Sumo Bots C

[Jim_R]

Discussion for Sumo Bots C.

Sumo Bots Wiki
Sumo Bots Image Gallery
The_Best_of_2010#Sumo_Bots Wiki
2010 Trial Event Thread

http://www.geekshavefeelings.com/projects/la-cucaracha-sumo-bots-2007
http://www.geekshavefeelings.com/projects/harpy-sumo-bots-2009

[paleonaps]

Didn't Sumo get axed, according to the rule book? The Table of Contents doesn't have it.

[AlphaTauri]

Naps...that's last year's Table of Contents. None of the new events are in there, and the bolded ones are last year's new events.

[paleonaps]

No..... really?
Wow.
/Facepalm

[zyzzyva98]

This is one of the more anticipated events, I doubt they'd axe it, especially if they want to get interest up.

[paleonaps]

I hope the bot can be bigger than 2000 grams. That's really small.

[ichaelm]

Nice, glad to see this up

I just wanted to say that in the past month or so, between summer school and vacation, I've been part-time working on the art of etching PBCs! I divided the sumo bot circuitry into two parts: first the logic board, which interprets the signal from the radio receiver and sends signals to the next part, the power board, which is basically a motor controller. Surprisingly, my power board worked perfectly with the medium-power motors I have. It was basically two H-bridges using P and N MOSFETS, with some simple circuitry with a NOR gate chip so that I could theoretically control four states: forward, backward, shorted, and disconnected, with only two parallel signals from the logic board for each motor. It was also pretty small. The logic board, however, was pretty much a failure, mostly because I tried to do it completely in analog. It was impossible to make it precise and consistent. Next time, I'll try a more accurate digital design.

The actual etching-a-PCB part was a pretty cool experience for me. You start out with a flat piece of plastic, fiberglass, or whatever, that's coated on one or both sides with copper. Then you have some sort of etchant, which dissolves the copper on contact, and another "etch-resist" chemical, which you can apply to your board in the exact pattern you need to make the traces and pads. I don't have any of the fancy equipment that I really wish I had the money for. The only things I actually bought were the etchant, which was some ferric chloride from Radio Shack, and drill bits. The internet says that you can get great results by printing your design with a laser printer on glossy paper, and then ironing it on, but I couldn't get that to work for me. I experimented with a lot of different resists, and the best one I found was nail polish! I was able to apply it with a normal sewing needle in very fine patterns, and then either wait for it to dry or pop it in the toaster if I was impatient. The etchant needs to be hot to work, but ferric chloride makes some nasty fumes when it gets hot, so I set up two fans blowing out the windows of my kitchen while I used the microwave. Then I left it in a glass bowl outside. It took a lot of trial and error, but once I perfected the nail polish it worked consistently. Then you need to drill the holes, assuming you're using through-hole components. Some components, especially in the DIP package, require REALLY thin drill bits. I got some that were gauge 70 or so, or around 1/32 inch, and some others gauge 65 and 60 for bigger things like the power transistors. Then, I had to practice soldering again XD. That wasn't that bad though. In all I was able to make a pretty good quality through-hole PCB in around 6 hours.

This is definitely not something you need to do to build a sumo bot. I obviously went a bit overboard with the do-it-yourself attitude, partially to save money and partially because I thought it would be fun. And it was! But you can easily have some factory make a PCB for you for somewhere around 100$, I think. Or you can buy pre-made things like motor controllers and programmable microcontrollers like the basic stamp, and just program it and hook them up to each other. Or you can use a proprietary full system like vex. And the simplest option would be to buy a RC car and just add a shovel or something XD It all depends on how much time, effort, and money you want to put into it.

[bridgebuilder25]

Does anyone know approximately how much torque a 2 kg bot can take until it becomes overkill? In other words, when will the wheels start to slip?

[Ubermensch]

So has anyone implemented a braking device in their 'bot? You know, maybe a rubber-coated arm that can be lowered in the back of the 'bot to create traction and prevent it from being moved. I know this sounds like a "cheap" strategy, but I was thinking that an extremely light 'bot with a powerful and effective braking system could be used to "stall" out the opponents. It wouldn't be cheating per se, as the lighter 'bot will always win if there is no outcome after a certain amount of time.

[ichaelm]

I've built an electronic braking system that shorts the motor terminals, which is pretty different than what you want to do. My device definitely works, but I can't really tell how well it works yet, because I have yet to put it in a functioning heavy robot. I think your idea is definitely viable, especially if they increase the weight limit this year. But if your robot is already light, you might not need that kind of physical braking, unless you're using bad motors. My guess is that a lighter robot could make good use of electronic braking, but I don't have any test results to back that up. I may try that idea later, and if I do I'll post my results here.

[Ubermensch]

While re-reading the rules to Sumo Bots, I realized that there is actually a rule specifically against stalling tactics such as the one mentioned.

[paleonaps]

Yes, stalling is not allowed. I think that it needs to be said that these matches are usually over very quickly. A complicated strategy is not always the best- sometimes just having a strong bot and driving well is the best approach.

[old]

Anyone have any idea about whether 2.4 Ghz spread spectrum radios will be allowed? In the past our team has had terrible trouble with getting jammed (accidentally or purposefully? I don't know) so we would really like to go with 2.4Ghz since this is essentially impossible to jam.

I was looking over last years rules and though there is a stalling penalty it is written to cover the case where a robot is avoiding contact, not for simply standing still once contact is made. It is hard to say if a judge would invoke that penalty for simply standing still once contact is made.

Does anyone know how often matches were won based on the weight of the Sumobot? The last time I watched this event was several years ago at Nationals. That year the overall winner had a very simple but fast little vehicle. They were able to get the advantage by out maneuvering the opponent, not by any great pushing power.

[ichaelm]

2.4 GHz was allowed last year, so I assume it will be allowed this year. There's no reason for them to ban it.

In my opinion, the best possible design would be the maximum weight, have near instant acceleration and braking, have a very high speed, high traction, and would easily get under any other robot. But usually, people decide that you have to choose between power and maneuverability, or compromise. My goal is to see how close I can get to achieving both of those qualities at a very high level.

[starpug]

2.4 GHz was allowed last year, so I assume it will be allowed this year. There's no reason for them to ban it.

In my opinion, the best possible design would be the maximum weight, have near instant acceleration and braking, have a very high speed, high traction, and would easily get under any other robot. But usually, people decide that you have to choose between power and maneuverability, or compromise. My goal is to see how close I can get to achieving both of those qualities at a very high level.

I concur

"The spirit of the rules" is to attempt to push your opponent out of the ring using your bot. Attempting to win by building a light robot and avoiding getting pushed out of the ring until the time limit is up and the bots are weighed so you can win with the lighter bot breaks the "spirit of the rules." Not that it matters as a good wedgebot can beat a bot that is attempting to exploit the "Lighter bot wins rule" any day.