Home        Store        Docs        Blog

SolarSurfer Continuation Project

(Kevin) #1

Now that I have a few projects completed, I would like to turn my attention to something with some more persistence in the marine environment.

I really liked what BlueRobotics did with their SolarSurfer Project (https://hackaday.io/project/1677-solarsurfer) back in 2014 when their thrusters were first coming online. The project was novel for its time with incredible long range, satellite communications, command suite, and good instrumentation capable of doing real work.

In one of the build logs, the team outlined how they would like to scale up the project and get multiple units roaming the ocean for persistent operations. https://hackaday.io/project/1677-solarsurfer/log/10122-long-term-design

I know BlueRobotics is a bit tied up getting BlueROV2 units out the door, but I would like to take a stab at building my own SolarSurfer based on everything that is on the GitHub repository. Too much went into this to let it go away and having an unmanned vehicle running in the ocean that I can command from home is simply too awesome to pass up. There are also plenty of defense and oceanographic applications and this could be seen as a cheaper alternative to a Wave Glider or Autonaut.

With that, what are the current thoughts on something like this? The SeaCharger, a different variant, made it to Hawaii a few months ago. http://www.seacharger.com/ That one had a conventional steering set up with a propeller and rudder and seems to do rather well, but lacks sensors other than a GPS.

Starting with a hull, I’ve been looking at some type of HDPE/ABS paddleboard. Lifetime has one that is fairly flat and might make a good starting point: http://store.lifetime.com/products/blt/pid-90579.aspx



Surface Vehicle with Underwater Camera?
(Anthony White) #2

Great idea Kevin!

I’d love to help in your build - and even more so in later testing!

I’ve lurked in the forums while only interacting with you guys in person ( you, Rusty, Jonathan, Adam, etc.) for far too long. This is the perfect topic for me to start participating in the discussion, and hopefully even provide a great opportunity for anyone in the community!

As far as the hull for your build - I agree, a large paddleboard makes a lot of sense. I think the original board used in the SolarSurfer is still available… and I might know of another local source of an even larger (but heavier) SUP… I’ve seen inflatable ones for fairly cheap, and the cost / low weight is attractive in that paradigm - $250 instead of $600. Mounting things could get a little tricky, but maybe straps could both hold on panels and whatever skeg?

Heck, even your survey ASV could probably fit a couple solar panels, no? I believe panels like these are most appropriate - high power, light weight, cheap, water-proof (able?), etc. The SeaCharger looks to have used similar.

Batteries and electronics - I’m thinking basing this entirely on tube enclosures would be interesting, and have some potential stability advantages for what we lose in the convenience of a pelican case. If whatever enclosure(s) containing these bits were considerably negatively buoyant, and strapped directly to the underside of the board, the necessity of a long skeg mounted thruster for keel stability might be eliminated. This might help it not get entangled in ropes or drifting seaweed.

For propulsion, using motors and props vs. a T100 or T200 might get us a much higher top speed - the top speed I’ve ever got from any small scale design with the nozzles has been only 1.5 m/s… Perhaps we can experiment with a weed-eating prop rather than cage protection solutions that add such drag / biofouling area…

Now, as far as a testing and logistics…

I’m lucky enough to have found the opportunity to help the Catalina Sea Ranch in all sorts of things starting about 5 months ago, in addition to my day job making swarming ocean profiling vehicles. It resulted in Rusty and I setting up the NOMAD buoy & its website- which would now serve as the perfect testing-host for a prototype vehicle. It provides a high-speed link to the internet via Verizon 4G, shared to the ranch area via a wifi network. After a trip I’m making tomorrow morning, it’ll have an additional updating camera stream.

I’d like to extend the invitation to anyone who wants to work on something like this - I’d love to help you test your solution offshore! Even if you are not local to the SoCal area, if you have the urge to build something like this on your own but no way of doing testing - this is your chance!

Your SolarSurfer vehicle could do it’s first longer term trials tethered to the NOMAD, patrolling some transect of the ranch - 6 miles off-shore!

Whatever assistance I can provide - from the obvious NOMAD communication and attachment methods, to rapid rhib boat transport to the ranch - seeing a vehicle navigate, sending back sensor and camera data for long to indefinite periods of time would be absolutely incredible. Who knows- when a reliable solution has proven itself, CSR could be interested in funding continued operation!




(Kevin) #3

Awesome, thanks for wanting to join Tony! I thought this would have been up your alley.

I feel like I have a pretty good grasp of the hardware involved for a project like this, but I severely lack in the software and programming area. I looked at all the resources on the GitHub repository, but I don’t know where to start getting the Commander and API built, etc.

Hull: Interesting idea with the inflatable board, the prices are nice, but I have some longevity concerns with that material being on the water for that long. We would also need to apply some anti-fouling paint and I’m not sure that would be suitable for it. I don’t mind paying a bit more for a rigid board, but I don’t think a full fiberglass one is necessary.

Yes, I could go ahead and rig up a frame for my ASV for the solar panel, but I wouldn’t want to leave it on the open ocean unattended, it has no self-righting mechanism if it capsizes.

Solar Panels: I’m going to have to do some research in this area. There are a few marine grade ones out there that are completely sealed, but I haven’t found one that can produce the power we need. We may end up using a “weather-resistant” one like what you have found. If we use an 11ft paddleboard, we’ll have some serious surface area.

Battery: What do you think we should be looking for? We sometimes have a heavy marine layer, so the more capacity we have the better. Also, if we are operating near-shore, we should probably have some propulsion at night. Not full power, but at least a rough position hold or loiter.

Enclosures: Moving towards tube enclosures might be a good idea here and it would certainly keep everything dry should it get flipped or take on higher sea states. Enclosure shape might limit our battery type/shape though.

When you get a chance, can you draw out your submerged enclosure idea? I still think we need some sort of a pendulum for self-righting and I like how the BR team angled theirs in their concept rendering. Maybe an inverted A-frame design that has take-down pins so it could be disconnected for transport?

Propulsion: Propellers with weed-cutting props is what I was thinking. Trolling motors obviously figured it out. What do you think the largest prop an M-200 could turn at low RPM? I don’t think we need high speed, but efficiency and torque are most likely the keys.

California Sea Ranch: I really like this idea, it’s a great offer! What requirements would CSR have for mobile monitoring or a vehicle that could patrol the area?

Using the NOMAD buoy would also give us more options for communications. If could use the wifi, that keeps operating costs very low. Because we would be within line of sight to shore, we could use 4G cellular as a failover, and keep Iridium in reserve. I would still like to keep the transmission packets low.

(Kevin) #4

Here’s where I’m at for some basic design changes. Forgive the crudeness, just putting some ideas down.

I want to move the battery, type to be determined, to underneath the board. This eliminates a heavy topside and improves the righting ability if it ever gets knocked over.

Instead of the two Pelican Cases forward and aft, I think just having one in the middle would suffice. I’m also thinking about putting the sensor/comms mast amidships as well.

For the prop…could we try a small trolling motor propeller? Those are designed to be weed resistant.

(Anthony White) #5

Sorry for slow response!


Agree on the longevity / durability concerns on an inflatable version. I think we could re-use the first surfer’s board, panels, radar reflector, etc. where applicable (?) I’m all for maximizing our power generation capacity though - many small asvs use small wind turbines…


As for battery, I’d think ballasting the vehicle with 18650 cells would be awesome. I’m aware of a pack design we could lay hands on that fits nicely in 3" compartment, and ~doubles capacity in same volume of the green 4S hobbyking lipo you know from bluerov2. 6 or 10 of these would give our vehicle serious capacity!


I can throw some stuff in cad (go sign up for onshape.com if you haven’t already to access) - but to summarize my general structure idea - A single long 3" tube or perhaps 2 such tubes held in a frame as you have pictured, or even up against the board itself! If run on centerline, it should prevent flip-over in anything but the craziest sea state.


I think an m200 could spin a trolling motor prop, but I don’t know how well… I’ve got shaft couplings and a jig for attaching m4 props to the BR line of motors, so perhaps we could use some exotic RC boat prop - if metal we could even file and sharpen the blades. Placing it at the rear of the enclosure should also help, as you pictured. I’ve never seen a trolling motor prop handle swamps or some seaweed 100% successfully, but hope to be impressed.

The Catalina sea ranch would really love to have monitoring of the area for other vessels. They have an AIS receiving system setup, and the Nomad buoy transmits it’s location for large vessels, but monitoring for tampering, buoy loss, boat activity etc. is the current desire. A third (underwater!) camera should join the first to on the Nomad’s update system soon- theres no reason a roaming, ASV camera couldn’t be implemented on the same site! Throwing a Celsius temp sensor and perhaps a cyclops fluorometer I have for testing on it might enable the platform to have some scientific / biological justifications as well. The tricky part on this would be sensor drivers…


As for electronics infrastructure - I’m thinking the vehicle uses a pixhawk + rasperry pi 3. We can use the 900mhz radios common to planes/copters to get 1km+ telemetry link easily - and for more than the $20 that’d cost other systems offer as far as 40km range! The Nomad would host an instance of Mavproxy to forward mavlink communications from the pixhawk to the internet, for remote interface with our own local instance of QGC. Some vpn action might become necessary here.

Do you know from your rover / ASV experience with pixhawk if looping missions are possible? I think they are… a patrol back and forth between 2 or eventually 4 or more waypoints would be a good demonstration mission I think - obviously a position hold mode activating if batteries run low or by default at night?

Meanwhile, when in range of the Nomad’s wifi signal, the Rasperry Pi could handle upload of patrol pictures, logged sensor data, etc. It would simply join the network and upload files to a local ftp for viewing. Having installed a few cameras so far, setting one up to play nice with the Pi is trivial - and I don’t think we’d need more than a few simple scripts to handle logging and upload on connection.

The only other electrical system I can think of in this project would be the charge controller. Rusty did some neat things with the SolarSurfer - but if nothing else some off the shelf mpt solar charge controllers with load connection should handle this easily - the tricky part is finding ones compatible with our 4S voltage - or anything not lead acid for that matter… Definitely a point to research.

I think the single case in the middle makes sense - with an antenna mast next to it to minimize co-ax cable run length. A radar reflector could stick up on the tail?

How about both above and below water cams!?!?

We should pick a weekend to play with the components on a table!




(Kevin) #6

Ha no worries, I see the feed from the second camera on the NOMAD, looks good! And this is just us throwing out ideas for a new prototype, although my higher-ups are getting interested in the possibilities either for CSR or beyond.

Thoughts on the wind turbine…Very interesting and I know ASV has one with the capability. There are quite a few small marine turbines out there and it may make for a good auxiliary power source should the sun not play nicely. Do you think the power produced would be enough to offset the drag from the wind? We don’t have much sail area right now.

Yep, we are going to need some serious capacity. I forget, is the 3" battery you mention, a Li-Po? If so, I don’t think we can charge that at sea because it has to be balanced, etc. We might be relegated to making our own out of Li-ion or LiFePO4 to fit a 3" or 4" enclosure. The good thing is, if it’s being used as ballast, weight is no longer an issue and we can make it as long as we need.

(I’m on Onshape now) Ok, I see where you are going with mounting the battery enclosures to the hull. It sounds good, but we’ll have to try some flip tests. I am envisioning the vehicle being survivable to Sea States 4-5 (8-13ft waves). If they are mounted directly below the hull…where do we mount the thrusters? We should have some hull separation.

I agree on the electronics architecture with a Pixhawk and a RPi 3 and using Mavproxy to forward communications. However, we won’t need the 900 mHz SiK radios because we can just do comms over WiFi to the NOMAD. BR had already figured out the command interface with the SolarSurferCommander so I recommend using that. We’ll still have some work there though. I would still like to have back-up GSM and iridium comms in case comms on the NOMAD fail and we need to tell it to go somewhere and wait to be picked up.

Looping waypoints is definitely supported in the ArduRover software, at least the documentation says so. The last time I tested the AUTO feature with a few waypoints, it completed the mission but then went into HOLD mode. I probably missed a parameter, but I have seen boats do looping missions. Looping boat mission.

Yep, I’m fine with the dual masts, forward for comms, aft being for the radar reflector/and stern light. With the vehicle (not vessel) being unmanned, we are technically classified as debris, so we don’t have to show lights. However, if our mission is patrol and protection of infrastructure, we should definitely look like a vessel in accordance with COLREGS.

Yes, I was thinking of cams below and above, that would be very useful! That way CSR doesn’t have to use a UUV/ROV to do shallow inspections. This has a lot of good utility and I am totally onboard. There’s something to be said for something that doesn’t cost as much as a Wave Glider.

Give me a few paychecks and I’ll start buying parts and then we can start piecing this together. We’ll need a workshop big enough and I have to figure out some logistics for transport for a paddleboard.

(Anthony White) #7

Wind turbine - might work out well, but maybe better to consider it an add on after initial baseline position hold power is established. Also, should be an easy add on once other stuff is figured out.

The battery I mention is 18650 cells, which are LiPo. Balancing is easily achievable with protection circuits that batteryspace.com and others sell. The trick is finding a solar/wind charge controller with a 16.8V charged battery pack voltage set point. (?)

Definitely agree on flip tests, and don’t expect tucking tube up under board will result in a self righting system. However, it might take quite a lot to submerge over half of board (left or right side) and cause a side rollover… front flipping could be a risk? (rollover about pitch axis.) Going to start an onshape document now - available here.

The motors/prop(s) (do you think we should use thrusters instead?) could tuck behind the tube and thus blocked from impact with large debris? The tail of the surfboard would serve well to prevent the high velocity water sucking air into the prop.


We could do comms over wifi to nomad, but the cheap cost, longer range, and appeal of having separate and dedicated systems rather than connection sharing between mavlink packets and image upload make it seem simpler to not - at least to me. Once we start getting into the weeds of development it might be better to combine them?


Agree on lights - white on radar reflector, red/green on bow should be easy. Only on at night. Controlling this from pi driving FET/relay I imagine?

Didn’t even think of the shallow inspection capacity the underwater camera would bring! Was more thinking of the odd wildlife sighting. There is an amazing amount of marine life starting to hang out around the shellfish lines. Long term we’ll need a solution for biofouling on that camera’s dome though…

I’m happy to help on costs, and Rusty mentioned much of the old solarsurfer may be available for re-use! As for workshop, we could use mine or find some more middle-ground in terms of travel. My truck should haul this thing no problem?





(Kevin) #8

The model is looking great so far in CAD! That definitely helps with the visualization.

Ok and I can see where this is going now, we are working on a relatively short range. We may as well “geo-fence” the vehicle in between the buoys at CSR. With that Wi-Fi for primary comms should work out fine for range, and have the SiK 900MHz radios for backup. QGC supports switching between two links…but I’m not sure what happens on the vehicle side. I have a proposed link diagram attached. How is NOMAD connected to shore/ the internet?

And I’m fine with the above. I couldn’t get the SolarSurfer repositories to install properly off GitHub. The code might need to be updated because they were throwing a bunch of errors when I got to the “npm install” even though I updated phantomjs and npm.

I’ll start researching chargers with variable set points, if they exist. What about regulating 24V down to 16.8V?

We might be ok for front flipping. If the seas are heavy enough for a front flip…there’s probably bigger issues.

Now that you mention it, we might be fine with T-200s tucked behind the battery tubes with the 3D printed guards for good measure. I found that mounting the thrusters 2/3rds of the way back from the bow worked out well for control on straight aways and still having enough response to rotate on the axis. Would you consider having 2 battery tubes, one on each side?

No idea on light relay, I’ll have to look.

For the biofouling…I figure someone is going to make a trip at least once a month, the dome could be wiped off then and the propellers checked. We have the benefit of being close to service, unlike some of the other long range solar powered vehicles. I also don’t know how we would be able to handle 2 cameras yet.

I think we should probably use your workshop if you have the room…and your truck is bigger than mine. There are a few paddleboards on Craigs List up in LA, so it shouldn’t be too hard to find a cheap one. I see them going for $250 instead of $600 at retail.

I’m really excited about this one! It’s where AUVs want to go and I don’t know if ASVs have demonstrated it reliably enough to actually be useful. It would be very cool to have one or two of these patrol the area, and be available to be taken control of to do light inspection work…or go look at the wildlife.

(Anthony White) #9

Currently the Nomad uses a mofi router to connect to Verizon’s 4G network, with sufficient bandwidth for hd video streaming and good signal strength. It hosts a WPA protected wifi network with internet access, but the external antenna could use an upgrade, and we could also improve range by moving the router’s second diversity antenna to the exterior of the buoy. I like and agree with the link architecture diagram! I’ll work on one for the system internals this week.

I’ve started a spreadsheet to try and track parts, costs, etc.


I’m convinced the duct of a t200 will slow us down, and limit us to about 2 m/s. Some protective features and bare, high pitch props should let us hit 4 to 5 m/s peak?


The charge controller in that spreadsheet has a configurable load output that can come on at sunset and off at sunrise - drive our navigation lights automatically.


I’m hoping we can either call the vehicle in to remove biofouling, or even have it drive into some buoy mounted brushes perhaps?


(Kevin) #10

Ok sounds like you have a good handle on the data link.

I have some different components I’d like to try to use, so I added another tab for what I’m thinking. If anything we can always have 2 prototypes running around.

I haven’t found any model boat props that I like and I’m thinking more along the lines of good endurance and efficiency rather than speed. What about the propeller model you made for “X” company? That fit a M-200 I think. From there we can probably just make a 3D printed mount in a few pieces. I was looking at this to see how their seaweed guards looked. http://asvglobal.com/product/c-enduro/

Interesting thought to call the whole vehicle in for servicing/cleaning. I’m on board with that, it’s only 6 miles out to CSR. If we do it that way, it’s definitely going to lose WiFi signal and most likely the 900MHz as well. Might be worth it to put a 4G modem on it for those service runs. I don’t think we’ll need the Iridium.


(Anthony White) #11

Only diff. I see is the pixhawk source? General wisdom seems to think those are more reliable, I’ve had great success with the cheap knockoffs though…


I can definitely print those props, or perhaps a modified, even larger version of them. I’ll set that going tomorrow.

ASV is using the lower half of COTS outboards coupled to BLDC motors on the inside - looks like the enduro now sports a kort nozzle of sorts as well. I only peripherally saw elements of the enduro when I worked at ASV/C&C, but think our general approach to this vehicle is very different - not based on PLCs and din-rail mounted components but more more miniature and cheaper hobbyist/consumer grade parts.

Throwing a mofi on the vehicle would be easy, but an iridium would help for the worst case drift out to sea failure mode. If we have one available with messages, seems using it during the prototype phase if nothing else will let us rest easier.



(Kevin) #12

Sorry about that, working on picking components as I go. As for the Pixhawk 1, just a personal preference to use genuine parts if I can find them. I updated the component list and component diagram, but it is still a work in progress.

I agree on how we are in the “more miniature and cheaper hobbyist/consumer grade parts,” I was just looking at how they did their propeller guards. Besides the comms, I think the weed guards are going to be key. Right now, I’m thinking of using your propeller design, the M-200s, a motor mount that looks pretty much like the new VideoRay thrusters, and a cage around the whole thing. The space between the rotor and the stator will also need a sleeve to keep things out, that was where I found the eelgrass always liked to get caught on.

Sure, I’d love to have Iridium on it, the messages are just more costly than cellular. I have more applications for a vehicle like this after CSR, so I’d like to keep it on if we can.

Here are the proposed data link modules right now for telemetry:

  1. 900Mhz Radio - Medium Range, Medium Bandwidth. Telemetry only, needs a compatible receiver radio. No charge.
  2. Wi-fi - Short Range, High Bandwidth. Can do telemetry and video. No charge.
  3. 4G Cellular (Independent of NOMAD) - Medium Range (LOS to a tower), Medium Bandwidth. Can do telemetry and video. Charge per minute of data.
  4. Iridium - Long Range, Low Bandwidth. Telemetry Only. Charge per credit used (every 50 bytes).
Edit: What do you think about using a RockFLEET unit? http://www.rock7mobile.com/products-rockfleet I like how it can switch between GSM and Iridium.

(Kevin) #13

If we can’t get the original SolarSurfer software working again for the Iridium comms, it looks like the Mavlink/QGC team is working on a solution:




(Anthony White) #14

Agree with your connectivity option list. 4G cellular with verizon has first 10 gigs included in monthly fee (~$80 to 100?) then extra after that. Iridium will hopefully nicely work out within QGC somehow?

Mofi router Nomad uses is SIM4, just noticed they have version that breaks out GPS, would be nice but likely not integrate with pixhawk - redundant source, just needs a second gps antenna?



Got a CW and CCW prop printed and epoxied, ready for whatever our first setup is.

Email’d these people I could pickup with truck and pay cash this week:

http://losangeles.craigslist.org/wst/spo/5862975362.html $325

http://losangeles.craigslist.org/lgb/for/5861843298.html $260


I think doing a layer of fiberglass epoxy and cloth around the foamie would give us some ruggedness / long term buoyancy.


I’ve got a piece of t-slot that might make a good keel / thruster mount, use some simple hardware to bolt it to an underside plate.


If we go the rifle case + solar panel mount on top, were we thinking batteries there? I can make a small 4" dimater enclosure x ~7" tube with IP camera and temp sensor, hang the available Cyclops fluorometer off of it as well. Place at back, directly below tail mast / radar reflector / diversity antenna mount? Do a second small mast with antenna and wind speed / direction sensor in middle?


Gotta spend some time on the onshape model to capture this stuff…


We should put together an parts list to get from Rusty - motors, enclosures, esc, etc.




(Anthony White) #15

Rifle case selection - I figure bolts and washers through bottom of this to thick, heavy plate on underside that hosts gussets holding keel bar, or hinges for 2 point keel. Maybe some tensioned wire lines to slice / guide seaweed as well?

A small

B medium

C large

I’m thinking a lighter and cheaper frame material to hold solar panels would use DIN rail, rather than aluminum extrusion. Just need to get panels and figure mounting hole spacing? Can paint it to resist corrosion.


(Kevin) #16

Yeah hopefully the Iridium will work out with QGC. I’m looking into how to chance the MavLink transmission rate so it doesn’t cost us $50 to run for 5 minutes.

I like that MoFi, that’s a really good find! and now that I see it has an integrated WiFi, I’m wondering if there is a transceiver out there that does all three of what we need (WiFi, 4G, Iridium)? This goes beyond the RockBlock. Switching back and forth to the lowest bandwidth is what worries me a bit.

Great job on the props and those paddleboards look nice and fat, good for us to drill into. If you want to coat it, I leave that to you, I’m a bit inexperienced in that area.

I’m onboard with your small enclosure at the stern under the aft mast.

After our discussion, I’m really thinking about having two rifle cases mounted topside where we can put the panels on top and all the electronics and batteries inside. I think Sealed Lead Acid is our best bet. Heavy, but stable and far easier to charge. we have quite a few options and capacities with Universal Battery.

I’m onboard with your small enclosure at the stern under the aft mast. We’re going to have some drag…but we need the sensors.

Regarding the underside, we can either mount the rifle cases (I’m thinking Pelican 1700’s) to a heavy plate, or maybe two parallel rails? Might cut down on the drag a bit. If we move all that battery weight topside we’ll need a pendulum keel again.

The DIN rail is a good option, but the aluminum extrusion gives us some pretty good building points. I’m undecided right now.

I also found these cool masts: http://scanstrut.com/products-22/gps-mounts-category/tb-01-t-bar-detail Might work for at least the aft one. I’ll have to see about the one amidships.

(Jim N.) #17

Hay Kevin, you guys need a programmer with autonomous systems dev experience? Cuz I know a guy :slight_smile:


(Kevin) #18

Hey Jim, thanks for the interest!

Yes, in fact, we could use some help with programming. Communications is one of our biggest issues right now. Tony and I are considering methods to switch comms between Wi-Fi, 4G, and Satellite Short Burst Data (SBD). The satellite is the most problematic to get working with Mavlink. We’re thinking that we might need a python script to switch to the satellite receiver when the vehicle has a loss of 4G comms (out of range).

I’ll draw out the architecture when I get a chance, but the basic premise is that we have Mavlink data being pushed from the Pixhawk to a companion computer (RPi3). The CC then pushes to an Amazon CE2 Instance running Mavproxy and then we can link up into that IP on our ground control station. A guy by the name of Antimatter Crusader has done it on the Ardupilot forums and I’m just waiting for a tutorial.

(Jim N.) #19

Kevin, sounds good. I’m familiar with mavlink, etc. I also started working with Piccollo SL and the Iriduim service, so might be helpful. Let me know.


(Kevin) #20

The retrieval report from the SeaCharger project was an interesting read: http://www.seacharger.com/tracking.html

It lost both blades on the propeller as well as the rudder and the hull was all covered in bio-fouling.