QOTW 08-24-03: Old School

[Andrew]

I wanted to clarify what I said earlier.

I like the fact that there are no restrictions (other than exotic materials) on the raw materials allowed.

Here's a "for instance."

Back in the "old days," you had to buy any flat stock that was thicker than .25" from Small Parts, pay exhorbitant shipping, and hope they didn't run out before you needed it.

Allowing us to buy any flat stock, any thickness is not a significant concession. But, it saves teams both money and trouble.

Here's another possibility. Let's say that "hex stock" is not on the list. A team with the facilities and manufacturing can take round stock and create hex stock. The teams without such facility have to compromise their designs.

By basically saying, "Go down to your local Metals Supermarket and buy any raw materials you want," FIRST is opening up design options to resource limited teams.

Another example...gears.

Gears and gear stock are relatively inexpensive. If you can buy gear stock and fabricate gears to your specifications (a facility which is within most teams' grasp), then you have a much larger variety of design options.

If you disallow buying gear stock (as was the case in the "old days"), then only teams with the ability to cut custom gears out of round stock could have this flexibility.

By opening up the number of raw material options and putting a $3500 cap, FIRST actually did more to level the playing field than in the days when we were forced to buy only from Small Parts.

This is a good thing!

On the Pneumatics...
Relying on inspectors to find safety issues with non-approved pneumatic components is a formula for disaster.

Relying on all teams to only use pneumatics that are rated to the correct specs is a formula for disaster.

It would be nice if FIRST put together a catalog. However, (this is from someone who has done just that), you are asking some poor FIRST engineer to allocate about 2 weeks of his time, full time, to developing such a catalog. Give these guys a break!

Perhaps if the FIRST community got together and developed such a catalog and provided validation for all the parts, such a thing may come to pass.

Even then, you have inpsection issues. The larger the catalog, the more difficult the inspection.

[Matt Leese]

Because I wasn't paying very good attention, I missed this thread when it first started. A few people have already heard my views on this but I think I'll state them here anyway (I've been meaning a good post on this since about Nationals but it always got pushed to the bottom of the queue of things I had to do).

I think that the loosening of restrictions on what components can be used is the complete wrong direction for FIRST to go. In general, by loosening the restrictions it has made many parts of the competition easier. My problem wouldn't be with the competition being easier if that's all that happened. Because the competition becomes easier, more teams design more complicated robots. Basically, this raises the bar for everyone meaning that if you want to compete you must have a more complicated robot.

Now, that may not sound like a bad thing (after all, what's wrong with a competitive environment?), but the key comes down to what happens when a more complicated robot is designed. In general, the more complicated the robot the less involved the students will be with its design and construction. It's a simple issue of skill level and experience. As the robots get more complicated, it's obvious then that there will be less student involvement and more engineer involvement. This is not the direction that FIRST should be headed.

Now, because of the above, it may seem that I'm against complicated robots. I'm not. I don't think that teams should be forbidden from designing complicated robots. I just think that perhaps FIRST shouldn't encourage it and make it easier.

Matt

[kmcclary]

Originally posted by sevisehda It seems like I have one of the first [dissenting opinions]. Pretty much now a team can put just about anything on there bot except for depleted uranium and a few other exceptions.

I do feel that the main "exotic material limits" are IMHO reasonable, and should remain. In essence, they're limiting outrageously expensive materials, or materials that require "expensive tooling" or specialized expertise to use. That would give "rich teams" access to an unfair strength to weight ratio material that your average team couldn't afford. That is the equivalent of "more weight allowance" for them vs other teams, which we shouldn't allow.

A fair rule, as long as FIRST keeps it current by staying abreast with innovations in "cheap fab" technologies. Once a new technology becomes widely available, cheap, AND technologically reachable by any team, IMO it should probably be allowed. Ex: Super glue is now cheap and available everywhere, but it was once exotic, expensive, and in the past may have given a rich team an unfair advantage. If every hardware store in the world suddenly started carrying carbon fiber tubes, cloth, and a squirt tube of some new mega-glue for it for a buck, we might wish to reconsider allowing carbon fiber technologies into the contest.

Similarly, if some new megabuck "supermaterial" appears, IMHO it should probably be limited until it too becomes potentially "reachable" by all teams.

Originally posted by sevisehda
In the 'industry' most of the time there are limits on materials. Often companies only do business with certain suppliers. Or like someone pointed out only use tested materials.

"Use only QUALIFIED or TESTED materials", definitely yes, but other limits? VERY rare. The market STRONGLY awards competition, innovation and shifts in technology to reduce cost, weight, construction time, etc.. Only in VERY specific industries or in deliberately and artificially NON competitive marketplaces (for political or other reasons) are you normally limited in your suppliers. In fact, limiting competition in suppliers is contrary to normal business practice and motivation, and in many circumstances even illegal.

It IS understandable though, when a contest is SPONSORED by a company, that we have to use their products! But WHY can we drop thousands on fancy gearboxes or pick up a dozen more IFI control bricks, yet can't even buy a 50 cent air fitting from Bimba, a foot more of the SAME tubing, or another valve from Festa when we're short? There are no "additional actuators" involved, and NO safety issue in that case.

Originally posted by sevisehda
On the electronics side often there is a technology cap that stops you from upgrading components after a set time.

HUH??? Sorry, but that doesn't make sense to me. As anyone that has ever bought a home computer (or otherwise experienced Moore's Law) can see, the evolutionary rate is HUGE in electronics. What you can get, or do for a fixed amount of money expands virtually weekly, and doubles in a year or so. We've OFTEN had to stop mid design cycle and redesign, because a part (or all) of the original design was suddenly superseded by the appearance of a new electronic device or component that changed the entire cost vs profit equation. THAT's the real world for you.

BTW... Speaking of which, I fervently hope they upgrade the RC's CPU next year, or let us use some better outboard smarts. Autonomous Mode needs a major computational boost to be much more than crude, and it would be nice to have it integrated into the RC instead of outboard.

Originally posted by sevisehda Pneumatics: Hands down its mainly a safety issue. If teams could make there own or modify components there would be a risk of breakage. My suggestion is FIRST have a catalog of approved pneumatic components in which teams can buy more of items.

No question... I totally agree! But instead of messing with compiling a catalog (FIRST is severely understaffed for that), IMO it would probably be easier to simply specify a long vendor list or "an ISO 900x vendor" and/or "a [specific industry reg agency] qualification or pressure spec".

Originally posted by sevisehda Electronics: Increase the budget. Maybe something like this, 1000$ Max for electronics, 3500$ Max for materials, 4000$ Max Total. No I didn't do the math wrong, this way there is a 'shared' pot between materials and electronics. While still maintaining a lower budget for teams with limited resources.

I still don't see ANY reason to differentiate between the electronics and mechanical budgets. Expand or contract it as you wish, but make it UNIFORM. In industry, we're REWARDED whenever we can remove a piece of expensive machined hardware from a system and substitute its function in software, a cheap chip, or alternative technology. That's probably one of the BIGGEST economic motivators for both optics and electronics growth today. We're replacing mechanical switches, and complex mechanisms right and left with simpler and more reliable optical switches and micros.

Why are we SO stuck in a "make it work via complex mechanics" mindset??? I'd bet if allowed, a poor but computer and electronics savvy team could probably do WONDERS with just the kit, mostly wood and screws for structure (and enough $$ for misc hardware or structural metal where needed), a laptop on the bot, and using the rest of the budget for purely sensors and computational electronics. THAT would be a robot *I* would like to see!

You've got some great points, but I still think FIRST's simplest and best move would be to make a SINGLE "uniform" overall budget, to be split between mechanical, electronics, safety qualified pneumatics (no new actuators), sensors, (or whatever) as you wish, and allow us to buy electronics from anywhere. Say $3500 total, as before.

BTW, That would also allow for including Hobby Robot vendor electronic parts many of us already use! Most of those things are unavailable from either Digikey or Future/Active.

Aside: In a dream environment, I do wish they'd include in the kit: some ROTARY pneumatic cylinders as an option in our kit's "cylinder shopping list", and maybe toss in couple of small 12VDC solenoids you can drive with a Spike. (Should this be a different question or thread?)

[edit - added my reply to another poster]

Originally posted by Matt Leese
I think that the loosening of restrictions on what components can be used is the complete wrong direction for FIRST to go. In general, by loosening the restrictions it has made many parts of the competition easier. My problem wouldn't be with the competition being easier if that's all that happened. Because the competition becomes easier, more teams design more complicated robots. Basically, this raises the bar for everyone meaning that if you want to compete you must have a more complicated robot.

You can to some extent limit complexity simply by adjusting the "uniform limits", such as the additional hardware budget, weight, and size, versus what you're given in the kit.

Unfortunately, the more you limit things, the more you also limit the variety in the robots seen. Less weight allowed implies simpler payloads, since less stuff can be on the chassis. Less budget means robots are more likely to rely only upon kit drivetrain parts (or <shudder> the PLASTIC gears), making them look more alike. Etc.

Note I'm NOT suggesting reducing anything! I think the weight, size, and dollar limits are fine for now, and I for one LIKE seeing all the cool ideas! I'd just like to see the limits become UNIFORM, by dropping budgetary distinctions between the technologies.

[end edit add]

- Keith

[Matt Leese]

Originally posted by kmcclaryYou can to some extent limit complexity simply by adjusting the "uniform limits", such as the additional hardware budget, weight, and size, versus what you're given in the kit.

Unfortunately, the more you limit things, the more you also limit the variety in the robots seen. Less weight allowed implies simpler payloads, since less stuff can be on the chassis. Less budget means robots are more likely to rely only upon kit drivetrain parts (or <shudder> the PLASTIC gears), making them look more alike. Etc.

I'm going to have to respectfully disagree with that statement. I think we see many more similiar robots (basically, pusher bots) since we've had relaxed rules than when we had stricter rules. Compare the robots from 1998 to this year. Because you couldn't just build this overly complex drive train to just push everyone around, you had to work at designing a good mechanism to score.

I realize that the game design is partially at fault for this but the fact that a good drive train meaning you win has quite a bit to do with it also.

Matt

[kmcclary]

Originally posted by Matt Leese
I'm going to have to respectfully disagree with that statement. I think we see many more similiar robots (basically, pusher bots) since we've had relaxed rules than when we had stricter rules. Compare the robots from 1998 to this year. Because you couldn't just build this overly complex drive train to just push everyone around, you had to work at designing a good mechanism to score.

I realize that the game design is partially at fault for this but the fact that a good drive train meaning you win has quite a bit to do with it also.

Matt

Yea, well that game WAS too "swampable" by a simple "big tank" kind of bot.

Hmmm... How do you feel FIRST should "fix" that? Are you suggesting somehow restricting drive train complexity in the future? Yes, some of the major gearboxes are mind blowing and out of the reach for some teams. But OTOH, for me seeing innovative stuff like Thunderchicken's CVT design is half the fun! I'd hate to lose that via something like defining a specific drivetrain in the rules!

Rather than limiting complexity and innovation, how about focusing more on game challenges where the type of drivetrain used in and of itself doesn't matter (or at least not swamp the game)?

I liked the game this year. I hope we see more "ramp top fights", or at least "hill cresting" behaviors of some kind in the future. Talk about leveling the field for drivetrains! I noticed that a simple two wheel two caster robot could EASILY flip and take out a "tracked monster" at a hill crest discontinuity, simply by being AT the far edge BEFORE the tank came "over the hump". As the tank crested the hill, its balance was extremely unstable, and the underside VERY vulnerable to a "little push"...

I hope this year for a "rock paper scissors" game of some kind, where one team can't swamp the game because you need your partner(s) to cover at least SOME of the possibilities.

Hmmm... How about including pie slice segment shaped CONICAL hills some year (or a "plus sign" KOH, with filler slices on the octagon sides)? No side walls needed. You could run around it, or climb it at all sorts of funny angles. Balance, CG, and getting hung up on the sides' "mini peak edges" would ALL be a concern for Autonomous Mode! The drivetrains to automatically handle THAT one would be VERY interesting to see!

- Keith

[sevisehda]

HUH??? Sorry, but that doesn't make sense to me. As anyone that has ever bought a home computer (or otherwise experienced Moore's Law) can see, the evolutionary rate is HUGE in electronics. What you can get, or do for a fixed amount of money expands virtually weekly, and doubles in a year or so. We've OFTEN had to stop mid design cycle and redesign, because a part (or all) of the original design was suddenly superseded by the appearance of a new electronic device or component that changed the entire cost vs profit equation. THAT's the real world for you.

I think you somewhate reinforced my point. Unlike FIRST which lasts 6weeks most real world projects last years. Imagagine if after 2 years of developement a sub-team wanted to upgrade there system. This upgrade may require changes throughout the project. This one change could undo most of the previous work. So in some projects a barrier is set to stop 'upgrades' to prevent having to redesign things over and over.

---

Reading this post has given me an idea. How about a handicap/bonus for teams based on there robots cost. The max budget still being 3500$. The first 1000$ is free. The final score for each round is multiplied by 1.5. For every dollar used the multiplier is reduced by 0.0002. So if you used your entire budget you'd get a multiplier of 1. but if used none of it you'd get one of 1.5. It would make scoring more difficult but it would encourage teams to spend less(often if not always encourage in industry). Also this would definately help out teams with lower budgets. Many other design competitions often have some budget aspect to them so its not a new idea.

My 2 cents.

[Madison]

Originally posted by sevisehda
Reading this post has given me an idea. How about a handicap/bonus for teams based on there robots cost. The max budget still being 3500$. The first 1000$ is free. The final score for each round is multiplied by 1.5. For every dollar used the multiplier is reduced by 0.0002. So if you used your entire budget you'd get a multiplier of 1. but if used none of it you'd get one of 1.5. It would make scoring more difficult but it would encourage teams to spend less(often if not always encourage in industry). Also this would definately help out teams with lower budgets. Many other design competitions often have some budget aspect to them so its not a new idea.

My 2 cents.

Are you volunteering to be the one to count every nut, bolt, screw wire, and connector, measure all the extruded aluminum, steel, and plastic, and then verify the listed prices of each part?

FIRST needs to operate at higher efficiency with as many or less people than we've seen in past years. For that reason alone, I expect much more freedom in robot design and part usage. The less we have to worry about kit legal parts, the less they have to worry about kit legal parts -- and they already have a lot to worry about.