...yes, it's that time of the year again, everyone! Time for Ry to show up for a few weeks before inevitably disappearing for about a year after life distracts again! This time, I lend my pseudo-intellect to the world of combat robot design! While I've personally only screwed around with hack job hobby grade bots built out of RC parts and that have never seen legitimate competition(just fun with friends), I do have a decent amount of history in amateur & professional auto racing(even worked for Michael Shank Racing, a professional road racing team that won overall at the 2012 24hrs of Daytona[of course, I only worked for them for the '11 Daytona 24 *KILL ME*] and is currently Acura's factory team, racing the new Acura NSX in IMSA competition) where ounces count, cutting edge composites & mechanical designs are developed & utilized, and where stresses are MAGNITUDES greater than anything seen in robot combat, plus some maintenance positions at a few manufacturing facilities that have necessitated the use of my fabrication skills, not to mention the work I've performed on my own cars & bikes(like rebuilding the engine of my '81 CB650), so one could say that I have a large amount of indirect experience that is applicable to the limitations & necessities of combat robots.
And so, with my nephew having recently gotten his license(...it's ****ing with mah head, man - I ain't THAT old, I swear!), a job, and expressing interest in building a robot, I decided to organize my thoughts & opinions about combat robot design into an 'in general' detailed 'philosophy'(as much for myself as for my nephew), with a heavy influence from my experience in high end automotive racing dictating the minimum standards that I would expect of myself and my robot designs(and some around here know of my many unique concepts that I still believe are realistically plausible - I STILL want to build a real life '
Unlimited Limits'). Needless to say, it's a bit incomplete and a work in progress, but I thought some here would find my technical vomit interesting at least(get ready, she's a long one!)...
[note: this is biased towards the BIG weight classes, not hobby weight robots - weights & kinetic energy don't scale linearly thanks to physics, so many of these statements simply won't be true for hobby scale bots]
Design Goals
- Maximize simplicity vs benefit; if benefits of a system do not outweigh its complexity, reliability, or ease of repair, it should be redesigned & simplified
- Weight reduction; all possible areas while minimizing impact on strength & durability, and meeting target kinetic energy goals. Truss style lightening holes in structural components. All edges radiused. *Mass does NOT equate to strength - intelligent design & engineering!*
- Modularity; robot systems should be modular where possible, intelligently designed. Modularity should NOT take precedent over strength, durability/reliability, or simplicity, and should be applied where reasonable.
- Shock mounts; where possible, chassis joints/modules should be shock mounted to absorb as much energy as possible before it's transmitted to sensitive components. Intelligent design; shock mounting should NOT decrease mounting strength/durability to the point of unreliability.
- Minimal friction; high quality bearings and/or oil impregnated/self lubricating bushings in drive systems & weapon systems(high kinetic energy spinners should consider high quality bushings - significantly stronger & more reliable than the comparatively fragile bearings). Due to the lower RPM, drive systems can use dry bearings for ultimate low friction, but will have to be checked for play & replaced as necessary with regularity(not suggested for the high RPM[read: HEAT] of spinning weapons, should be cleaned & lubricated regularly for maximum efficiency, reliability, and life). Sealed bearings should not be used in combat robots, the seals significantly increase friction & are intended for extraordinarily dirty environments, such as industrial manufacturing applications
- Hardware; high quality black oxide(not stainless - susceptible to gulling, stripping) hardware is suggested in strength critical applications, raw graded hardware OK in all else. Thread locker(KNOW YOUR COLORS), antiseize, or locking hardware used on ALL fasteners, quality lubricant used on torque-to-spec bolts if used(critical for accurate torque - see: race engine building, lube used where a torque wrench is used). *NO HARDWARE STORE GARBAGE* graded bolts mandatory(better to 'waste' unnecessarily high quality hardware than cheap out and find out that you need it later - quality hardware is cheap insurance). Also, hardware should be sized for the application - BIGGER IS *NOT* BETTER. Gun drilling is possible for weight reduction with negligible affect on strength, titanium hardware is also a weight reduction option(and a wallet reduction method...)
- Minimize weapon reset/spin up time while still meeting kinetic energy goals. Applies to ALL weapons; spinning weapon spin up time, pneumatic/hydraulic weapon reset time, etc.
- Redundancy where possible. Multiple drive motors, or, for a spinner, multiple smaller weapon motors(w/ one way bearing/clutch systems) with individual electrical systems instead of a single large motor & electrical system.
- Cooling; heat sinks & airflow accommodations(venting, cooling holes, louvers etc.)where appropriate, possibly cooling fans(high pressure computer fans are a cheap, lightweight, airflow optimized option). Given the ridiculous amount of robots that have LITERALLY went up in smoke in the two recent seasons of BattleBots, this MUST to be a design priority.
- Motor protection; clutch(slipper, centrifugal, etc.) incorporated into weapon drive to protect motors & weapon drive systems upon impact. Also applicable for high powered drive systems(Vladiator and similar). See hobby grade RC cars, go-karts and similar for examples of intelligent clutch systems that both transmit torque while protecting drive systems.
- Wiring; all joints soldered or positively connected via high quality multi-pin connectors. All wiring & connections should be securely mounted, nothing free or loose.
- Electronics; all electronics securely shock mounted, nothing directly attached to chassis/mounting module.
- Power transmission; over sized shafts on over sized bearings/bushings, all power transmission & idler shafts gun drilled chromoly steel or titanium. Power transmission shafts MUST be splined, double-D, hex, or similar - NO KEYS, WOODRUFFS, OR SET SCREWS, these are NOT designed for torque transmission/torsional/shear loads and WILL be a weak point(they are also not a reliable mechanical fuse either).
- Mechanical fuses; these are pieces of hardware within a mechanical system that are deliberately designed to fail above a specific amount of load(such as a shaft that is intentionally weakened by being necked down to a specifically calculated diameter that will fail after torque load surpasses a specific value), with the goal being to accept a minor system failure to prevent a catastrophic failure that will cost significantly more spare parts, time, and money. This should not be attempted without some sort of legitimate engineering oversight(with calculations to back up the intended goal) as this is a pretty simple concept that is extraordinarily complicated & difficult to implement. Without such oversight, it's extremely likely that one will either undershoot, creating an unnecessary failure that will likely cost a win or overshoot, in which case the fuse will never fail before the catastrophic failure that it's trying to prevent. There is NO margin of error here.
Robots CAN *NOT* be accurately designed on paper; do not spec out the largest motors, batteries, strongest wheels, thickest armor, etc. with the belief that your paper calculations are accurate and will put you right at the weight limit - the many points mentioned above(hardware, wiring, heat sinks, cooling fans, shock absorption, etc.) *WILL* leave you desperately cutting weight in every area, resulting in a flawed robot. CAD designs that don't take the aforementioned into account will create the same problem. Unless you have a significant amount of experience in combat robot design, LEAVE WEIGHT TO SPARE, you WILL find ways to make use of every ounce under the limit.
Design Principles
Chassis - components assembled into individual modules
- chassis formed by modules being interconnected with countersunk holes, tapped holes/rivnuts & screws, to ease chassis repairs
- permanent joints should be attached with appropriate structural adhesives instead of welds(if budget allows) as this is stronger than any other method(including welding) and is lighter weight(see: pretty much every fully aluminum automobile on the market - adhesives replaced welding about a decade and a half ago)
- each module intersection either solidly mounted or shock mounted¹
- robot potentially completely reconfigurable into various designs - varying length, width, possibly even number of drive motors and amount of driven and/or total wheels(such designs must be impeccably designed & implemented or will likely be mediocre in all versions - advanced builders only, very risky)
Material - Metal²
- Composite²
Armor - armor shock mounted¹ to chassis where applicable
- armor thickness dependent on location(important to maximize use of available weight limit)
Style - conventional individual 'piece' armor with multiple mounting points, typically mounted directly to chassis
- single/multi-piece 'shell' armor, indirectly mounted to chassis by way of some sort of strut system, possibly using 'leaf spring' type struts that could be preloaded against each other during attachment. Likely less reliable than conventional piece armor due to less secure mounting. Risky, could leave your bot nekid if mounting struts/attachment points fail mid match
- combination shell + piece armor
- special 'heat sink' style armor to slow spinners(?)
- kevlar fabric attached beneath hard armor to "totally not entangle" spinners(?); obviously questionable legality("it's intended to stop piercing weapons!"), pretty damn expensive sacrificial armor, questionable effectiveness without testing, use at your own risk...
Material - Metal²
- Composite²
Drivetrain - independent drivetrain modules(see: Carlo Bertocchini's BattleKits for basic concept)
- structure: heavy wall carbon fiber square tube or extruded aluminum for durability of module
- drive: timing belt or motor mounted gearbox
- highest quality ball bearings(ceramic?) where possible
- drive & idler shafts
- chromoly steel(4130), titanium
- all shafts gun drilled
Wheeled - two 4 wheeled modules or four 2 wheeled modules
- wheels do not need to feature the same level of grip; a 4 wheeled robot can be designed to rotate around its front wheels by using lower grip wheels on the rear, allowing them to slide easier. This can be used as a method of tuning the drive
Shuffle - composite(delrin?) feet, rubber coated/soled
- with appropriate 'sole' material, can be one of the highest grip forms of drive due to the large amount of surface area compared to wheels while still allowing for very high speeds; great for pushers/rammers, as illustrated by Drillzilla/Dreadbot(the same robot, competed in 'Robot Wars' & 'BattleBots' respectively) before rule changes dictated that 'shuffles' didn't qualify as 'walkers'(which has, thus, made walking robots basically unfeasible - as it currently stands, no amount of extra weight will make them effective)
- a form of drive waiting to be exploited that is constantly being looked past since most see them only as a former 'walker' exploit, overlooking the speed + traction advantages that they offer
Tracked - a very advanced, complex form of drive that is very risky due to the difficulty of perfecting its design. A perfectly designed tracked robot can work as well as a well designed wheeled or shuffle bot, but anything but a perfectly designed track drive will inevitably lead to reliability issues that will leave you without propulsion, often times for reasons
- heat treated chromoly steel or titanium treads are critical for a durable, reliable track drive, rubber coated/soled
- also a very high grip option, dependent on the 'sole' material, due to the large amount of surface area compared to wheels
Weapon systems - independent weapon + weapon drive module
- high quality bearings in drive system, oil impregnated brass bushings in high load pivot points
- all shafts gun drilled
Static - Material: mild steel, chromoly steel, titanium, carbon fiber, kevlar
Wedge/Scoop
- integrated chassis/armor or hinged
Rammer/Blade
- integrated into chassis/armor or independently attached(removable)
Electric HK¹ Spinner
- 2-3 smaller motors, double voltage, with clutch/one way bearings
- conventional disc/bar w/ SURS*
- possibly 3 piece sandwich construction for simplicity of SURS*
- capable of being mount vertical or horizontal
- centrifugal clutch or slipper clutch
LK² Spinner
- conventional disc/bar
-
- hammer mill
- spring loaded 'hammers'
HK¹ Drum
- 2-3 smaller motors with clutch/one way bearings
- conventional w/ SURS*
-
- hammer mill
- spring loaded 'hammers'
- sandwich construction w/ large diameter hammer pins supported by bushings
LK² Drum
- conventional
-
- hammer mill
- spring loaded
Pneumatic HP³ Flipper
-
LP⁴ Flipper
-
HP³ Axe
-
LP⁴ Axe
-
Mechanical SE⁵ Flipper
-
SE⁵ Axe
-
Electronics - electronics mounted in independent module
- individual multi-pin connectors for wiring of each module(combining wiring of each module into individual wiring harnesses), using
Weather Pack connectors or something similar
- electronics & batteries shock mounted
- resettable breakers on all module circuits, no fuses
- solenoids to reset breakers during match?
Speed Controllers -
Batteries Drive
-
HK¹ Weapon
-
LK² Weapon
-
Miscellaneous
- heat sinks & ventilation for all motors, batteries, and electronics, possibly fan assisted cooling using computer fans
- countersinking/flush fasteners where possible
- drilled & tapped connections where possible/minimal use of nuts
* SURS - Spin Up Reduction System, sliding weights integrated into spinning weapon, spring loaded towards center mass to maintain greater weapon weight(kinetic energy) while reducing spin up time by tuning the sliding weights by operating RPM(the higher the RPM required to move the sliding weights, the quicker the spin up) |
¹ HK - High Kinetic energy, weapon system 1/2+ of total weight, slower spinup, heavy impact - Nightmare |
² LK - Low Kinetic energy, weapon system less than 1/2 of total weight, quick spin up, smaller impact |
³ HP - High Pressure pneumatic, damage oriented - The Judge, basically anything from Inertia Labs |
⁴ LP - Low Pressure pneumatic, point/control oriented - |
⁵ SE - Stored Energy, spring loaded weapon for example |
Thoughts, explanations, reasoningShafts - Gun drilling - the root or center of a shaft contributes very little to torsional strength while generally being the initial point of torsional failure(the material shears at the root/center, with the initial crack working its way outwards at 45 degrees across the width of the shaft, as shown
HERE); gun drilling significantly improves strength-to-weight(reduces weight significantly while having negligible affect on strength), 10% reduction in strength 25% reduction in weight rule of thumb, improves strength to yield(allows greater degree of twist before shaft yields - yield is the point at which a material will no longer spring back, its strength falling off a cliff), reduces rotational mass. Not meant for pins in shear or bending, reserved for torsional applications.
- Splined, double-D, hex, or similar - keys & set screws are used to locate components on a shaft, they are not designed to transmit torque(support shear/side loads), which is why they're known to fail when used in such an application. Splined shafts are generally stronger, with the greater the number of splines always being stronger than larger, low count splines.
Proof of concept - see axle shafts in racing applications, such as high level drag racing, rock crawlers/rock racers/desert racers(Trophy Trucks)/Ultra4(combination rock crawler-desert racers), Formula 1/IndyCar, etc. You'll see gun drilled shafts with high counts of small splines(
IMG of a BIG gun drilled 40 spline axle shaft from a Trophy Truck - note that the outside diameter is SMALLER than the diameter of the splines,
this is actually stronger than if the shaft was the size of the splines because it removes the stress risers that are created at the splines, and it also reduces a little more weight. Also note the size of the gun drilled hole, that has little overall effect on torsional strength, believe it or not - *'CAUSE SCIENCE!*)
Timing belts - Positive engagement of drivertrain, not dependent on friction
- Torque/shock absorbing properties with very high durability(high stretch before failure), limits the amount of abuse drive motors receive
- Lighter total weight than chain drive, significantly lighter than gear drive
...TBC...
...there you go, guys. Like I said, it's incomplete, a work in progress, and certain sections I simply need to do a little more research on before adding my thoughts, like batteries & electronics. My knowledge of those dates back to the original BattleBots, so, to me, Victor 883 speed controllers are the sh**! Lol, but, given how many robots have gone up in smoke lately, I'm inclined to say screw the 'latest & greatest s00per d00per advanced' electronics, and stick with that old school stuff - because, assuming they're still available, there's no reason that they couldn't still work perfectly fine AND they've seemingly proven to be significantly more reliable than whatever has been letting the smoke out lately. And, call me old fashioned, but, regardless of what's available today, I would much rather control a high powered spinner with a good ol' solenoid. They don't necessarily need advanced control, solenoids are dead nuts reliable, and they can transmit truck loads of current & voltage without breaking a sweat.
I also can't see a legitimate argument for hydraulically controlled anything in an intended competitive robot, there just aren't enough readily available small scale hydraulic components that are designed for the pressures necessary in this sport, and the most successful example, the 'Robot Wars' halo bot,
Razer is pretty well known for operating at pressures WELL exceeding what was actually legal in RW(don't confuse 'Robot Wars' for a fair, legitimate competition, guys, there's a LOT of 'pro wrasslin' going on in what was an entertainment biased TV show first, competition second...).
Also, note that the 'HK/LK' comments about weapon weight vs robot weight, the '1/2 total weight of robot' remarks are pretty much place holders at the moment, more research necessary to determine how much of the weapons system weight the "typical" high kinetic energy spinning weapons take up(not to mention a high energy drum system will require less weight than, say, a
Nightmare-esque all offense vert spinner) - definitely open to advice here.
You'll likely notice that some of my superscripts are off as well, again to be corrected in future revisions.
Thoughts, comments, opinions, pure hate, post it all below, I'm curious to hears some thoughts. Mind you, these guidelines aren't mandatory for a GOOD robot - the vast majority are using Home Depot grade hardware, most aren't modular, most aren't using any kind of shock mounting anywhere(I'll expand on that later with some CAD screen shots), basically no one is gun drilling shafts(EASY weight savings - c'mon people, use your noggins!!!), so on and so forth, but IMO it's these details that can separate a GOOD robot from a GREAT robot(think Team Whyachi or Inertia Labs stuff - that's about as high end, well engineered, impeccable fit & finish as it gets!). Delusions of grandeur I'm sure, but, if I aim high enough, my failure should leave me somewhere around 'average', right? right???