We still needed to add leaves, Eva had been working on some nice hard acryll plastic leaves, they looked quite good, now the challenge was to attach them. We had to lasercut again, make the final arms more fitting. We did that, it was alot of work to properly attach everything and put all those tiny screws back in place. But the fina result was one again an improvement. The biggest challenge however would be the extra weight taxing the opening and closing mechanism. For that we really did have to fix ourselves a stronger servo engine.
So we brought all our gear to Hengelo and started putting it all together, messing with the code a bit and solving some minor problems. We drilled some holes and added nice speakers. After our pitch we heard from Renée that he really wanted to see more interactivity, in the form of a pir specifically, I must have put this idea in his head by making a suggestion myself. This was just a little idea however and with time and energy being low his preference for adding a pir was somewhat unwelcome. However, I’m kinda glad I went to school to fix that pir thing even though I already had one hell of a long day. That last few hours in the lab turned out more fun than expected.
Finally after doing extra work on reinforcing the design it was strong enough to bring it all together. Ids got busy on soldering the lights into the head while I thought of a clever way to attach the tip to the serva base. We didn’t wanna do this too permanently so I decided we should drill holes in the pipes and just let bolts hold the design in place.
It made for a very stable firm yet adjustable solution. While we brough the lights, the ultrasonic sensors and the servo together it was time to make the sound work. After a long struggle it turned out we just need to use two arduino’s. One master one slave.
The slave had the mp3 shield detached and only needed to start playing on the master’s command. A nice and viable solution. The ultrasonic sensor gave trough many strange values though, it was still a struggle requiring multiple professionals to implement a system that properly deals with these odd values. But we made it.
In the end it was all together and neatly working. Ofcourse we still had to implement Luca’s photon, but that was no problem really, we needed some help, but the worst was in the past already.
Now we had to bring it all together at the workplace.
In order to properly run tests to combine the mp3 shield with the ultrasonic sensors and the servo I had to create a really creepy prototype of the soon to be insides of our flower installation. It was time to fix the major issue where all the servo engine and ultrasonic sensor code I’d written before stopped working as soon as I tried to make the mp3 shield play a song, this likely had something to do with certain functions using the same communication ports. So I got one of the experts in and started to wrestle with this problem, taking a working code and adding more and more components till it stops working. We didn’t really fix it yet. But we did get alot of sensors working together.
So our first cut had a lot of issues, some parts we could keep, but most had to be redone because for some reason certain parts weren’t the same size exactly, which messed up the mechanism. Then we put it together, this time we had a monster of a mechanism compared to the cuter old one. But it was necessary for the final design. This time we were gonna secure all bolts with Loctite superglue after making sure the design is as firm and stable as possible.
Later we had to change some more and remove some screws again so we could attach Eva’s acryll plastic leaves.
We went back to indesign and made ourselves another design, this time one that does support the six leaves we wanted on our model. We already cut out the first parts of this new version, but some mistakes have been made. So we still have to do a little more. We will keep you updated!
Since we need a rather wide range to sense presence we decided we need to use more ultrasonic sensors pointing in different directions. This was gonna be a bit harder in the code because it required me to turn an already quite long code into a function. There was alot of room for errors and many errors I had to deal with. It wasn’t very fun and I still don’t have everything working. But I do have two sensors both sending their individual data. So that’s something. But the code that I used to get an average number is no longer sure which data to pick so that doesn’t quite work. But a small victory was at least achieved. Two sensors! Now to make that three…