Exploration Electronics (Trident Kickstarter Update #14)

Original post : July 13, 2016

Hello everyone!

It has been another intense month for us at OpenROV and once again there is a lot to report. In our last update, we told you about a plan that we had in the works to use a modified 2.8 OpenROV running experimental Trident software and using a Trident camera to explore the shipwreck of the SS Tahoe steamship. That went very well and I’m extremely excited to tell you more about it. As a testament to how tools like OpenROV can influence the Citizen Exploration movement, the Tahoe expedition made it into the New York Times as well as into numerous other news and media sources. We also got some incredible footage of the wreck with our new camera… but let’s get into that later.

First, I want to geek out on the recent technical developments that we’ve made with Trident’s hardware.

Computer

The big news: We have successfully ported OpenROV Trident software to work on Raspberry Pi and have selected the Raspberry Pi 3 as the main computer onboard Trident. This is especially big news for developers. The Raspberry Pi community has over five million users worldwide, and it is by far the most popular single board computer for hobbyists. That means that we will be tapping into one of the largest embedded system development communities in the world.

Alt Text
Raspberry Pi 3 Model B

Additionally, Trident will now have a lot more computational horsepower. The Raspberry Pi 3 sports a 1.2GHz quad-core ARMv8 CPU, 1GB LPDDR2 RAM, embedded WiFi, and excellent peripheral support. Compared to the Beaglebone Black (which we’ve used for all OpenROVs up until this point) which has a 1GHz AM3358 processor and 512MB onboard DDR3 RAM, that’s a significant step up in performance and capability. Just after the Raspberry Pi 3 was announced, we began to research whether it would be possible to port our existing OpenROV software over to the Pi. After some final testing last month, including testing done during our Tahoe expedition, we made the decision to move our system over to the new architecture, and we couldn’t be happier. Having multiple cores gives Trident much better video and data processing capabilities which will continue to come in handy as we release software updates. The built-in features will make the entire system more solid from the get go. Moreover, Trident will be (in our opinion, at least) one of the coolest devices that uses the Raspberry Pi. We’ve designed the Trident software to include an integrated development environment so people wanting to write their own plugins and modifications can edit their code directly on the vehicle through their web browser with no additional software needed. Fueling a strong developer community has always been core to OpenROV, and we think our integration of the Raspberry Pi 3 will move us even further in that direction.

Motherboard

Another exciting milestone that we reached recently is that the first prototype motherboards for Trident have arrived. Walt Holm adapted many of the features from his 2-series controller board design into this system, and we’ve added a lot of other custom subsystems for Trident’s particular use cases. The prototype board is designed to wrap around the vertical thruster cavity and will be pinned in place between the two main parts of Trident’s shell. This first prototype only has basic systems onboard. We still need to layout our custom ESCs (which will be optimized for ROV thruster behavior), power management, and communications subsystems, all of which are being completed now.

Alt Text
Raspberry Pi 3 on prototype Trident Motherboard with camera and ESCs

Hardware Testing

On the hardware side, we’ve been doing a lot of iteration before turning the key on injection mold tooling. We flew the latest prototype of Trident -complete with internal lattice pattern, Trident motherboard, custom Trident battery packs, lights, and camera- for several hours during our Tahoe expedition and we were really happy with its performance. Later on, we did a dive to just over 100m and noticed a leak. When we recovered the ROV, the prototype urethane cast shell had held up perfectly but it seemed that one of our passthroughs or o-ring seals had failed. We had hand-sanded the o-ring grooves since the urethane casting process had left some imperfections, so we think a dimension being out of tolerance as a result of that may have caused the leak. Although that failure mode would not transfer into the final injection molded version of the design, we’re carefully reviewing the failure to make sure that we understand what happened before moving forward. That being said, the fact that we were able to take a urethane prototype that deep without material failure in the shell is extremely promising. We can’t wait to do more deep water testing with our prototypes. If you’d like to see video of the earlier shallow dives we did with the Trident prototype (all filmed with a v2.8 ROV we had), we’ve posted a 2-minute YouTube video of that as well as a recording of the video stream coming from the Trident itself.

SS Tahoe Dive

We also created a highlight reel of our dive on the SS Tahoe. This dive was made with a modified OpenROV 2.8 with external lights, using the same camera we will have in Trident. What’s more is that the entire dive was live streamed over the internet as we explored the wreck using beta software for Trident. We accumulated more than four hours filming around the wreck as well as additional time spent exploring a nearby debris field. It was an incredible experience, and I’m extremely excited about watching other people’s dives happening around the world in real time when thousands of Tridents can do the same thing. For more information on how the expedition went as well as some history of the wreck, check out our OpenExplorer page for the expedition. Below are some screen grabs from our favorite moments during the dive:

Alt Text
Bathroom sink filled with sediment. We entered this room from a hole in the ceiling of the ship where the roof had been torn off.
Alt Text
Bow of the SS Tahoe with forward davit (ship’s crane) visible. The Tahoe rests on a 30-degree slope.
Alt Text
Stern of the SS Tahoe. “TAHOE GLENBROOK” can clearly be read. Glenbrook was the town in Lake Tahoe that the ship was launched from.
Alt Text
Wooden door leading out of one of the ship’s upper-deck rooms.
Alt Text
Deck hatches. Some glazing still in place.

You can find the full-length recordings of our dives on the Tahoe (as well as other OpenROV footage) on our YouTube Channel. Also, check out our Facebook Page for links to other press we've gotten from the expedition.

Experimental Trident Camera

On the topic of HD video streams, we’ve got more exciting news: later this month, we will announce an upgrade kit for the 2-series ROVs that will allow them to use the HD Trident camera that we’ve developed. Not only will this system provide extraordinarily good video quality, but the camera’s built-in H.264 compression encoder will also enable in-browser recording (no software will be needed) and eventually live streaming just like we did during the SS Tahoe dives. This system is very experimental. The main point in creating the upgrade is to allow us to get some early feedback on how the system works, but if you’re dying to try out the camera, the 2-series kit can now support it. We will be selling the experimental upgrade in our store for about $300, but if you’re a Kickstarter backer and would like to try it out, please send an email to [email protected] with the words "Trident Camera Discount" in the subject heading and we’ll give you a discount.

Alt Text
Prototype HD Trident Camera upgrade for 2-series OpenROV E-Chassis

International MATE Competition

To finish this update, I’d like to take a moment to talk about education outreach. From day one, OpenROV has been about connecting people who want to explore, while developing the resources to make exploration something that anyone can do. Many of us at OpenROV got our start with ROVs in high school as part of the international Marine Advanced Technology Education (MATE) competition. For more than a decade, MATE has brought young people interested in exploration technology together to develop technology with humble resources. We at OpenROV owe a lot of our experience and knowledge to MATE, and this year we were honored to visit their competition at NASA’s Neutral Buoyancy Lab in Houston, Texas. We helped film and livestream the event using one of our ROVs and we also teamed up with the Schmidt Ocean Institute to gave away an OpenROV 2.8 to the first place team of the competition. More than 600 teams from around the world competed in the MATE competition this year and it was incredible to see what these students came up with. If you’d like to know more about MATE (and we really encourage you to check them out, especially if you have kids) you can visit their website at http://www.marinetech.org. You can also watch our entire recording of the event - filmed with an OpenROV 2.8 during a single five-hour-long continuous dive.

Open House

As always, we are extremely grateful for the support we’ve received from our community. We hope to learn as much as possible about the 1,324 of you who backed our project, and we hope that all of you can get to know us as well. To that end, we’ll be hosting an Open House at OpenROV HQ in Berkeley on July 16th, and we encourage anyone interested in seeing the project in person to come. If you’d like to visit and haven’t already RSVP’d, you can do so here. If you can’t make it, but would still be interested in telling us a bit more about yourself, we are also planning to send out a survey pretty soon with questions about your interests in the project and how we can support those interests. We hope to send that out in the next few months so keep your eyes peeled!

We are striving to include everyone in the process of developing what we believe will be a revolutionary exploration tool. Together, I hope we can create something amazing.

~Eric