Register To Comment
Page 1 of 3 123 LastLast
Results 1 to 10 of 24

Thread: What to build?

  1. #1
    Okay, partly because I genuinely want advice, and partly because I think this might be an interesting discussion (and full of hints and tips), please humour me for this essay - and tell me your thoughts...

    Im a beginning roboteer. Ive been hanging around the forum for a while, and following the show since the beginning - with the obvious planning of my own robots that seeing Robot Wars tends to induce in the mechanically minded. I know my Bosch 750s from my Eteks, my regulators from my actuators, my PICs from my MOSFETs, my planetary gearboxes from my harmonic drives (anyone using one of those?) - at least in theory. Ive got several elaborate designs which Id like to build, each unusual and potentially effective, but also, without exception, *complicated*.

    Given that I know my place, and that Id rather learn before I build a super-robot which takes vast amounts of time and money (because getting it trashed because of a stupid mistake will be very annoying), common sense says I should build something simple and traditional, but more likely to lose, first. Because I know Ill lose, the robots out there are no longer season 1 quality and neither are the drivers, and its better for my ego to be thrashed in a robot I know isnt very good anyway.

    But what to build?

    Conventional wisdom says start with an antweight or a lightweight, but I cant see much being reusable for the heavyweight I really want to produce. Besides, I probably couldnt see it well enough in the arena, and Id rather have a bit of room to work with. Also its easier to drag victi^H^H^H^H^Hfriends in to help with something bigger. So humour me, if only because Im in this bit of the forum, and assume I want to cut my teeth on a heavyweight.

    Conventional wisdom next says, build a box on wheels. The problem with that, of course, is that there are at least two boxes on wheels on the live circuit anyway (obviously I mean SMIDSY and King B) - with Tanto and others upcoming - and the chances of me doing anything to any of them with a poorly built first robot are tiny. I may expect to lose against everything, but Id want a finite chance of getting lucky. More importantly, a rambot kind of relies on being put together better than its opponent when it smacks into it - and somehow I doubt thatll be the case.

    Thwack bots (T-Wrex, Overkill, Stinger) may be slightly better in terms of not taking the whacks they give out, but are harder to drive, and Id really struggle in a small arena to do anything useful with it. Besides, sit and spin is a boring tactic, and overhead thwacks arent so effective. Plus: live event = railings. Tend to be trouble.

    So although Im willing to be persuaded, Im probably looking at active weaponry.

    That leaves:

    Pneumatics (flippers or axes - TaN/Huggy Bear/Complete Control style clampbots probably require more driving skill than I have for now). Every book on the subject says stay well away from pneumatics until you know what youre doing, or youll kill yourself on them. Ah. Problem there... Otherwise I like the flipper idea. Axes are interesting, but modern armour (other than what Ill be able to put together) kind of means I dont expect to get far with one.

    Hydraulics (pretty much crushers, unless anyone has other plans). Like pneumatics, but more dangerous - and I know least about it. Kind of restricts you to entering an extreme damage class, with a spindly appendage sticking out of your robot while people with spinners are wandering around. Not a problem (much) if you have driving and engineering skill, but thats what Im here to get.

    So, unless someone talks me into pneumatics, that leaves electric weaponry. Which means:

    1) Spinners. Same problem as hydraulics, extreme damage only, and you need to build them really well so theyll take a hit. Oh dear. Besides, Im from the Rex Garrod school of not demolishing the opposition, so Ive never really like the idea. Since this is for live events, Id probably not be allowed to run anyway.

    2) Lifters (like Panic Attack and Biohazard - and arguably Storm 2, although I dont buy it). Okay if you have a lot of driving skill and a solid robot. Oh dear. Also kind of reliant on the presence of a pit, which is by no means universal in a live event.

    3) Electric axes. Hard to get through current armour, and anything powerful enough to bypass this problem (e.g. Beta) is going to be expensive and hard to build.

    4) Saws. With the possible exception of the Pussy Cat blade (which is arguably a spinner), pretty much incapable of cutting into any current armour. Rare these days for a reason.

    5) Esoteric stuff (drills, disembowellers, stampers, stompers, flingers, wheelcutters... dont ask me to explain all that) - not very tried and tested, probably wouldnt work very well, might not be arena safe, and might do expensive damage to opponents if they *did* work. Not for a first robot, methinks.

    So, I think Ive just said why I wouldnt start with *anything*. Hence my dilemma.

    People have always said start with a box, because its the obvious thing with which to start. So, to the general assembly: is that still good advice? With all the combined knowledge of the roboteering community, is the easiest thing to build really the simplest? What stands up best to being built badly? Im not expecting to win, just to put on a good show - and that point of view is another reason why a (dull) box is a bad idea. What has the greatest chance of producing an interesting fight, per unit competent builder?

    Dragging tiddlywinks into this: experts at tiddlywinks win by playing very simple shots very well. Leave yourself in a position where you never have to *play* a difficult shot, and youll never *miss* a difficult shot. A novice advised by an expert is a dangerous opponent. What is the robotic equivalent - the minimum input/chance of failure for the maximum return?

    What, with everything you know, would you decide to build if you had a competency failure?

    Anyone?

    --
    Fluppet

  2. #2
    Mr. Fluppet.

    I too am starting out on the path to enlightenment/non-destructive-carnage, and Ive got a handful of ideas Im considering. Like you Ive considered all the weapons and sizes and Ive come to a few conclusions.

    1. Weight. Im building a feather for the pure and simple reason that I cant afford to build a heavy. I couldnt even afford one speed controller never mind two, and a mixer, and the 2MAh batteries, not to mention £100 PER MOTOR etc etc... Im not building a lightweight because theres so few of them and you never see events with a lightweight class. Middles are basically smaller heavies with most of the same problems, and Im just uninspired by ants.

    2. Weapon. Pneumatics would be nice and Im not ruling them out for the future. Hydraulics - ditto but harder to implement, more dangerous etc etc. Axes/hammers - ineffective without pneumatics or money (as you said). Spinners - Im not after wholesale destruction (of me or others) so theyre out. That leaves me two options. 1. An electric lifter/flipper. This is relatively cheap to make, should be easily able to lift a feather (another problem with a heavyweight is the extra power needed for this kind of thing) and can be used as a srimech if well done. THis is probably what Ill go for. 2. An electric crusher/gripper. Sounds unlikely but its a possibility Im considering. It wouldnt do as much damage as hydraulics etc and therefore should be allowed in the low-damage competitions. Im not going for a rambot as I already have my motors and Ill be woefully underpowered and too slow to push around the featherweight equivalent of Storm II (they should do an ant called Storm in a Teacup ) etc.

    3. Shape. Build around the weapon is the advice I heard. I have a few ideas buit its going to basically sit on a rectangle of box-section steel and the frame can be whatever shape works, with a few rules (to let the weapon do its stuff etc). For control Im having the wheel span nice and wide, and probably a castor at the front, but with (hopefully) and adjustable zero ground clearance at the front and maybe skirts around the sides and back. I like the idea of a plough shape, but itll basically be a funny shaped box. It wont be invertible (I plan to use Dominics motor mounts and wheels eventually and Id have to make it very thin to be invertibe on the same wheels) but it will have a srimech of sorts.

    My most likely point of failure is going to be the electronics which Im designing and building for myself. Its a full speed only speed controller, but with two stick control (one up down, one left right) and spin on the spot capability. At the end of the day tho, itll be driven by microswitches and relays and Im designing it in such a way that if the electronics fail I can take em out and JUST drive with Microswitches and relays, and only lose the on-the-spot part.

    Basically Im keeping it as simple as I can on the assumption that at some point it will fall apart and break something, and I dont want that to mean a rebuild, or buying something else pricy to replace it. It should also be able to take a short flight and land on concrete and be ok, but Ill test that at some point no doubt. (another testing issue with a heavy - how do you throw it to test it? Unless you live near a big flipper)

    Ill stop there as my rambling is refusing to come to a sensible end point.

    -- Kev

    EDIT : that all sounds like Im saying dont build a heavy which Im not. Im just saying Im building a feather for these reasons, and heres my thoughts on a first bot. You never know, my feather may go overweight and become a lightweight yet.

  3. #3
    Dont feel too impressed by other robots and do try to design/build one that can take a few hits and deal out a few as well. If you feel constrained by a budget too small go at least for a feather or antweight and get the experience, you will learn a lot out there and have the chance to exchange ideas and tips with other roboteers. You might even be able to acquire parts for a heavy in time when you start to know certain roboteers, or better parts for your existing robot (be it heavy of feather, or even ant, we dont care). Some people know the cheap addresses, or might arrange something for you that you can afford. Talking is very important here, as are having a broad mind about everything and showing an open attitude as well.

    Shape: there are so many geometric shapes, choose and combine... strange and weird designs may pleasantly surprise onlookers and may even confuse opponents. With regard to the frame I would suggest support beams when and where you can have them, triangular support shapes tend to be the strongest and are recommended. (If you dont understand what I am referring to, send a private message and I will make a drawing for you.) Not everyone can buy or -even more difficult- install hardox or titanium armor on his/her robot you know ! Most often it is the frame that literally holds the robot together, along with strategically placed bits of tougher armor here and there. Armor does not need to be on the outside of the robot either... protect the real expensive and/or vulnerable parts deeper in your robot and you might walk out of a fight with less damage than you think.

    Weapon: you might be surprised at what a simple weapon will do. Damage counts ! Being agressive counts ! Strangest weapon I have seen so far: a ships propeller, strange but true and it is a German robot too. They have a long history of surprising their enemies. (See for instance the Blitzkrieg where they used floating planes to drop troops instead of parachutes !!!)

    Driving: experience is the key word here. Live events give you the opportunity for gathering experience in both driving and fighting. Do not expect to win at your first battle ! (Though it may be a really big nice feeling if it happens.)

    I may not be right, I even may make mistakes in my explanation, but this is how a non-technical person sees Robot Wars. Have a go at it.

  4. #4

  5. #5
    The dictum with RW these days seems to be: Whatever you do, do it properly.

    I would be inclined to build a flipper- theres no better way to learn about pneumatics than to play with them (in a safety concious way of course!) and they really do become very easy. Yes, it is dangerous, but not if you take a common sense approach, and realise that if you dont know what will happen if you stick this tube in that fitting, dont!

    LP pneumatics are obviously preferable to HP for a first-time robot, and can, in many circumstances, be just as effective, if not quite as spectacular. If done properly (read: a la M2) then they make excellent weapons.

    I can see the logic in wanting to start with a heavy, but feathers really do have some massive advantages over heavies, not least in size and expense (tho by no means complexity). They really are scaled down heavies. Antweights, im my 4 years of experience with them, are not the place to start if you have heavyweight aspirations. The technology and construction styles and techniques, materials and so on are very different to feathers, lights, middles, and heavies, which all have very similar construction techniques.

    Fitness for purpose: this obviosly encompasses combat, but also, as is all too frequently overlooked, maintainability and repairability. It makes an enormous difference when in the pits. Thankfully it doesnt happen so much anymore, but it used to be the case than crystal changing was a 20 minute affair with some robots.

    Materials- Fitness for purpose, not just fitness period. Ill illustrate this with an example from the semi shown on saturday: St Argo was well built and nicely machined, but I beleive (and correct me if Im wrong) its wheels where a hard grade of aluminium, eg 7075 or 2014. Yes, these materials have a high Youngs Modulus, and are reasonably tough, they tend to snap and shatter instead of bend and be gouged when hit by spinners. This isnt very desirable for wheels- as you saw, a massive chunk of the wheel was snapped off by a spinner. If we then look at TANs wheels, they are made of a plastic (I forget which- appologies) which doesnt snap off in large chunks, instead close inspection of them reveals spinner tooth marks- where they have been gouged, but the wheel is still in tact. So while St. Argos wheels may on paper appear stronger than TANs, it doesnt make them more suited for their purpose. This priciple applies equally to armour and chassis.

    In short, Id build a pneumatic feather as a first machine- its advatages include:
    a)You will learn about pneumatics as well as electronics- so 2 disciplines- which will allow you to make a more informed choice when building your heavyweight. (the same could be argued for a hydraulic feather, but hydraulics are expensive, difficult (ie you have to know what youre doing) and very heavy).
    b) Flippers need good, powerful drive, and need to be very well driven, so you will have the benefit of learning to design a high performance drive, and learning to drive it well at live events. (again, allowing you to make a more informed choice when building a heavy). I dont want to start another spinners dont need good driving arguement, but I think it is true than flippers tend to need slightly more accurate positioning, and rely on good driving, slightly more than spinners.
    c) You have all the adavtages mentioned previously of building a feather.
    d) Feather flippers are fun!!!

    Hope thats of some help, and Id be glad to talk to you more about any of that at the social (although there will be far more qualilified people than me there so you may be better off talking to them :-)

  6. #6

  7. #7

  8. #8
    Babeth rule 2: Air can be armor.

    RCC2 is very well armoured then Babeth.

  9. #9
    Thank you, everyone! To keep this going, I should probably provide a few more parameters for leverage, at the risk of making this less applicable to other readers of the thread.

    Firstly, budget. Id like, at least for a first robot, to keep it as far under a thousand as possible, but Im presuming something in the a few hundred range. This presumes I can re-use as much as possible, though - I cant afford to make a habit of it, unless I get sponsored. I suspect thatll cover a basic heavy if Im prudent, although obviously were not looking at titanium and Magmotors. Im not counting petrol/accommodation etc. in that.

    If Im going with a heavyweight (Ill come back to that in a minute), that breaks down as follows.

    Im presuming that its worth my while to buy a couple of modded Bosches. Partly for ease of replacement parts, partly because I might be able to sponge a temporary replacement in the pits if Im really stuck (and equally loan things out, obviously), and partly for ease of matching the motors. Im sure I could get some decent motors for less from a scrap yard, but the first time they get wrecked I can see myself needing to re-do my entire gear train for a different motor. Any opinions - is this an unnecessary investment, or would the second-hand route be a false economy? I dont have an infinite amount of time (any more) for scrap hunting - although since this is why I can now afford to start building I shouldnt knock it.

    Im guessing, unless Im lucky in scrap, that Ill be buying gears/sprockets/chain/bearings (depending on the drive train). Im hoping theyre not ridiculously expensive. I dont have a machine shop (Im also budgeting for a cheap MIG welder, although I have some basic other tools thatll cope at a push until I start with the hardox), and there are the same issues with replacements otherwise.

    The frame and armour, on the other hand, Im thinking purely of the scrap route. Box section and steel/ali plate are much of a muchness at my stage of procedings, and a grands worth of titanium pepperpot after the armour gets trashed wont do me as much good as something I can fix up with a sledgehammer and some judicious welding, or replace for a pittance. Well see about wheels and weaponry. I dont think thats contentious.

    Control might be, however. As I said, most of what Id *really* like to build is... interesting. That includes the locomotion. Even if I start with a simple bot, I dont want to spend a fortune on a Vantec or 4QD that I cant re-use.

    I have passable electronics knowledge (Im really a software kind of guy), but I might be able to cobble together something that would work well enough in a given application, even if it wouldnt impress David Gamble much. I know Im going to need PWM (and some speed-dependent voltage ramping/current limiting) and not just on/off, btw - Im trusting some appropriate microcontrollers can sort out the hard bits. (Kev - you probably wont smoke your motors like a heavy would, but are you *sure* you can get away with on/off control?) Id rather trash some PICs and a few FETs than a 400 quid unit, and it makes life easier for some intelligence in the control system too. Am I biting off more than I can chew?

    The same applies to the radio end of things - I have no problem building a custom controller, and ideally Id want to send a data stream to the robot, which can then sort out the mess. Is my best bet to buy a radio modem, or should I actually build the transceiver myself? If so, would I be trusted to get the frequencies right? Is this way overboard for a simple first bot?

    Okay, done giving my circumstances.

    Regarding feathers, Eddy and Kev: Im still concerned about how much robot is reusable, and there are scary people talking about translational spinners in that forum. If Im treating the motors as an investement, and if I have no problem with the physical size of the robot (I own a car which would fit it, and since Im of a size where Id only count as a heavyweight because Im a walker I could probably move it around okay - I need the exercise!) are the other benefits still more significant? I can see people from work helping on a heavy, but not on a lightweight. Plus custom electrics only seem insignificantly light weight on a heavy.

    On the other hand, Id rather lose my finger (scuse the pun) to a pneumatic accident with a feather than break my jaw on a heavyweight...

    Pneumatics: given advice and due care, is this advisable for a sufficiently paranoid novice? Is it going to cost a fortune? Several of my more interesting designs have pneumatics in them, so the extra experience would be welcome, but is this too much to start with? If not Ill look at electric lifters, but I doubt I can make an effective one - other than, possibly, a dull Biohazard clone, which would be trashed by Storm2.

    Marios modular idea is interesting (and yes, live events only) - although Im hoping that I can do some remodelling with MIG and an angle grinder; I certainly want to expand the design as I try to learn about new techniques. Ive got robots designs which are pretty detailed, but for something I want to be a simple trial run, I dont want to over-think it - thats why Ive not built yet, after all!

    On the other hand, Ive always believed that robots can get trashed, but ideas cant. Therefore a robot should be designed very carefully, but built as cheaply as possible. Im deliberately going against that here to get some more parameters for the design - Ill still build cheap, but Im not particularly building to win. Doesnt mean I wont think about it, but itll stop me straying too far from the beaten path (even though I generally favour innovation over success; besides, I can hardly ask people to give me ideas for a completely novel robot design in this thread).

    Shape: Wedges are clearly good, and skirts might be wise too (if I get the hang of attaching them). Novice question: in TV footage all wedges look like theyre on the ground - how much is that really the case? Ive never been close enough to a running robot to tell, and people keep talking about a given robot being lower at the front. Getting under a better robot does seem my best chance of doing anything much.

    I had a long debate with people on the RW forum last year about triangles. I know the principle, but Im not sure I ever got a straight answer as to whether some people had a more rectangular design deliberately to allow for the frame corners to rotate to absorb impact - where a triangle would have to fracture. Andrew Marchant assures me that there are no true right angles left on Tornado, but its pretty stiff anyway. Any more thoughts on that?

    Eddy: Do flippers need good, powerful drive? By modern standards? If so, a couple of Bosches might not cut the mustard these days. Enough to control the robot well and shove another robot while flipping it, but Id have thought orders of magnitude behind ram bots?

    Babeth: to John Reid, all opposing robots look like nails. Who am I to argue? For the rest of us, the trick is designing a universal hammer. As for being too impressed by other robots - the roboteers have my respect for building anything (particularly for so long as I havent), but that doesnt mean I dont look at every robot with a view to working out how I can beat it!

    Rather than address everything else everyone said, just presume Im grateful so I can stop rambling before everyone loses interest.

    And yes, Eddy, looking forward to talking with you at the social. :-)

    --
    Fluppet
    He does go on...
    (Yargh. Ive just seen this in the preview, rather than a bit at a time in the entry box. Must learn to be more concise.)

  10. #10
    Have a look at Roger Plants website, http://www.bigcheeserobot.comhttp://www.bigcheeserobot.com - he has combined gearboxes and axles and motors to fit them, and suitable speed controllers.

    2 axles: 30 gbp
    2 reliance motors: 50gbp
    2 Curtis 1208C speed controllers: 100gbp
    Total 180 gbp for all the power transmission components you need, plus a bit for postage and packing.

    This will not give you as much pushing power as Storm II but at this stage I wouldnt worry about it. It will almost certainly be more power than your wheels can put down onto the floor. You say you are more interested in innovation than winning, which is a commendable attitude, and you shouldnt worry about being underrated in terms of power. Reliability is more difficult to quantify and is, in general, better with tried and trusted components like these rather than bleeding-edge technology.

    I would advise against building your own speed controllers. I tried this, and it was a very interesting project, but there were a lot of difficulties and I gave up once I realised I could get the Curtis controllers for 50 pounds. You may find that a lot of your existing knowledge of circuits doesnt hold when youre into hundreds of amps. Of course, you may know enough about this field to make a good job of it. Mine lasted about 30 seconds in its first proper fight, which was a severe disappointment considering they let down the time Id put into the rest of the robot.

    Control electronics are much easier (from my point of view) and theres still room for quite a bit of creativity there if you just treat the speed controllers as power output stages. Just make dead certain it fails safe!

    Best of luck.

Register To Comment

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •