Quote:
Originally Posted by linuxboy
I understand most of this is probably not possible for one reason or another (mainly safety, and I think too many robots would get bypassed, and there would be a lack of support staff, since every robot could be different), but, it does seem like something that would be a lot of fun for the Control System guys on each team to implement, and probably gives more real world experience with Control System Design.
- Oliver
|
While I understand the focus FIRST puts on safety I don't feel it's the greatest impediment to what you describe at all. In fact, most of the likely injuries you might sustain are similiar to those we already can experience if we don't put safety first. Since you aren't suggesting adding additional mechanical risks directly to the systems, I foresee that perhaps the additional safety issues might be exposure to chemicals common to the photographic methods of making printed circuit boards (photographic chemicals, etching solutions, tinning solutions or baths, acids...all of which can be found in your high school chemistry lab usually anyway) and cleaning the printed circuit boards after assembly (flux remover...luckily we don't use trichlor for that). Mechanical safety speaking the only thing that the existing electronics can achieve with relation to the robot's physical components is to stop or start attempting to move on field command and while that's a sort of safety feature on the field, let's be realistic that under non-field conditions the robot can move unexpectedly anyway and there are probably far fewer people there watching (less eyes, more risk). I will confirm that when I was in high school we manufactured printed circuits boards with Sharpie pens drawing on the copper clad boards, ferric chloride heated by a can of Sterno as an etching solution and a drill press (needless to say even my hobby process of making printed circuit boards has dramatically improved on that).
The safety issue fully considered for relevance. I think anything that requires electronics design, production and assembly would likely have these issues:
1. This would require a full circullum of education that might require at least 1 year to provide, not 6 weeks. You must assume a common demoninator of experience with this sort of thing. I'm all for the commitment frankly, but we need to be realistic about the time frame.
2. This might require the handling of additional chemicals and therefore the disposal of said chemicals which could be a regulatory issue (I know for example that my college no longer makes printed circuit boards because of the disposal costs). Course you could just send them out, but that takes at least 1 week generally or the costs go way up.
3. You touched on this already so I'll expand on it. This will make the rules more complicated and the review of the robot's compliance more difficult. Essentially the reviewers will be faced with the need to analyze your circuits. Not sure how much a problem that might be. Perhaps that could be dealt with by early submission of the schematics to some extent.
4. Even assembly in today's environment can be a bit of an issue. We are no longer in the days were everything is through hole printed circuit boards. Today surface mount is most common (except in military and aerospace use where it depends on some factors). Therefore you'll see teams with the need of tools to assemble surface mount printed circuit boards. Sure there are do-it-yourself solutions but all resources like that take up space, take up finanical capacity and create a division of capability. So one needs to consider that factor even if we all share openly.
5. I'm not too terribly concerned that we can all hack up powerful computing solutions. Frankly, the community beyond FIRST is extremely adept at coopting technology. In point of fact, the most common coopts tend to revolve about the programming aspect not the actual manufacturing that is common to electronics. For example, that is why we often see communities build up around Arduino boards, which really, are just carrier boards for Atmel microcontrollers. People can just go buy the Arduino carrier boards, they therefore avoid much of those electronics. However, once they try to interface that assembly to the real world that avoidance can be a major problem. I am concerned, in essence, that many teams lack the knowledge and experience to properly design functional and half-way reliable power controlling systems for motors. It's not a trivial effort. There are many factors one must fully consider. Take for example the issues teams face while implementing the Jaguars, the reaction to the challenge of doing that, and magnify that a few fold.
I will say that the power electronics issues might be mitigated by the availability of various speed controls as we have now. This introduces a set of known limits on the amount of power that can be delivered to an electromechanical actuator. I am all for that if that is the solution. This is to say not just one choice of speed controls and switching devices but a wide selection. Then if teams are up to this challenge of making their own power electronics, and willing to suffer the burden of manufacturing (like WildStang and the swerve drives), the whole community can benefit without creating a difficult situation for new teams to compete within (new teams are already really in a mechanical arms race with the existing teams as the existing teams are more likely to have full machine shops, a programming arms race as existing teams have already dealt with this challenge and a CAD/CAM arms race as the existing teams have probably at least touched on the effort). The key point being to present the idea for review and approval to the long list of available approved solutions we can pull from in advance so they can be bought, built (if the design is 'open-sorurce') or put in the kit of parts.