Largest spinning disc on a lightweight!

[jim]

I knew Y-Pout was a bad example after I posted it!

I think it could be done, using brakes as Daniel suggests. I was thinking of quickly firing a brake onto the floor rather than the wheel though.

Its obviously possible to make a low speed translating spinner, its just a question of how fast you can go.

*Still believes a true spin-drive system is impossible*
Ah, well, if it was possible, thered be no fun in building it

Mike, Mark, I had never heard of either of these bots.... I will have to do some research.

[bugs]

Im not sure, but i thought teslas tornado breaked on the motors.
If you read the facts & Specifications it mentions the active breaking of the speed controlers

the only way to really have a heavier disk than halfwit (who was 12.2kg at Preston, and as such the disk was 12.2) is to have a walking spinner. If anyone can come up with a walking meltybrain (transitional spinner) where all 24kg spins, and getts over 1000 rpm, Ill be very impressed

Now *thats* a challenge I like. :-) Almost enough to overcome my dislike of spinners. Possibly I should actually get as far as building something normal, first, though!

I like the idea of the infra-red LEDs on the transmitter for direction finding. Having done a bit of a web search (only just back at work, and still I spend all my time waiting for things to build...) Tesla Tornado does seem to use braking, although whether its physical brakes or whether it just means that the engines brake it I dont know.

Ive finally managed to source a bit of stuff about CycloneBot (you get really weird things if you put Cyclone Bot into Google - I presume there *isnt* an official web page?) at http://www.geocities.com/Hollywood/Set/9859/sc4_1.htmlhttp://www.geocities.com/Hollywood/Set/9859/sc4_1.html. If its accurate (well, GPS is nonsense, but extrapolating), its direction finding system is genius: use a magnetometer to measure the Earths magnetic field direction, then steer it with the compass points. Cool.

Ive been having a bit more of a think about translational spinners, just to divert the thread off-topic (as is my want). So far the mechanisms I can think of for actually inducing translation (disadvantages and explanations in brackets) are:

1) Speed up/slow down the motors (efficient in weight, but hard on the motors - speed limited by the motor response?)
2) Tilt the wheels with a cam/navbot as in Y-Pout and Why Not (navbots waste weight, are fallable, problems sitting still, tiltable wheels weaker than fixed ones?)
3) Tilt the wheels manually, with a motor on each (complex, waste of weight, but no nav-bot and more speed control)
4) Manage the speed of the wheels without changing the motor speed (e.g. by having an eccentric gear somewhere in the drive train which gears the wheels up and down in sync with the rotation of the robot, or have a cam-controlled driven differential; mind you, this might just result in the load on the motor changing rather than achieving translation, but a flywheel effect might offset that - also a very complex [and heavy] drive train)
5) Change the size of the wheels, e.g. by variable pressure pneumatic tyres (might have the same problem of just changing the strain on the motor, although probably less so; does require some waste of weight getting the tyres inflation mechanism working - but the words CO2 and doesnt count towards weight limit spring to mind)
6) Brake the thing on the ground (inefficient, possibly unreliable, probably bad with ramps)
7) Cheat - give it Razer/Groundhog style wheels and shove it around with a driven navbot

...and the means of controlling the direction of locomotion:

a) Navbots (waste of weight, possibly fragile, invertable ones hard, but they can shove a bit and drive a cam directly)
b) Gyro-navbot (not in contact with the ground, motor-driven to counter-rotate and keep the gyro still; still a waste of [less] weight and cant do any shoving, although the gyro could be mounted on a cam)
c) Emissions from the transmitter (prone to reflection/interference problems, but no waste of weight)
d) Optical profile matching (measure the light levels on the way round and deduce when a circuit has been traversed; complex and large changes to circumstance [being shoved around by Storm] may confuse it)
e) Track the wheel rpm (error prone with [common] skidding, but very simple)
f) Magnetometer (I *like* it) or, at an unreliable push for outdoor events, GPS tracking

Given that I have a pathological dislike of spinners, anyone care to add to the above (with any designs theyre not keeping secret)? Know thy enemy, and all that. :-)

Incidentally, there seems to be a bit of confusion on-line as to what exactly a tornado drive is, and whether it differs from melty brain. Can anyone elaborate?

--
Fluppet

[jim]

Whatever you use to move the robot will have to be fast - at a rather leisurely 1000 RPM, the robot will make a complete revolution every 60 ms.

Andrew, do you have any other source for the magnetometer idea? I have doubts that youd be able to detect the earths magnetic fields inside a metal box full of motors but it may be possible.

For simplicitys sake I would favour a dead reckoning scheme based on wheel RPM, but with additional feedback to the driver - a set of LEDs which pulse once per revolution to indicate the current heading; the driver can then provide fine control of the heading if need be.

The only source was the comment in the above link:

To tell the robot which way to go, the driver uses some kind of global positioning system that informs the robot which direction is magnetic north.

The idea of GPS is nonsense (at least indoors), so I took the above to mean a carefully positioned magnetometer - but it occurred to me, too, that the interior of a robot isnt exactly free of distracting magnetic fields. Neat idea if that *is* how they did it.

I missed off the simple idea of just sitting a gyro in the robot and tracking rotation rate with it. I suspect off-the-shelf gyros are limited in the rotation rate they support, though - Ive only seen this reported (in one of the web sites Google gives you with a search for tornado drive) for a thwackbot with a low RPM. This assumption was the basis of my idea for counterrotating the gyro, but if there are standard gyros out there which can accurately represent the difference between 1000rpm and 1001rpm then it wouldnt be necessary.

I like the LED feedback idea, although it could get a bit hypnotising!

As for speed of locomotion handling, certainly a problem. The wobbly-wheel method used by Team Whyachi probably puts less strain on the drivetrain than the alternatives. Looking on the spec page suggested above, the Team Whyachi spinners are capable of appreciably higher rpms than CycloneBot (1000rpm for Why Not and 900 for Y-Pout, vs 400rpm for CycloneBot), but Teslas Tornado (with wheel braking?) has reached 870 rpm according to its web page. I wonder where the limits actually are on the different drive options. I dont know whether the various people suggesting translational spinners - if thats what they meant by them - in this section of the forum (I *so* dont dare build a feather...) have any estimates of RPM yet. Mind you, of course, the faster the spin the less the tweaking required per revolution to get the same final rate of locomotion.

Theyre certainly an interesting technical challenge. Id still like to know whether anyone has a solution for the batteries self-destructing, though!

--
Fluppet

My kids just went 5-0 and were crowned US SouthEast Division Champs (in the 30lb US featherweight class) with their new Magmotor powered spinner.

Totally Offensive swings a 22 6.9lb tool steel blade at about 3000rpm. The impacts are huge. One of them sent the bot about 20 feet through the air.

Im working on a frame redesign that will get it down to UK featherweight standards and may bring it over for the FRA world championships, if she-who-must-be-obeyed will consent.

http://www.madoverlord.com/robots/offensive.thttp://www.madoverlord.com/robots/offensive.t for a build report.

On a related issue, we in the colonies are not used to fighting on wooden arena floors (everything here is steel plate). Whats the surface like, what clearances do you really need, etc?

[woody]

Surface is/was painted chipboard......clearance needed .... Ummmmm This varies as the arena floor distorts under the stress of the battles.

You may well have seen some bots.. with little or no clearance .. come to a full stop as they hit a raised sheet.

*double post*