Build Log: How NOT to build underwater drones


Hi Everyone!

Once again, welcome to the Build Log! This is a continuation of the previous story (link to previous newsletter releases), where we showcase all of the ROVs we created before designing the CPS 5 (the drone that actually works well). After you see all of the prototypes we made that failed, you'll know what to avoid in your own ROV journey.

Fourth prototype

After we tested the third prototype of our ROV (mentioned in the previous newsletter), which was very much a proof of concept, we decided to make an ROV which is very similar in the overall design but much prettier (the previous prototype was very boxy, as you may remember). We ended up designing the best shape we could at the time. Again, this was supposed to be a fully waterproof, 3D printed body, which is notoriously difficult to do.

Going back to the design process, I think we've done this the wrong way around. Instead of designing the drone around the components that needed to fit inside (electronics, ballast), we designed a body which was "large enough". We'd figure out how to put all of the electronics inside later on. Because of this, packing all of the electronics inside of the drone was really difficult - as you can see on the picture below:

We were so determined to make this version work that we forgot that the design might have been flawed right from the beginning. We've made about 3 prototypes of this drone. The last 2 prototypes worked okay, but minor repairs and "drying" the inside of the drone needed to be performed between each swim. It was still much better than all of the previous ones.

We also noticed a much bigger issue - the "3 thruster" configuration of the drone had a lot of disadvantages - the drone wasn't nearly as controllable as we thought it could be. The main issue was that with this shape it wasn't really possible to get the center of mass below the center of buoyancy so that the drone could stay level in the water. Here are some pictures of the prototype that worked the best:

Fifth prototype - The wooden drone (Deskodron)

So once again, a little frustrated with the fact that we couldn't get the previous design to work - we decided to change the design completely. We decided to eliminate the fancy shape and use a very simple shape made from commonly available components (the frame was made from wood). And we decided to fix the control issues by using a 5 motor configuration.

For the electronics, we used 2 very simple 3D printed boxes, which were located in the front (the battery) and in the back of the drone (contol electronics). This setup worked very well, with only very minor leaks at small depths. This prototype was the biggest success yet. Most of all, the self-leveling worked flawlessly - this time the drone was rock stable underwater. Here's the video from testing this drone (it's one of the first videos on our youtube channel):

video preview​

CPS 5

Then, we started designing what we later called the CPS 5. This drone was supposed to be everything that the previous prototype was but it was supposed to have "the fancy shape" and "the fancy electronics system". The first concept looked like this:

Only when we started the design of the CPS 5 we came across the ArduSub project, which really simplified a lot of our electronics work. We almost didn't have to write our own software for the drone. We created about 5 different CPS 5 prototypes before we released the CPS 5 course that allows you to build your own underwater drone.

Here's some issues that we faced along the way:

  • Battery managment systems failing - switching to balance charging the drone between missions,
  • Endcaps leaking, water leaking through cables - spent a couple of months developing a reliable endcap design,
  • Electronics not servicable - integrated the Molex connector (for unplugging the electronics quickly),
  • Drone negatively buoyant - solved this issue by using some buoyancy foam and making the electronics pipe a little longer

Here are CPS 5 prototypes (all of them are different builds):

We definitely learned a lot here, and our later designs were going to get even better :) If you want to get that knowledge as well and build the CPS 5 yourself, check out the online course course here:

Cheers,

Piotrek, CPSdrone

PS: You can access all of the previous releases of the newsletter here.

The Build Log by CPSdrone

Subscribe if you already are following us on YouTube!!

Read more from The Build Log by CPSdrone

Hello Reader, Welcome back to The Build Log! We're almost ready to take the submarine on its mission under the sea, to explore a crashed WWII B-17 bomber. In this newsletter, we'll show you the tests we did before the mission, to make sure our underwater localization works well enough. Quick recap - our localization system is based on an external compass in the AUV's antenna - that provides the current heading (along with sensors in the Pixhawk), and a DVL (at the bottom of the submarine),...

Hello Reader, Welcome back to The Build Log! Last week we showed you the upgrades we made to our autonomous submarine. This week, we'll show you the actual mission we're going to be attempting with the submarine: We want to take a picture of a crashed WWII B-17 bomber. The wreck sits at the bottom of the Adriatic sea at around 70 meters: Because of the depth the plane is at, diving there is very hard and requires advanced scuba diving certifications. The mission we are attempting is also very...

Hello Reader, Welcome back to The Build Log! We are preparing our autonomous submarine for a very exciting mission in the Adriatic Sea, and in this e-mail we'll show you all of the upgrades we have made to the submarine to prepare it for the mission. 1_ Double thruster upgrade As you might have read in one of our previous newsletters, we had some problems with the submarine turning to the side while flying at high speed. This happened because of a phenomenon called "torque roll". The one...