Sunday, October 4, 2015

For real this time

We took the ROV to the San Marcos River today at La Isla. This is sort of how it went, though it doesn't tell the whole story.

The reality is that for most of the time we could only see a blue abyss, even in bright daylight and clear water.

By the end we had collected some water inside the main tube and called it a day.

Steering was still a huge problem. The ROV always wants to go left. No matter what. We have swapped motors, propellers, tried inversion settings and updated the software to try and solve this problem to no avail. Our newly joined member Evan suggested we try a fixed-position rudder. We may just do it. I am willing to try almost anything given how bad this neutrally buoyant brick flies.

San Marvelous

Road Trip!  DannyE drove us down to San Marcos, TX to test the ROV out in the crystal clear water of the San Marcos River on Sunday.  It went surprisingly well, as far as things for the ROV project are concerned.  First of all, the day was gorgeous:

The dive staging area

We set up shop on an island in the middle of the river, by Rio Vista park.  The island is accessed by a small concrete bridge - carrying a bunch of electronics across that bridge was fun stuff.  The ROV may be waterproof, but most of our gear certainly isn't.
In any case, after setting up and and clearing our pre-dive checks, we chose to launch from this choice looking spot:

Our launching point - a tiny pier and some beautiful water

The goal was to dive in the still area between the fast flowing water on the southern side of the island and the slower (yet still powerful, compared to the ROV's motors) current on the northern side.
And dive we did!

Thar she blows!

Safe to say the San Marcos River was far more powerful than our little robot.  We explored the immediate environs but were swept away by the current repeatedly.  However, we seemed to draw quite a few curious onlookers, including human, piscine, and testudine ones.

Oh, and we saw a really special creature breach the water at one point:

Oh hi

We took a lot of video both from the ROV and the GoPro attached to it.  We'll be editing video and throwing it up here soon so that everyone can see what we're seeing while the ROV is underwater.  We think you'll like it.

We are also really looking forward to diving in some calmer waters more suited to the ROV's power level - namely, lakes like Lake Buchanan.  We are in the planning stages of an exhibition there soon, to look for remnants of the lost city of Bluffton.  Whether we find it or not, we're finally able to use the ROV out in the real world and it feels pretty darn good.  More to come, soon!

Wednesday, September 23, 2015


Welcome back!  We have updated the blog name to reflect the new mission of the group.  Just because we've moved on from our original goal of exploring Jacob's Well, our focus on doing citizen science and exploring with the OpenROV project is still alive and well.  We're exploring a lot of interesting possibilities for underwater dives in the Austin area, including underwater ghost towns in the Highland Lakes chain, and are excited to start planning some of these expeditions out.  We plan to continue blogging about both the technology challenges we encounter with the OpenROV platform as well as detailing any finds we make along the way.  Thanks so much for reading the blog so far and we hope you continue to follow us on our underwater journeys.

Monday, September 14, 2015

Three good tests

The OpenROV team says "you can build it in a weekend" but to us the reality was much closer to 18 months. I think this is an example of how "dreamy" the OpenROV folks are. They have very lofty ideas but execution always leaves something to be desired.

A short example, the next version up from 2.5.1 is version 30.0.0. Maybe they fudged the version number, something that could be corrected, but they went ahead with it. It isn't all that revolutionary or a silver bullet for the 2.5.1 problems, just an incremental change.

Ah, the perils of being an early backer of an open source project. The upside is that things improve all the time, like software. The downside is that as things improve, you are stuck with the "old world" hardware and software you started with. We bought LiPo batteries for our 2.6 kit, and very soon after we found that there was no way to use them, the OpenROV team switched to a new LiFePo battery type requiring all new ROV geometry and chargers. We decided to stick with LiPo and went with a much higher current 26650 battery.

Finally, we are ready to actually dive our project OpenROV. It doesn't fly perfectly but we think it is about as good as it can get with this platform. We've fought software, firmware, hardware and collaboration issues to get here. Now that we're ready, the place we thought we were going seems to have slipped further out of reach.

Sadly we probably won't be going to Jacob's Well. This is partly because we can't rely on the ROV being able to get back out of the well and the JWEP makes our efforts redundant. It is also worth mention that we couldn't get ahold of anyone at JWNA or Wimberly Valley Watershed to even discuss our project.

Sunday, July 26, 2015

Don't use SUDO while underwater

This will be a text only update.  If you're wondering why - well - somebody erased all the latest GoPro footage (that would be me, James.)  We had a nice little test dive going this past Saturday - a tune up dive to tweak the controls (they're still difficult to use) while we wait for the summer swimming season to die down and for someone/anyone at Jacob's Well to respond to our inquiries.  Early in the day, I noticed we had already filled up the GoPro storage - I had neglected to erase the last few dives worth of footage, so I decided to format the card and just capture more footage for the rest of the day.  Great plan, right?
Unfortunately, we got no more footage that day.  There was a leak and the ROV started going nuts on us.  We thought perhaps our code tweaks to the motor controls tanked it, but it wasn't actually errant sudo usage but a bad o-ring that sank us.  Anything that can go wrong on this project will go wrong.
After letting the ROV dry out, it was working fine again, but we were done for the day.

So next up, if anyone reading this knows someone that can get us diving access to Jacob's Well, that would be amazing.  So far, we've gotten nowhere with them.  I'm not above just driving there with the ROV in the offseason and diving it, but I'd rather do this with the cooperation of the county and all.  We're as good as we are going to get piloting this thing - it's time to dive.

Tuesday, April 14, 2015

Bout Time

Well...we finally got this robot working.
I can't quit you

If I had known...if we had known...before starting, how many difficulties we'd have in just the building  of the robot, we'd have been stunned.  The OpenROV team had me hooked way back when, with their kit that "can be built in a weekend."  Here we are a year later, finally getting to that point.  There have been a lot of lessons learned in this endeavor and I won't detail all of them here, since we've been covering them throughout the course of the blog, but a few things are worth touching on:

  • Hobbywing ESC's are a pain in the arse
  • Danny E likes to cut, burn, and hammer on things
    • But he makes them work
  • Piloting a robot in the real world is NOT just like a videogame
  • Building robots is even more fun than you'd think, when all is said and done

We navigated a little homebrew obstacle course in the pool to do some pilot testing and get a feel for how the OpenROV will handle in tight spaces.  It's a little scary how tough it can be to pilot in the pool; some of that is due to the currents being pumped around, and the relative lightness of the ROV.  But it's also just tougher than it looks.  We bonked the GoPro on the obstacle course more than once, and sometimes even getting our bearings straight was a hassle.  But we we able to fit the ROV through some tight spaces and we think we're ready to do this.

Next step is to make this official with the Jacob's Well people; we are hard at work on securing a date and time to take this little guy out there and see what we can see.  It's finally coming together!  

Saturday, February 21, 2015

The best ESC in China

Oh, the pain you have caused me, hobbywing.

I declare that Hobbywing is *NOT* the best ESC in China. At least not all of them are.

We've replaced these already with identical parts only to have the same problems again. Now the trouble seems to be one ESC losing calibration.

Our little ROV has been through a lot. Just getting it to the test dives has been a huge challenge. We have seen software, firmware and hardware problems in and out of the water. Operator error has contributed. Our group has become fatigued with the motor issues.

This motor had failing bearings that were potted in, requiring replacement.

Richard potted a replacement motor and then we damaged one of the brand new motor bearings.

This probably happened while removing the c-clip that holds the shaft in place.


Arthur provided some much needed new bearings. Danny soldered the motor onto the ROV. Alas, there was still no joy to be had. Now the motors are behaving unpredictably again.

After removing power via the battery tubes some, all, or none of the ESCs lose their calibration settings (they all kept their programming, so yay?). This is evident because the motors will then not spin backwards. Re-calibrating the ESCs works and stays fine through reboots but does not survive pulling out the batteries.

More and more the Blue Robotics thrusters are looking like our next steps, probably even with their ESCs. but we still plan to limp the ROV out for another test dive very soon.