Categories

On we go, Expedition rocks

I think things are unravelling now that we’ve got the ball ready. Indeed, we’ve settled on an architecture, and decided what we were going to do in terms of hardware, and managed to get well ahead with that. The hardware we’re aiming for will, we think and hope, enable many cool applications, especially games, that will make RoseRolls the true Pokémon evolution of the Sphero. But I for one don’t care much about how our work compares to Sphero or whatever. And now that we’ve decided to use the hardware through the serial port using the Sphero API, it’s clear we’re building upwards. So, let’s do this.

Now that the Sphero study and test phase is over (or just about), and that we know where we’re going, we’re getting into the more concrete stuff. It may be more of a technical trial ahead of us but as I said earlier I think the hardest was to decide exactly what our RoseRolls was going to be like – and that’s behind us.

So, what does the more concrete stuff look like ? Well, first of all, it’s tackling the design lab, placing the components on the pretend board and route all signals from one to the other.

This deserves its own paragraph. It was truly an adventure of its own. So first we had the schematics – that was only hard because all of us were all over the place trying to come up with a viable way to fit all the connections around the STM32F405. That done, the schematics went along fine with the necessary time and work. But the really interesting bit – the adventurous one – was placing the parts on the board using Expedition PCB and routing it all.

24 hours ago (after a boring final debate at the Hôtel de Lassay where Polytechnique won because a great black guy had a brighter shirt than any of the bow-tie wearing HEC dues and lady) I was at my desk, giving my first real attempt at routing this stuff. I had given a go at the tools during the corresponding lesson, and had performed a few tries on my own – but nothing serious, especially since I didn’t realize what the task really was – not just drawing lines of one color or another. So I started placing my components, was told to make it all much tighter, which I did (by a factor of about 6, which only tells you my board was originally huge and more an electronic desert than the crowded city it’s supposed to look like), and then I placed my first trace. Oh God. I didn’t know what world I had entered.

At first I tried to preview all  that I was going to do and when it was all clear I did it – and I was using the fixed trace mode. With that I routed about half of everything. It was a real pain in the ass. By 3 a.m. I was to my knees figuring out crazy ways around and over every single obstacle given the already present traces. And upon zooming out, I thought : “This is a dead end, I don’t know how to carry on.”. The answer to my problem was simple : GLOSS ! Better than Zeus’ thunder or Vader’s lightsaber, the gloss feature is the ultimate weapon to shoot your way through a crowded PCB and make it to whatever other pin you’re supposed to connect to. I mean, you readers probably know it well – but I’m a newbie (or I was one then), and one that wonders about everything too ! Everything on your path moves aside, everything is optimized in terms of distance, and all with horizontal, vertical or 45° slanted traces. Amazing. After a time of practice I had gotten the hang of the basic behaviors of gloss (whose main behavior is : “Let me arrange everything for you !”) and decided to give a fresh new go at my routing. Like mr. Polti later told me “You have the means to answer your own questions and solve your own problems”. When he said that, I wasn’t sure it was going to work out – but it did, and beautifully so if you ask me. My teachers may think otherwise – and in that likely case, I await advice with great impatience so as to improve as much as I can. That’s the whole point of this course anyway, and I think it’s working out – one just needs to keep a tight balance between “learn on your own” and guidance from those who have experience – and I take this occasion to thank Guillaume for showing me the way to go when optimizing a ground plane. Guillaume, you’re dedicated, and I learn a lot from you – so yeah, thank you, because now I know and can do new stuff, stuff I like – and that’s priceless. And yes, I think you’re right when you display a loud awe (like only you does) in front of Expedition’s power, cunning and deep intelligence. Call me a geek, but I think this is one of the life-defining things – as in, my dream is now to work with Expedition again in my professional life.

I thought of posting a screenshot of my very first board placing and routing but I’m thinking it would be nothing you to you, readers, so I won’t. Instead, I’ll tell you we’re almost done choosing RoseRolls’ electronic parts, we didn’t even forget a potentiometer to calibrate or IRDA emission power, a JTAG port (though as soon as we can we’ll implement a bootloader in order to update the firmware via Bluetooth without opening the shell), and lots of other nifty things. Designing a board is indeed an adventure ! We hope to get this ready  by the end of the week – and for now, I’m off to what I believe is a well deserved sleep – be it just to stop this article of mediocre prose. TTYL, all.

Charles THIN

ROSE 2013

Commentaires fermés.