The Forever ExpeditionJanuary 1 2015
You are set up to receive WePay payments.
Here are the details to your account:
Name: Andrew Thaler
Account ID: 309591383
While California continues into its unending epic drought, we on the east coast experienced a thousand-year deluge that flooded out entire cities, and left huge chunks of North and South Carolina under water. We got off fairly unscathed, but, just to give you an idea of how intense it was, here's a video of me, driving my OpenROV through the front yard.
After two successive dives, including one where we flooded at 70 meters, the Colonial Lee is ready for some TRR (tear-down, rebuild, revive). The Niskins and servo kept firing all the way to the bottom, though we had to do a power cycle on the final dive, which triggered the servo early and only partially. It was the only misfired Niskin of the entire cruise.
Over all, this was a triumph. I'm making a note here: "Huge success."
The final cast.
At last, we come to our last station. The OpenCTD has gone down on four casts this trip, ranging in depth from 13 meters to 140 meters. I have a lot of work left to do with regard to processing data, assessing how accurate and precise the probes are, and fixing all the issues that I'm sure will emerge as we dig into the data.
But for now, an end-of-cruise teardown indicates minimal water intrusion and no degradation of internal systems. Data logging was consistent and without critical error. We also detected no mineral oil leakage during the casts, which is important for both equipment housing and environmental concerns.
Rapid prototyping at sea.
One of the secondary missions for me during this cruise is to assess the utility of using a 3D printer for rapid prototyping at sea. I intentionally brought minimal gear for my deployments and printed out parts as needed. In addition, I'm working with the other scientists to fulfill needs (you always leave at least one thing back on shore).
The first fab job was a filter holder for Dr. Tristan Horner. It doesn't look like much, but this tiny piece holds a very delicate water filter that went over the side on our last CTD cast to provide a water column sample to compare his discrete samples against.
The portable oceanographer.
In preparation for small boats ops, I put together a field backpack with all the tools of an open-source oceanographer: OpenCTD, OpenROV, mini-Niskins, bait bucket for camera work; plus, of course, a notebook and laptop to run the system.
All crammed into, onto, and around a small backpack.
There's something liberating about making one of your expedition goals "push your gear so far it fails, then figure out why." It means that complete destruction is valuable data.
Which is where we are with the underwater servo, a prototype gifted to me by the good folks at OpenROV central command. It triggered beautifully at 45 meters, collecting yet another discrete water sample. Unfortunately, as my robot breached the surface, we were greeted by a sheen of food-grade mineral oil, a clear indication that the servo had failed. Closer inspection revealed that the o-ring around the servo collette had been pushed in, breaking the seal. Not much I can do out here to fix that.
Meanwhile, my personal ROV seems to have the particular tendency to flood at exactly 50 meters in very cold freshwater, so it's also receiving a little TLC.
Even with these setbacks, we had an awesome dive. At about 38 meters, we hit a layer of mysid shrimp jauntily swimming towards the surface. It lasted for all of two meters, a reminder that even seemingly stable lakes are heavily stratified.
We have some other plans for the robot, so, for now, the Niskin bottle has been removed.
Tonight, before closing up ship for the evening, we took my OpenROV out on the deck and tested the complete Niskin/Servo?ROV-mount system in the water for the first time. The rig obviously makes the bot heavy, but the servo works perfectly and the bottle slammed shut on first signal. After driving it around for 15 minutes, we returned to the surface and confirmed that the Niskin bottle was sealed tight.
We have our first OpenROV-powered water sample!
Tomorrow, the OpenCTD goes out on the line.
Checking in from the Bean Town airport.
Important advice for anyone going on expedition (and really, just good travel advice in general), Chicago is one of those airports that are consistently delayed. Tight connections are a bad idea. I always budget a 3 hour layover if I have to fly through Chicago, just in case. Today, it paid off, as my flight out of Richmond didn't even take off until well after it was scheduled to land. And while everyone else on my flight is standing in the customer service line trying desperately to get a different flight (also often impossible at Chicago), I get to enjoy a drink and dinner. So there's that.
Traveling is a mess, and it's an even bigger mess if you're both an international traveler and an environmentalist. There's no easy answer. Flying is nasty business and there is not shortcut to absolving yourself of carbon guilt. Most of the people I work with accept it as a necessary evil, genuinely believe that the work they do will ultimately yield greater benefits to the environment than the harm they cause by flying. I tend to agree.
In the last couple of years, I've been trying to remove extraneous travel from my professional life. Teleconferencing technology has helped quite a bit, but for times when I have to travel, I've engaged in the infuriating, often baffling task of excising "stupid routes" from my itinerary.
A stupid route is a route that takes you thousands of miles in the wrong direction. A stupid route takes you from Richmond to New Jersey to fly to Atlanta. A stupid route sends you north to go south, east to go west. If I can't avoid flying, I can at least avoid flying in the wrong direction.
Unfortunately, stupid routes are cheap routes, and clean, no-nonsense flights that go from point A to point C though point B, which happens to be reasonably and rationally located somewhere between A and C are not cheap. No-nonsense routes are often twice as much as stupid routes, if not much, much more.
But it's worth it, at least to me. It's the cost of being at least a bit more aware of our impacts on the environment.
Someone should write some code that finds the least stupid routes for any given itinerary. I'd use it.
Checking in from the Richmond Airport. After years of flying question-free, in the last five trips, the TSA has asked to inspect the OpenROV at security. I find this comforting, but also, fellow ROV World Travelers, plan accordingly. The last two times has resulted in a 15 minute delay.
Down to the sea again.
Over the last month, I've been working frantically to finish some major hardware projects--both the 3D-printed Niskin bottles and the long-suffering OpenCTD. Both projects are now ready for field trials, so it's time to head out to sea. Tomorrow I'm heading up to Lake Superior to participate in the UNOLS Chief Scientist Training Cruise aboard the R/V Blue Heron. This is also a great opportunity to test out the OpenROV servo prototype in the field.
The Niskin Bottle and OpenCTD could become powerful tools for open science, citizen science, and oceanography for everyone.
If they work. Let's find out.
How Man and Machine Unlocked the Ocean's Depths
Check out my latest article on Motherboard!
"To understand the sea, to dive beneath the waves and experience its awesome, unforgiving wonder, was once the privilege of a few fortunate explorers, but new technologies are lowering the barriers to entry and creating a generation of globally-connected ocean explorers."
A 3D printable Niskin bottle for water sampling? Yes, please!
The Niskin bottle, a seemingly simple tube designed to take water samples at discrete depths, is one of the most important tools of oceanography. Coupled with a CTD, an array of Niskin bottles fit into the rosette, a Voltron-esque amalgamation of everything an oceanographer needs to profile the ocean. Niskin bottles are neither cheap nor particularly easy to use. A commercial rosette requires a decent-sized winch to launch and recover, which means you need a vessel and a crew to deploy. For Rogue Ecologist and citizen scientists, getting a high-quality, discrete water sample is a perpetual challenge. With tools like the OpenROV and the soon-to-be-completed EcoDrone, I wanted a Niskin bottle that was light weight and capable of being mounted on both underwater robots and quadcopters with ease.
After a few months of brainstorming and planning, I sat down this Friday and began building a 3D printable Niskin bottle that could be hand deployed or mounted on an OpenROV or drone. While this version is designed around a 1.25 inch acrylic tube, the trigger mechanism can be expanded to fit any size pipe. The trigger is driven by a waterproof servo developed by the good folks over at OpenROV. Everything else can either be purchased off-the-shelf or printed on you home 3D printer. Later this month, I'll be taking my prototypes out on the RV Blue Heron for field testing in Lake Superior.
Read full build instructions at Southern Fried Science: southernfriedscience.com/?p=18730
Not every great discovery takes place in the water.
While wandering though the shops in nearby Bolton Landing, I happened upon two nearly pristine copies of Life Magazine's 1960 issue featuring the dive of the bathyscaph Trieste into Challenger Deep. I've been searching for this particular issue for years. To find two in such great shape is an absolute treat.
Paying the Exploration Tax
Honeycomb was never meant to last this long. Custom built for the Bay Area Maker Faire, it had a special shell that prevented mobs of excited kids from catching their fingers in the thrusters. It performed exceptionally through its first Maker Faire and, several weeks later, was shipped of the DC for National Maker Faire. Over these two events, over 1000 attendees, most of them children, flew Honeycomb.
Honeycomb was flown by more people than almost any other individual OpenROV.
After two intense Maker Faires, this little robot was retired to my fleet, to serve as a backup machine to Colonial Lead, diving in when needed.
Unfortunately, she dove this morning for the last time. While attempting to examine the wreck of the Land Tortoise, in Lake George, New York, Honeycomb's tether broke. After a prolonged survey, it was clear that the little robot wasn't returning to the surface.
So what happened? Honeycomb's original tether was barely ten meters long. In order to convert the robot to field operations, we extended the tether to fifty meters, using spare wire lying around our workshop. It was clear that the tether had failed at one of the several solder joints. The failure was entirely my fault, for using hacked together materials rather than investing in a full length tether.
Colonial Lee, though it suffered a very minor leak after a fifty meter dive (the teaspoon of freshwater in the housing is the reason Honeycomb was in the water at all), is still alive and kicking, so our Lake George Adventure continues.
This is the Exploration Tax. At some level, you have to be willing to accept losses in order to push the limits in order to find the limits of these machine capabilities. Sure, these robots wouldn't be very good tools if we lost them all the time, but after expeditions stretching from Papua New Guinea, to coastal Florida, to Lake Tahoe, and the Chesapeake Bay, after more than 500 dives with OpenROVs from the Kickstarted 2.3 to the current 2.7+, I can accept that a lost robot is the price we pay.
What I can't accept is leaving trash on the sea floor or lake bed. We've left notes with the local dive shops and the Lake George Association to keep an eye out for the robot. It was lost on a fairly well traveled dive site, so I have hope that someone will recover it. And who knows, maybe it will slowly rise back to the surface and some intrepid kid will find it resting on their beach, waiting to be repaired.
Until then, let's hope we get another 500 dives before the Exploration Revenue Service comes a callin'.
Speaking of invasives.
We' spotted a few Zebra Mussels on our systems-check dive yesterday afternoon. Zebra mussels are a major species of concern throughout the Great Lakes and, along with rainbow smelt, Asian clams, watermillfoil, spiny water flea, and pondweed. and are among the most prolific invaders in Lake George.
Essential pre-expedition research.
Just because we're technically on vacation doesn't mean we aren't doing our proper due diligence before sending our robots into the water. Lake George has a ton of history and is facing significant environmental challenges. Team Forever Expedition is spending our evenings reading through the 2014 State of the Lake Report, which covers the last 30 years of scientific surveys of Lake George: fundforlakegeorge.org/stateofthelake
And, of course, familiarizing ourselves with all the local invasive species concerns and potential vectors: fundforlakegeorge.org/threats/invasive-species
The next major expedition.
Amy and I are on our way up to Lake George, New York to meet up with David Shiffman and spend five days on the water, searching for shipwrecks from the French and Indian War. Over 200 warships were sunk in this lake in the summer of 1758, the most famous of which is the Lost Radeau: thelostradeau.com, With our tiny fleet of ROVs, we're hoping to spot the remains of a few more.
Kirby, the world's greatest farm dog, will stay behind to tend to our livestock.
Simple. Ubiquitous. Cheap.
Technology becomes really interesting not when it's at the bleeding edge of development, pushing the limits of what's possible, but when it becomes embedded into society in ways that allow the power of mass adoption to reshape how we interact with the world.
I had one of those clarifying moments earlier this week coming back from a trip up the Delaware Coast. Sitting in the rest stop along the Chesapeake Bay Tunnel Bridge, we saw a large cargo ship begin its transit. It was sitting high in the water, and the shape was a little bit off--it didn't look like most of the big container vessels that pay the Port of Baltimore a visit. The shipping company, Wallenius Wilhelmsen Logisitics, was clearly emblazoned across starboard side (and, presumably, the port side), so I pulled out my phone and looked it up. Within 30 seconds, I had ship tracks for the Oberon, cargo data, specifications, and it's next port of call.
Five years ago, accessing that information that quickly would have been unimaginable.
That moment comes on the heals of this report, coordinated by SkyTruth, on crimes aboard ships that routinely flaunt the law: nyti.ms/1fYzIL3
On the high seas, these vessel are nigh untraceable,, but perhaps with the technology we already carry around every day, we can make things like this just a bit harder to get away with.
At the very least, we have the power now to learn a little more about our local shipping channels.
Reflecting on three years as a Maker.
Three years ago, "maker" wasn't a word that had entered my vocabulary. Back then, I was a carpenter, a furniture builder, an occasional boat builder, heck, I even worked on multi-million dollar ROVs, but, somehow, I hadn't yet found the maker movement.
Then a couple of misfits from California kickstarted an open-source ROV and the rest, as they say, is history.
Sitting down and finally finishing the original 2.3 kit gave me the opportunity to reflect on how far I've come. My solder joints were such a mess that I ended up redoing all of them. The Beaglebone was fried--likely shorted the first time I tried to power the robot on. The tether was, well, not good. Comparing that build to my latest 2.7 (or the half built art bot; or the 2.6 currently out on loan to a colleague in Alaska) is like night and day. Even though my first 2.3 is imperfect, and will likely only ever be used as a showpiece rather than an expedition vehicle, it still the most important robot in my fleet. Tins Lee, serial number 116, was the robot that started everything else.
My first "maker" project was the Sea Leveler--a turn of the century water level gauge that I hacked to listen to twitter and rise and fall with fluctuations in the online conversation surrounding sea level rise. That project helped me learn Arduino, talk to Twitter's API, and run a stepper motor. Since then, I've moved on to bigger, more complicated projects, but it all began with that little kickstarter kit.
The Rare OpenROV Double-Phoenix
How many 2.3 bits can you find on this table? Answer: too many. Thanks to a quick tour through the OpenROV HQ Scrap Bin, I managed to scrounge up an extra brain for my Kickstarter original 2.3. This poor blue creature never quite made it into the ocean, so one of the major goals for the Forever Expedition is to finally get this little bot back in the field.
Digging through the wiring from my original brain was a humbling exploration in to just how far I've come as a maker in the last three years. 2012 Andrew's solder joints look terrible, the original Beaglebone was fried thanks to a brutally obvious short circuit, and at least one coil on each motor was damaged beyond repair. This new Franken-2.3 has a new Beaglebone White, new motors, and I've hacked in the Tenda topside adapter to replace the old ETS Baluns.
But that's not all. In addition to the 2.3, you might also spot a hodge-podge of parts from 2.5s, 2.6s, and 2.7s This hybrid beast will form the heart of an Art Bot that I'm putting together for an ocean acidification outreach project that I'm working on,
I've always wondered how many of the original Kickstarter kits made it into the ocean. If you have one that you've managed to take to sea, let me know in the comments!
So here's the story behind our antler-based tether management system: Karyn came down to pick up Indy, at which point I realized that we had no good tether reels--Colonial Lee has the big PVC one. My grand plan was to eventually use old 3D printer filament spools for keeping tether neat and tidy on the older bots, but alas, we had no spare filament rolls ready to go.
So, after 30 minutes digging through my workshop, we came up with the only quick and dirty solution, an old rack of antlers hanging from the rafters.
Did it work? That depends on what you mean by work.
Did it keep the tether tidy during transit? Sure.
Was is easy to manage unspooling and respooling the tether during ROV deployment. No. Not at all. Awful.
Usability? None. 0 out of 10. Would not recommend.
We took Karyn and Indy out to Beaverdam Lake for a quick training run. This also gave us the chance to test out the GoPro quickclip mount. While mounted underneath, it did alter the flight characteristics of the bot and caused it to pitch downwards, but after 10 or 15 minutes of practice, we had the robot compensated. The GoPro hung on and shot some amazing footage of a completely underwhelming muddy lake bottom.
But at least we didn't get trapped under a log this time.
Add a GoPro to you robot!
My bot project for the weekend was to design a simple way to attach and detach a GoPro to the OpenROV. The end product, based on my Quickclip system, works pretty well, though I doubt I'll have any time for sea trials before the Tahoe Haxpedition next month.
3D printer files are here: youmagine.com/designs/quickclip-openrov-gopro-mount
Some OpenROV resources for you.
In preparation for the workshop I led last week, I put together this quick, 2-page cheat sheet to cover the basics of ROV operation, pre- and post-dive checklists, and minimizing transport of invasive species. Feel free to download and use for your own dives.
PDF is here: http://bit.ly/1C1Kmn7
Have bots, will travel.
We're on our way to Miami to meet up with the Hammerschlag lab for some OpenROV-based shark surveying.
In case you were wondering, the Pelican 1520 fits two OpenROVs and just barely stows underneath the seat in front of you.
And yes, this will be the big field trial for both the tether floats and the QuickClip system.
Locked and loaded.
Next week I'm heading down to Florida to run an OpenROV for Science workshop with students and faculty at the University of Miami. Planning out this workshop revealed two large gaps in my ROV toolbox.
A good training hook. Usually when I teach people how to pilot the robots, I tape a rod to the front and have them dive through a hoop and collect a target on the bottom. This always felt a little kludgey to me.
A line follower. We're going out on a shark tagging expedition to see if OpenROV can reduce shark stress and researcher fatigue by allowing us to inspect drumlines in the water, rather than hauling 80 meters of line every time we pass. For this, I want the dive to be as quick as possible and don't want to waste minutes searching for the anchor, so I want something that lets the robot clip in to an anchor line and follow it down.
With these two needs in mind, I spent a long, snow-filled weekend in front of the 3D printer.
Introducing the OpenROV Quickclip payload system! youmagine.com/designs/openrov-quickclip-payload-system
In addition to the two tools described, this systems also includes a strong, easy to install, and easy to detach system for locking tools to the payload bay (as long as you install to spanning rods in the front of the ROV). There's also a quickclip blank, in both .stl and .skp formats, for you to develop your own tools for the OpenROV.
So you want your own tether floats?
One of the things I love about the open-source movement is that, as long as you have the right tools, anyone can produce the things you develop. Case in point: the OpenROV tether floats that I've been working on for the last month or so. Now that they're tested, iterated, and vetted, anyone can take the shape file and print their own (and you should!): youmagine.com/designs/openrov-tether-float
I recognize that not everyone has access to a 3D printer and, over the last few weeks, several OpenROV users have asked me if there's a way I can send them some tether floats. I'm not set up to be a manufacturer (nor do I want to be) but I do want OpenROV users to have access to these little floats (which I believe are the best solution for managing tether buoyancy). Fortunately, OpenExplorer has a way forward that supports me continuing to develop cool tools for the OpenROV and provide a little support to the OpenExplorer platform.
For every $4 you contribute to the Forever Expedition. I'll print out and send you 4 tether floats.
The details: When you make a contribution, send me an email at southernfriedscientist at gmail, with your address and WePay transaction ID. If you want them sent out of the US, add an extra $5 for shipping. Unless there's a surprisingly large demand, I'd like to stick with multiples of four (that seems to be the most efficient arrangement for my printer). Color will be determined by whatever filament is loaded when I print, but if you order at least 12, I'll let you pick out the color (currently I have blue, green, and reddish-orange). At the moment, I'm only set up to print in PLA.
More tether floats!
I've been plugging away on tether float designs for the last week and I'm finally at the point where I can confidently say that the Forever Expedition has produced the finest 3D printed tether floats available for the OpenROV.
Let me talk a little bit about the problems we encountered with the first few iterations and what we fixed. If you'd rather just get some awesome new floats, go here: youmagine.com/designs/openrov-tether-float and download OpenROV tetherfloat v2.
There were two related "performance" problems with the first design: 1. it was too buoyant and 2. it was bulky. How can too much buoyancy be a problem for something designed to make tethers float? One word: spacing. For neutral buoyancy, the old design needed to be spaced over 5 meters apart. This is great from an economy of materials perspective, but that meant that there was a long line of tether between floats. This line sags, drawing the floats together and creating a nice big loop to snag on things. Not great. The floats were also bulky, they created a lot of drag and could, snag on submerged objects themselves. These floats are supposed to make piloting easier, not harder.
But there was an even bigger problem with these floats, they didn't print consistently. Because we're using only 20% infill on the printer, there's not a ton of supporting material for the final few solid layers (you want your floats to be mostly hollow, right?) The practical result of this was that the floats didn't print well in bulk on low-end printers--there were small gaps that let water in.
What's better on the new floats?
1: Lower profile. Less chance of drag and snag.
2: Smaller. Easier to space out properly without getting a saggy tether.
3: Printable. Even in bulk, with no gaps.
4: Looks cooler. If that's the kind of thing that matters to you.
I'm still working on getting the print fidelity up for those of us with low-end consumer printers. After quite a few attempts where the quality is just not consistent, I'm going to go back to the drawing board and tweak the design so that it can be printed horizontally instead of vertically. This will probably involve changes to the wire slots but shouldn't change the buoyancy too much.
So printing out the 16 pack turned out to be a little too much for my little Printrbot. Too many transverse moves and too little adhesion on the footbed caused two of the 16 floats to pop off during the print. I'll leave the file up for those with a high end printer, but I'll also test and load a 9-pack this week.
After a long weekend tinkering, tweaking, and burning through just a ridiculous amount of filament, it's time for the grand unveiling.
3D printable tether floats!
Sure, you can send your ROV down without tether floats. For short dives it's fine. But the more tether you pay out, the more it weighs down you robot. 100 meters of tether is a lot of weight for your robot to work against. And a tether dragging on the bottom is bad news. Not only does it slow you down, but it can get snagged on even small obstructions, making flying through murk water a pain.
I got tired of using fishing bobbers to keep my tether up in the water column. They have a huge profile, create a ton of drag, and the metal hooks can cut into the tether's insulation, compromising the signal. It seemed foolish to keep using them.
After looking around, unsuccessfully, for a better option, I figured "hey, why don't we just print some?"
These tether floats are small, slot into your twisted pair tether, and are easy to remove. When printed correctly, they are watertight and pretty buoyant. In freshwater, 1 float per every 5.5 metes of tether will keep you neutral. In salt water, you can space them out even further, to 1 per every 5.7 meters. I like to keep the tether up and out of the robot's way, so make the first 10 meters of tether positive. Since there's no metal hooks to catch, your can trim your ballasting on the fly without fear of shredding your delicate cables.
You can download the STL files from my Youmagine page here: youmagine.com/designs/openrov-tether-float
*note, these have not been tested for long dives in cold water, yet. You should go for it!
Rebuilding Tinz Lee
As I mentioned earlier, I have three OpenROVs, a 2.6 and 2.7 which have done quite a bit of travelling at this point, and a Kickstarter 2.3, which has... not... left the box... or been completely assembled. Having now built quite a few OpenROVs, I've decided it's time for a challenge.
I'm digging back into the old 2.3 to join an elite, exclusive group of explorers flying big blue, the original consumer OpenROV.
She's going to need a lot of work. I want to keep it as close to stock as possible, with a few minor changes. I'm replacing the old power block with a marine distribution hub and swapping out the Molex connectors for DB15s. Other than that, I don't plan on changing much beyond what was shipped until I get her up and running. Until then, enjoy my frustration.
Colonial Lee gets its first chance to dive in saltwater.
Despite a small problem with responsiveness in the port thruster and some ballasting issues, Colonial Lee performed flawlessly.
And, of course, you must respect the OpenROV prime directive: if there's a child nearby, you have to let them drive.
The time has come to assemble OpenROV 2.7.
I built my first ROV in graduate school almost 5 years ago. It was a rough job, cobbled together with PVC and bilge pumps. It pulled a massive tether that fed a CCTV camera and drew power from the surface.
In honor of my graduate school mentor, I named it Remote Lee.
Since than, all my underwater robots have been puns ending in Lee. To commemorate our move to the Williamsburg area in rural Virginia ( #GoatsAcrossAmerica), the newest ROV will be Colonial Lee. She joins Remotely Lee, Tins Lee, and Independent Lee in our tiny robot fleet.
There are as many different ways to be an explorer as there are places to explore.
From the hidden estuaries of the Chesapeake Bay to the skies around abandoned factories to the microscopic fauna in your own intestines, the Forever Expedition will travel, question, and discover. Why do we need this? Not every journey emerges from months of planning, logistics, and implementation, with clear goals. The Forever Expedition is different. Its only goal is exploration. One week we could be flying OpenROV's through a freshwater beaver lake in Virginia, the next we might be diving shipwrecks off the Carolina coast or geocaching in Baltimore City.
This expedition is always underway. The journey never ends.