Our first wet test ever!
Yesterday afternoon, after a few dry tests performed in the safety of our houses, we had our first wet test ever. Our friends Henrik and Sophia, who live on the shores of lake Mälaren, hosted our team for the whole day and provided the much needed rescue boat. Performing the test was for us the closing step for milestone one. You can enjoy the gallery of pictures from the event here.
The goal of the test was to set a reference point for our future work. All of the our rig, electronic systems and software was brand new and never tested in a real-life situation, so we simply wanted to see which problems would have emerged by setting the boat afloat. We had the chance to perform two launches of Humble Floaty and both of them proved to be ripe with learning, although none of them ended up as expected.
The weather was fine, with strong south-easterly wind and calm water. Both sails were done with the boat controlled via it's native RC (onboard software was used only for measuring heading and wind direction and for logging purposes).
Chronology of the first sail:
- The boat didn't manage to leave the jetty against the wind, so we launched it from the leeward side of the jetty.
- The weight of the wind sensor, mounted on a custom mast on the stern of the boat proved to be too much for the boat in the rather strong wind of that day, so the boat completely laid on its side every time a gust of wind hit the sails. As we completely sealed the hull, the boat didn't fill with water and was able to right itself up every time.
- After a few flips and flops, the wind sensor mast fell off the boat and began trailing behind the boat, with the cable wrapping around the rudder.
- On a gust of wind, the sail rotated 180° around the mast, entangling the string used to move the sail around the mast. At this point we lost control of the boat and we had to send out the rescue boat.
After having recovered the boat, and removed the wind sensor completely, we proceeded to a second sail. This attempt has a less catastrophic beginning than the previous one, although the test also finished with the sail flipping 180° and the boat being drifted outside the RC range.
When we recovered the boat, we discovered that about the hull was half-full with water. Later on we were able to identify the cause of the damage: the board onto which the wind sensor mast was mounted, was kept in place by being "sandwiched" between the electronic box and the boat deck. With the rolling of the boat, the force exercised by said board on the deck were too strong, and the plastic of the hull cracked.
Here's a brief summary of the lessons learned from the test session, in no particular order:
- Cheap RC models are not a good prototyping platform. It is very difficult to find a sensible arrangement of the electronics, they are fragile, and their keel isn't able to counter-balance the weight of the electronics even in moderate winds.
- The design of the wind sensor must be way lighter than it is now. [The bulky construction of our first prototype was however more of an accident than of a design choice]
- We must force ourselves to call for software "feature freezes" som e days before a test. As it happened, we tried still 12 hours before testing to squeeze-in a couple of features that - not sufficiently tested - ended up impairing the ones we relied on.
- We need much more righting moment on the boat. This will probably be achieved with a combination of larger hull, lower centre of gravity, and heavier keel.
- We need some sort of screen lock on the FreeRunner. When rigging the electronics box is really easy to inadvertently touch the screen of the FreeRunner and alter the normal functioning of the software.
- We must limit the range of movement of the sail. The current control system is such that strong wings will manage to force the sail to an unnatural and non-recoverable position.
We are planning to introduce these changes in our next boat whose codename is Salty Breadboard.