Small Details Make Big Differences (Trident Kickstarter Update #16)
Original post : November 4, 2016
First off, check out the segment CBS This Morning put together about OpenROV and the incredible community of people who have gotten involved. We’re extremely honored to be growing this movement with all of you, and we believe that this is still the beginning of an amazing chapter for underwater exploration.
October was a busy month here at OpenROV and we're excited to be taking a moment to share everything that has been going on.
We've been placing orders for many of the sub-components that will be used to build Trident. Now that we're close to the point that we had originally wanted to ship, we believe it is important for us to keep everyone updated with where we are in terms of timeline. So far, there haven't been any major issues with our production plans, but there are some smaller delays that may add up. We'll be getting more explicit lead time information and making decisions based on that information in the coming month. We'll aim to have more concrete information about timeline in our next update in December.
We often write updates about major milestones, but in this update we’d like to focus on the incremental developments we've been working on throughout the project - the kind of things that are often discounted but actually have a huge impact on the overall quality, performance, and feel of a product.
We spend a lot of time at OpenROV developing ways to make operating Trident as intuitive and enjoyable as possible. Many of us have spent years designing and operating industrial ROVs, so we have a lot of opinions on how to do it right. When you are out in the field - particularly in harsh conditions - there are a lot of moving parts to manage, so every detail that improves user experience makes a difference. Here are some brief descriptions of just a few of the aspects of Trident that have received a lot of focus:
Although the size and shape of Trident may seem fairly straightforward, a lot of time and thought was put into getting the dimensions just right. Overall, it's size was constrained by wanting it to be big enough to hold the necessary on-board electronics and typical payloads that might be mounted to it while simultaneously wanting it to be small enough to be easily hand carried in a case or a backpack along with all of its support equipment (tether, computer, gamepad controller, etc).
Having a shorter height reduces Trident’s forward and aft-facing surface area, thus minimizing drag in the forward direction. At the same time, having minimal height allows the vehicle to be deployed in very shallow that would be much more challenging to deploy in with larger or taller vehicles. We also wanted the height of the sides to be small enough so that the vehicle could be gripped from the side.
Having a relatively large width provides several advantages. Externally, increased flat, downward-facing surface area allows more room for payloads to be mounted to the vehicle. Internally, adding width allows us fit electronics on a single internal printed circuit board which improves reliability. Finally, having a large width increases drag in the vertical direction to induce pitch during vertical maneuvers (which we'll get more into later). The upper constraint for the width had to do with a relatively obscure objective which was to make Trident fit down standard ice-drill holes. The original concept for Trident was developed during a three month expedition in Antarctica to study ecology under the ice, so sub-ice exploration has always been part of the soul of Trident's design. The largest common ice drill (used by field teams in Antarctica and ice fisherman around the world alike) has a 10 in (~25cm) diameter, so keeping the width less than that (with some margin for freeze-in rate of the hole) would allow under ice operations to take place without needing to spend tons of time cutting a hole manually with an ice saw.
The length of the vehicle was constrained by wanting the case Trident will fit into to be small enough to fit under the seat on an airplane. Worldwide, most airlines have under-seat spaces with a width of 20in or around 50cm, so a 40 cm length would leave enough room for padding.
Seawater and freshwater have different densities. In order to make an ROV that is neutrally buoyant in freshwater become neutrally buoyant in seawater, you have to change the effective density of the vehicle either by adding mass or subtracting volume (and adding weight is much easier). The difference in densities between sea and fresh water is about 2.5 percent, so for our 3.3kg vehicle, we needed to find a simple and elegant way to increase the mass by about 85g in the field for seawater operations. Originally, we considered adding metal plates to the bottom of the vehicle, but attaching those metal plates reliably would require fasteners, and fasteners would either require tools or they would need to be designed so that they could be tightened and loosened by hand. An old adage we like to repeat at OpenROV is that a designer knows he has achieved perfection not when there is nothing left to add, but when there is nothing left to take away. In the end, we realized that we could make the fasteners themselves be the weights. By making custom trim weights (two, with one on each side of the vertical thruster, directly below the center of gravity of the vehicle) we could reduce part count (making it easier to keep track of everything), reduce time to add or remove the weights, and eliminate the need to use a tool to tune for fresh or seawater. The weights are made from 316 stainless steel so they won't oxidize over time, and the threaded inserts they go into are also 316 stainless so no galvanic corrosion will occur.
In addition to our considerations about the dimensions of Trident, a lot of thought also went into its shape. We wanted a design that would be extremely hydrodynamic in order to allow for longer run time, faster speeds (which translates into better ability to penetrate through current), and excellent tracking along transects. All of these things can be improved by reducing drag. In fluid dynamics, a term known as "Reynolds Number" is often used to describe the relationship between dynamic and viscous forces as something moves through a fluid. Without getting too geeky about things, it turns out that in situations with low Reynolds Numbers (such as a small ROV moving through water), drag comes mostly from the trailing edge of the object, not the leading edge. In fact, the difference in drag between a shape with a sharp vs blunt leading edge is almost negligible. For this reason, one will notice that the leading edge of Trident is essentially flat to allow for a large camera window, but every feature in the tail of Trident is tapered down to a point. Additionally, the sides of Trident are designed to act as vertical fins which work like feathers on an arrow to keep it pointing in a straight line when moving forward (which also helps the vehicle fly more efficiently).
Finally, one of the most notable hydrodynamic aspects of Trident is its off-center vertical thruster. By placing the vertical thruster slightly forward of the lateral center line, we are able to take advantage of the fact that drag grows exponentially with speed. When Trident moves its vertical thruster slowly, the righting moment of the vehicle keeps it pretty much horizontal as it ascends or descends. When moving quickly, however, the higher accumulated drag aft of the vertical thruster induces a pitching moment and allows Trident to point in the direction it is moving vertically. The end result is that users can carefully inspect vertical surfaces like dock pilings and rock walls when moving slowly, but can rapidly move up and down the water column and follow the contour of the bottom when moving quickly. Flying Trident feels more like flying an airplane than controlling an ROV.
One of the most often-thought-of places for UI to be considered is with on-screen interfaces. When we started developing our OpenROV kits back in 2011, we created the design around the idea that the vehicle would be completely digital and would require no custom topside control console- just a common laptop computer would be needed to view live video and telemetry from the ROV and tell it where to go. Today, using a computer (or mobile device) to control something seems pretty intuitive, but five years ago - when embedded systems where just barely powerful enough to process video - going completely digital was a radical idea. Many people from the industry encouraged us to stick with analog video, but we had a strong feeling that the future of ROVs would be completely digital, and our decision to go that way has really paid off with image quality and video processing capability.
Another concept we started developing early on with our kits was the idea of hosting the interface for the ROV as a webpage from the ROV itself rather than developing separate software for the controlling device. Hosting a web page from the vehicle does not require internet, so even though Trident uses your web browser as a control interface, no internet is needed to use the system. Our web-server based architecture has tremendous advantages for fieldwork that are very apparent to us, but that may not be as obvious to others at first glance. Here are some of the advantages:
System Agnostic Interface: Since the only software needed on a device to connect with the ROV is a web browser, our interface is system agnostic- it can run on a mac, pc, iOS, or android device, all with the exact same software. This means we can develop faster, track bugs much more easily, and deploy updates as soon as they are ready.
Versions Always in Sync: Users don't have to worry about making sure the software on their device is a compatible version with what's on their vehicle, and when they update their ROV, the interface automatically gets updated with it.
Ubiquitous Redundant Controllers: Being system agnostic and requiring only a web browser means that any laptop or mobile device taken out into the field can be used as a redundant system for controlling the ROV. If you're 300km off the coast of Fiji controlling Trident with your iPhone and the phone's battery dies, someone else with an Android can pick up where you left off, and they don't need to have anything downloaded for the system to work. If you want to use the ship’s laptop instead, that will work too. Almost all mobile devices out there today come with Wi-Fi and a web-browser built in, so whatever device is around should be ready to work with the vehicle without any prior scenario planning required.
Be sure to check out the Kickstarter page for more updates about Trident's propulsion system, materials, and certifications (backers only)!
The coming month will be a very informative one. We plan to be getting a lot of new information about manufacturing lead times, compliance testing, and assembly process validation, and we’re also planning to get our hands on a lot of first article production samples of subsystems for Trident. Our strategy has remained the same as when we started: Work expeditiously to make the product solid, but don’t cut corners to rush something out the door. We hope that what we ship out will reflect the quality and attention to detail we’ve been focused on, and we can’t wait to hear what people think!
Finally, in case you missed it, we did a Kickstarter Live AMA (Ask Me Anything) earlier this week and answered many of your questions there. The format was fun and we’re planning to do more of them in the coming months. Stay tuned, and as always, let us know in the comments if you have any questions.