First open water adventure post mortem


#1

Hi,

Almost after a year of anticipation I’m very happy to announce my first open water experience with OpenROV #128. I was very happy with the overall experience and wanted to document my initial experience with it while it was fresh in my memory.

Location

Things That Went Well

  • Portability
    • I was able to get the rov, tether, and laptop into two backpacks. This made it easy to transport the rov from the office to the car and then to the river.
  • • No issues with coiled up tether
    • When I packaged up the tether in my office into the backpack I rolled it into a tight coil. I was worried about the coil creating some sort of inductance interference since in my office the tether was typically more loosely laid out. Happy to report normal internet connectivity to rov when tether was in a coiled up state.
  • Happy with hollow core tether rope shield
    • A custom modification I’ve made to my rov is to wrap the network cable tether in a “hollow core polyethylene rope” . I’ve been able to use that outer rope as the mechanism for taking the weight of the rov when being pulled / deployed.
    • When the rov lost power and I needed to manually retrieve it by pulling the tether what I was pulling on was the rope, not the network electrical tether.
  • No one pestered me
    • I was able to find a location that met all my needs. One of those needs was not having random people coming along and asking, “What are you doing?”. When I have all the kinks worked out I don’t mind answering those type of questions, but when doing something for the first time and working through the kinks I prefer not to be disturbed.
  • Location
    • I liked where I was able to test the rov. I was able easily walk down to the water. The water’s depth immediately dropped off two or three feet and there was minimal current.
    • I was under the shade of a large tree so it made it easy to see the laptop screen and not be too hot.

Things that Didn’t Go So Well

  • Brownouts and not rebooting
    • This is my most serious “Ah bummer…” moment of working with the rov in the water. Multiple times, in fact always, under keyboard control the rov would brownout and not reboot until I physically retrieved the rov from the water and disconnected the battery leads.
      • Batteries were charged, I carefully charged them prior to this open water deployment
      • Battery end caps were pushed *all the way down* and snug so that both packs were running in parallel.
      • New propeller design will help with efficiency and power draw
      • Using proportionally controlled input controller (joystick / gamepad) should alleviate binary ON/OFF keyboard input control.
  • Negatively buoyant tether
    • The rov’s tether is negatively buoyant by default and thus is dragged on the riverbed and got snagged up on rocks.
    • I had wrapped the tether with hollow core polypro rope is to help offset the negative buoyancy of the base tether. The rope’s specification indicates that it “floats”, which it does, but it does not entirely offset the weight of the tether. Tether dragged along the ground and got hung up
  • Keyboard control is not realistic
    • By default trying to control the rov in the wate with just the keyboard is a bit frustrating. One, it draws too much power and causes brownouts as the mode is 0% or 100% power.
  • “Porpoise diving” while moving forward
    • This behavior can only really be seen in a large open space and was the coolest to observe.
    • When the rov moves in a forward motion the port and starboard thrusters engage.
      • When port and starboard thrusters engage in the forward motion it causes a large volume of water to be drawn in through the vertical thruster port from the top down. The water being drawn in causes a sucking motion of water and causes the rov to pitch downward in a diving motion (rov dives deeper).

Rookie Mistakes

  • Forgot GoPro #1 in office
    • I meant to bring another camera to the dive site and had it all charged up, but I left it in the office
  • Didn’t realize IP address was going to change when no internet connection
    • I had written a custom web socket service that by default listens to the computer’s default assigned address, BUT in the corresponding javascript that connects to the service I had hardcoded an IP address.
    • When the computer rebooted with no network available its IP address changed to a different value of what I had in the javascript file.

What I Learned

  • Don’t rush it, make sure you test the conditions you expect in the lab first
    • See javascript issue above.
  • Diving while under thrust
    • “Porpoise diving” was an interesting characteristic that I wasn’t expecting to see.
      • Solving the problem
        • Horizontal wings mounted to side of rov frame
        • Putting port and starboard thrusters (and nozzles) on outside of rov frame to avoid drawing in water through vertical port
  • Buoyancy
    • By default the rov was positively buoyant and floated on the surface, when it dove a foot or two under water it did not immediately surface. It resurfaced in a slow manner, roughly gaining altitude about 1 foot in 30 seconds.
    • Happy with the default buoyancy state of rov
    • Need to dive deeper for understanding deeper depth behavior. This initial dive only had a max depth of 3 feet.
  • O-Ring and Watertight Seals
    • 100% watertight seals are hard to do and there is zero percentage margin for error. I was having issues getting a seal in my test tank even at just under 6” of water, but once I got 6” without leaking I felt diving two or three feet would be alright…I was right
    • How do you test and validate O-rings prior to the dive? Until I know how to validate seals and o-rings deeper dives seem off-limits.
  • Laptop
    • I love my laptop, nice big screen, multi-procs, but what a battery hog. No more than 90 to 110 minutes available.
    • Either get more batteries or some sort of field charger
  • Silicone Spraying of Brushless Motors
    • Onsite right before the initial water dip I detached the props from the port and starboard motors and sprayed the coils with silicone water resistant magic spray.
    • This process still seems voodoo magic to me, I’m just doing it for good measure
  • External Accessory Premounting
    • I should have mounted the GoPro to the rov in the office instead of trying to do it onsite.
  • Video quality and streaming FPS
    • I’m not getting a smooth 30fps from the rov’s web cam so trying to control it just by the laptop’s screen is difficult. The current conditions leave it to a very jerky video stream.
  • Fogging
    • For me the electronics tube constantly fogs up and makes the onboard video stream worthless.
    • I should investigate using an anti-fog coating, like something you’d use for ski goggles.
    • Fogging of the video stream is a definite show stopper, if I can’t see where I’m going I can’t explore.
  • Onboard LED Light Glare
    • In addition to fogging in the electronics tube when the LED lights were engage their luminosity added a significant amount of back reflection glare.
    • People have suggested spreading the distance of the camera and lights, this would mean a new design though.


For Next Time

  • Pack more AA batteries for wireless joystick
    • I should make sure to bring extra AA batteries for the Logitech Freedom 2.4 cordless joystick. It takes three AA batteries.
  • Try out 3D mouse
    • The 3DConnexion mouse I have can be used instead of the wireless joystick and is a nice alternative because it is a wired USB connection so batteries are not a concern.
    • It has proportional control like joystick and works in three dimensions

#2

Marcus,

This is a great write-up! Thank you- this is the exact kind of thing we can use for improving the ROV.

I think this deserves a very thorough response, but here are a few quick things that might be good to say:

1. Brownouts: if you only have a keyboard, I'd recommend only using throttle setting "1" or "2" (which you can select by pressing that respective number on your keyboard before the dive). We've encountered these issues as well, and are currently working on some ways to mitigate them with both hardware and software. I know that can be very frustrating.

2. Tether buoyancy issues: using polypro rope is a good idea, but as you found, it may not be sufficient. We're looking into alternate flotation that can be added to a tether when flying over rough terrain. Also, the homeplug adapters seem to work with even very thin wire (we've tested down to 30AWG!) so that may be something worth trying to keep the weight down.

3. Porpoiseing: Yeah... we've noticed that as well. It seems that this may be tuneable using buoyancy weights on either of the two cross bars.

4. Testing seals: I use a vacuum pump with a suction cup adapter and try pulling a partial vacuum through an open vent port after the end caps are in place (but remember to put the vent plunger back in before diving!) If the pressure holds, then I've got a good seal. I would also recommend visually inspecting that the O-rings are fully engaged before a first dive after the tube has been opened.

5. Fogging: we started using descant packs inside the main tube, and so far they seem to work great! Also, if you get a leak, they tend to absorb a pretty good amount of water and keep it from sloshing around too much (which may get the electronics more wet then otherwise)

6. LED Glare: some of this may be due to the plastic blinders being somewhat translucent. It may be worth trying to black them out and put some felt between them and the inside of the main tube.

Well, those are my initial thoughts which I hope are a little bit helpful. It was great to read your write up, and please keep up the good work and let us know if we can do anything to help you with your adventures!

Eric


#3

Hi,

If you haven't got access to desiccant packs you can make your own out of cat litter... ;-)
Found this helpful video on youtube, http://www.youtube.com/watch?v=ly-o7sAsFtI

Best thing is that you can make it into a shape that will benefit the shape of the E-tube.

Peter