacharger project did) on the cheap, which turned out to be the reason for my first 3 launch failures. It wasn’t until I was able to recover the pod and open it up did I see why it failed and found that 3D printed parts like to melt. In the end I just used the RC motor with no protection from the elements…which surprisingly was still operational the day LoCARB went silent!
The same holds true for using a dedicated MPPT solar charge controller. Honestly I found a $12 buck converter performed better than the dedicated and much more costly solar charge controller.
If I had just started with the basics and launched, I probably would have progressed much further and sooner. But then again, what’s the fun of that right?
Staying true to the vision of the project
In retrospect I do believe I’ve accomplished everything I wanted out of this project and held true to its core principle of staying low cost by using as many off the shelf parts as possible, and also by experimenting with different methods of how something like this could be built. There were several instances during the build where there were no available off the shelf parts but with basic skills, could be designed in CAD (I used TinkerCAD).
In retrospect, I feel the first version of LoCARB with the car battery performed just as well as the LoCARB refresh v2 which used a lighter 18650 battery pack. The first version was a pain to move around because of its weight and size, but in hindsight, I believe its simplicity and approachability wasn’t a bad place to start.
Experimenting and testing
A notable hurdle in developing the boat was actually from the limitation of not being able to test the boat for an extended duration in ocean waters. Seaworthiness was basically determined by how well the boat behaved in a lake which could not duplicate the conditions the boat would eventually come to experience. Ocean strength wind, current, and waves, introduce variables and conditions which I feel aren’t fully appreciated until they are experienced when running the boat much farther from shore.
I’d also like to note how amusing the project can become by experimenting with unconventional ideas like adding a second thruster or using random pieces of trash to build a solar platform. I believe one of my crowning achievements in this project was the effectiveness of using an aluminum speed angle to make a rudder skeg. That skeg kept the rudder from destruction despite being tossed onto shore 3 times! Best $4 ever.
Cheap out on certain parts but not others
I set out to experiment if I could build (DIY) high quality parts that would be costly (relatively speaking of course) to have produced (such as a magnetic coupling) and found that for mission critical components its worth the money to get it machined to be both precise and durable. This marginally increases the cost of the project, but also decreases the time and hassle of the part failing…and reduces the costly (in time and money) subsequent repairs.
Assuming just creates more problems
Whenever I assumed a conclusion for a problem the boat experienced, or assumed an approach would work for assembly, I was consistently proven wrong. Could I use PVC solvent in an enclosed area where epoxy has not fully cured? NOPE. Did I have to go on a rescue mission 6 miles out in the ocean to rescue LoCARB after it its motor disengaged after the first launch? NOPE. Did I make a materials mistake by believing I could 3D print waterproof end caps and a magnetic coupling which could withstand the heat generated by an RC motor? YEP. These were all assumptions which I guess needed to be gained from experimenting, but dang. Those were costly and time consuming mistakes.
If I did it all over again
There are definite design changes I would make if I were to make another boat. First, I would place less emphasis on the efficiency of a self-righting keel (although important) using batteries as ballast, and instead put more emphasis on keeping the electronics dry and accessible for testing and reconfiguration. Having had the privilege of building both large and small pods for different weight batteries, I can say that I would much rather have all the electronics accessible within one easy to access top-side waterproof compartment. I would still mount the motor on some sort of underwater extension (like say, attached to the bottom/middle of the keel), but all motor electronics would still be accessible from the waterproof compartment.
I would try to minimize the amount of wiring and possibly print a circuit board to keep everything tidy. There was just so much soldering involved in having so many sensors and wires, I got tired of it.
I would use a bigger watertight compartment for my electronics. It was a pain not being able to easily access the wiring and add things when I wanted because of space constraints. In the end it worked out ok, but I would have liked more room.
I would generate more power with an additional solar panel (and up the size of the battery) and increase the throttle to increase the speed of the boat. This bump in power generation would help overcome wind, currents, and unforeseen excess power drain from the system. It would just allow for more flexibility in boat operation.
I would swap my leak sensor for a wind speed/direction sensor if possible. Knowing what wind resistances the boat is experiencing would be nice to know in case the boat is behaving in a way which indicates a failure or malfunction when in reality, its just not moving because of wind conditions.
I would also implement a simple redundant circuit (by dumb relay) to allow the solar panel to supply power directly to the electronics (bypassing the power supply and battery pack) in the event the voltage of the battery pack dropped to less than ~5v. Cost would be negligible, but would ensure a redundant power supply during the day when sun power was sufficient for operation.
Be comfortable with people who are interested
We were and still are constantly amazed at the support and interest generated by this project. All our kids’ teachers and classmates know about the project and while it was in its repair and operational stages, kept asking for status updates. My wife would even walk into the break room at the hospital where she works and fellow nurses would be browsing the website asking about the boat.
On the last launch, some boat personnel on the dock asked if the boat was a school project, to which I replied “no, it’s just for fun,” at which they laughed and muttered to themselves that it was almost inconceivable (insert timely The Princess Bride reference). Kerry, the captain of Bare Bones who I met on a chance day while walking around Pillar Point harbor has now become a friend, and we plan to go for a harbor tour in his new-to-him yacht. Him and his buddies have been with me through all the launches out to sea which is something even my wife cant claim!
There is something about the need for assistance from people with skills in areas of expertise far outside ones own in a project such as this, which inevitably forces some good people mingling. This was something I didn’t fully expect starting the project, but really appreciated at its completion.
Get a mentor
Little did I know a quick email to Damon at the Seacharger project asking if he could share his code for Seacharger some 3 years ago would inevitably lead to him becoming (unbeknownst to him) a kind of mentor throughout this project. His experience and skills in engineering (and making spreadsheets that calculate things), is so far above what I am capable of, was such a necessary resource when I ran into trouble. His encouragement was also something which kept me going when I was frustrated. I think this entire experience was enriched tenfold for his willingness to share his thoughts and skills with someone whom he’s never met!
I believe the project was successful despite the operational failure and subsequent failure to make it to Hawaii. I believe LoCARB answered the question if you could build a low cost autonomous boat with off the shelf parts, with “you can, but it depends.” The cost of the project can be kept under $2500 if trial and error didn’t occur, but may lack in durability. I think for $3000 one could have a reliable boat, for $5000 a pretty nice boat, for $8000 a state of the art boat, and for $10,000 a luxury autonomous boat.
I wouldn’t consider an update from the boat at a later time impossible, nor would I say the boat (or whatever is left of it) could never make it to Hawaii. I do have hope that given enough time LoCARB will wash ashore some place interesting and maybe one day I may find it had eventually made its way to Hawaii…or even better, someplace farther away! Well, that’s the hope anyways.
3 thoughts on “The final final end of the LoCARB project”
Fantastic project, massive congrats on what you have achieved here, just brilliant.
Adrian – during the long COVID epidemic and the lockdowns, your project became a really great example of how one person and one vision can be an example of fun, exploration, and pushing to learn more. Congrats on your outstanding results.
Found this while randomly searching around for references on servos for autonomous kayak rudders. This is a great project, and something really up my alley in terms of interests!
Seeing that it’s been almost a year since V2’s demise, have you come around on the idea of a V3? I would love to donate a few bucks to the cause–you should start a kickstarter for a V3 build!