Goliath And The Rough Road To Space

No one said the road to The Hackaday Prize would be easy. Many of us have been following [Peter McCloud] as he vies for the Hackaday Prize with Goliath – A Gas Powered Quadcopter. [Peter] literally hit a snag on Monday: his own belts.

Peter had hoped to be performing tied down hover tests by Monday afternoon. Weather and a set of fouled spark plugs conspired against him though. After fighting with engine issues for the better part of a day, [Peter’s] 30 horsepower Briggs & Stratton engine finally roared to life. Then all hell broke loose.

[Peter] only let the engine run a couple of seconds before cutting the ignition. In his own words, “Things were running good until the engine was shutoff. At this point one of the belt started losing tension.”

goliath-1While the tight new engine was quickly losing RPM, the propeller and belt system still had quite a bit of inertia. As the video after the break shows, the belts started flapping and caught on the propeller blades. The front right prop tip caught the double-sided toothed belt, pulling it up and over the propeller. The other end of that same belt lives on the right rear prop. It too caught a propeller blade, snapping the composite blade clean off its hub. The bent steel pulley axles are a testament to the forces at work when things went wrong.

[Peter] isn’t giving up though. He has a plan to add belt guides and a one way bearing to the engine’s crankshaft. The one way bearing will allow the rotor system to overspeed the engine when throttle is reduced. The same bearings are commonly used on R/C helicopters to facilitate autorotation landings.

We want to see all 50 Hackaday Prize semifinalists succeed, so if you have any ideas to help with the rebuild, head over to Goliath’s Hackaday.io page and let [Peter] know!


SpaceWrencherThe project featured in this post is a semifinalist in The Hackaday Prize.

49 thoughts on “Goliath And The Rough Road To Space

  1. You could also do a guide/tensioner combo. You could use a free spinning roller sandwiched between two larger diameter disks on a spring loaded armature. Of course just finding a cylindrical object with raised lips on either end could be even simply.

  2. Always leave the gas can near the death machine you are testing (just kidding around, I have done worse, no harm no foul). You might consider adding some cables over the motor connecting and bracing the opposing prop arms to keep them steady. Possibly doing the same on the underside. Good luck! I’ll be following this one.

  3. Damn I don’t think I have the adventure levels required to even attempt something like that. I’m just working on an automated lawnmower cobbled together out of some electric weedeaters and those have me on edge :P

    1. Ditto — this might be a good opportunity to switch to variable-pitch propellers.

      Alternatively, the vane-control scheme seems like the highest risk / most unknown element of this setup. Perhaps it would make sense prototyping that with a simple electronics drive system first? Almost all of the hardware and software lessons should carry over directly.

  4. I’m not an engineer, but nothing about this looks right, structurally.

    that pre-punched angle is crap. If I was seat-of the pants building this, i’d be using .028″ or .035″ 4130 with lots of n-bracing.

    it’s apalling that this thing is occupying a space in the top 50.

  5. Another good reason to avoid hand crafted custom prop’s – way too time consuming for the results – and when you break them, it’s way more effort then ordering another set from Hobby Lobby.

  6. I’m not sure I would be reusing the belts after this; if they are anything like automotive engine timing belts, any weird stress on them calls for a new belt. In an automobile, you’re not suppose to bend them too much, put them in a knot or anything else.

    This accident definitely stressed them greatly, and while they may be ok, they very well could break when least expected (like during the actual hover test or worse, during a flight test).

    1. Out of ignorant curiosity, does anyone know how much efficiency would be lost/gained with driveshafts and right-angle gearboxes compared to pulleys? I’m not a mechanical engineer, but I do know that cars can lose several percent or more in the differentials. Are belts any more efficient? It seems that there would be losses due to internal stresses of the belt bending.

      1. I’m guessing it’s more an issue of weight,belt/pulleys weight a lot less then solid drive shafts and gearboxes. Plus, it’s not needed in this application. Here, the belts are driving a prop that moves against air. Only if the belt slips does it lose efficiency. The car needs to transfer the engine power to the drive axle which has the weight of the car and the friction of the ground to overcome.

        1. The belt bends (on the pulleys) and thus heats up (they do get warm under the right conditions), so there are losses…there’s a very good reason you why belts are suitable only for lower power applications, even toothed belts have some losses associated with the heating.

          And as for not needing the shafts – I personally think that shafts and gearboxes are a must for multirotors of this configuration, because a belt like this will naturally vibrate(unless you use a crapload of tensioners and whatnot, which will increase weight and loss), both from itself and from induced harmonics from the engine, adding more vibration atop what the piston engine produces, causing more stress on the frame and making the gyros job even more difficult…
          Appropriately sized plastic gears and gearbox housings made out of composites or 3D printed plastic reinforced by wrapping fibers around them (heck, even the shafts can be made out of composites), the weight would be on par or even better then the belts, the only problem is the cost and time it would take…

          1. Belts aren’t restricted to low power applications. Go look at the top fuel motorcycle drag bikes. Connecting the engine to the remote gearbox is a gilmer or similar belt. I have seen multivee’s on large machines with hundreds of hp also. The belt actually helps cushion the drivetrain from the shock loading of the engine.

            If you go to gates online, you can specify how many hp and what speed you need, and designflex online will spit out an appropriate belt.

        2. Actually belts are inefficient, but not quite as bad as bevel gears.
          The best from an efficiency point of view is a well maintained drivechain.
          Go look at all the sportsbikes, check out the final drive train, then compare that to touring bikes, shaft drive is more the norm for exactly this reason.

          However, if you had drivechains, this would become the super death machine from hell. I’ve been chainwhipped once on a bike and they have chainguards and link explosion when something goes wrong is also very very bad. I don’t want to stand near anything using chains to drive rotors in this config.

  7. I think Peter gets an A for concept, A for effort to actually build a working prototype and enter the Hackaday Prize (the rest of us armchair QB’s really should not be so critical) and finally an A for “gonads” Lest lighten up o the guy..He has the right idea…create, test ,analyze, revise, re-deploy, test, etc. you get what im saying! “Those daring young men and their flying machines” is a great old movie from back in the early 1900’s. Everyone interested in flying went through similiar hardships and critisim’s like Peter is seeing here, but htere were literally thousands of folks trying to build the first flying machine… Then wholla the wright brothers perfected the final design to become legend. Hang in there Peter and keep pushing on what you believe in! That is the Hackaday way right?….I mean innovation happens a little at a time and look how much you are learning!! Great Stuff everyone!!!! PS. I would however suggest putting a cage around that machine to prevent any flying debris from hurting anything and lastly perhaps a handy fire extinguisher? REALLY COOL PETER! Check out that old movie mentioned above as it will make you problems experinced today seem a little easier to forget. THX JOHN

    1. Those Magnificent Men in Their Flying Machines was a movie made in 1965
      http://www.imdb.com/title/tt0059797/
      It depicted a fictional time in the early 1900s though. Heck it is shot in color, you didn’t really think it was made in the 1900s did you? It is a funny movie though. it’s supposed to be, it is a comedy. According to the trivia page on IMDB the planes were real, and based on real planes too. I thought all of the aerial stuff was totally faked. I’m going to have to watch it again now in light of this new revelation.

  8. As an RC heli pilot, for the prop arms, I’d go with thick enough pipes, just as tails are made in 700-class model helicopters.They enclose and protect the belts, which make a comeback recently in RC heli world. Also, I agree that tensioners are a must-use here. Current RC helis use them in main rotor drivetrain, coupled with high-grade toothed belts. In 3D 700 class helicopters, momentary power of brushless motor can rise up to 11kW, that’s half of power of this gasoline motor.

  9. I gotta wonder why the hell he went with a Briggs 4 stroke engine. He easily could have produced the same power for half (or less) the weight with a 2 stroke engine from a dirt bike or something similar. B&S makes big, heavy and poorly optimized motors primarily for stationary equipment or vehicles where weight doesn’t really matter (lawn mowers, small tractors, etc.). They’re reliable, but that’s probably the worst choice he could have made for the powerplant of a vehicle that has to fly.

    1. http://www.gamasutra.com/view/feature/132500/dirty_coding_tricks.php?page=4

      ” We cut megabyte after megabyte, and after a few days of frantic activity, we reached a point where we felt there was nothing else we could do. Unless we cut some major content, there was no way we could free up any more memory. Exhausted, we evaluated our current memory usage. We were still 1.5 MB over the memory limit!

      At this point one of the most experienced programmers in the team, one who had survived many years of development in the “good old days,” decided to take matters into his own hands. He called me into his office, and we set out upon what I imagined would be another exhausting session of freeing up memory.

      Instead, he brought up a source file and pointed to this line:

      static char buffer[1024*1024*2];

      “See this?” he said. And then deleted it with a single keystroke. Done! “

    2. I agree that a 2 stroke would be lighter, but if he wants to over-rev the engine with zero throttle, it has to be a 4 stroke in order to stay lubricated.

      I suspect that he chose the motor he did because it has a horizontal stroke. You can’t just lie a dirtbike engine on its side and expect it to keep working, and making a gear linkage wouldn’t be feasable.

      What would work great is a small wankel rotary engine. Light weight, 4 stroke and very high RPM would make it perfectly suited. Only poor fuel consumption would let it down.

      1. The only time the engine would be turning for any length of time at closed throttle would be while the blades are spinning down. Motorcycle engines are designed to operate without lubrication for brief periods while coasting down hills or decelerating.

        The beauty of two stroke engines is that you can turn them on their side, or upside down or whatever without affecting the operation of the engine whatsoever. The only thing that would need to change is the orientation of the carburetor, which would simply require a custom intake to overcome. Two stroke engines are ideal for small helicopters. That’s why they are typically used in IC model helicopters.

  10. did the faa abolish the 50 pound weight limit? because it looks just the weight of the engine and steel bars combined should exceed the weight limit or did the faa write the limit as a cash cow (betting on passing of law).

  11. There are so many things wrong with this…. Dump the erector set angle iron and learn to tig weld. The biggest problem is there is no way to control the thrust of the props. So no stability. Please stop and increase your knowledge level before you cause an incident that will make the powers that be clamp down in drone development. All we need is this thing crashing into a playground full of kids. A bunch of crying overprotective anyway mommies would be the worst PR possible.

    1. I think you’re being overly harsh here Mark. Go back and read [Peter’s] project logs. He plans to use a vane system to control pitch, roll, and yaw.
      Vanes are probably not as “positive” a control system as variable pitch props, but they are proven in full scale designs- the easy example here is hovercraft yaw control.

      A few years ago anyone involved in the R/C hobby would have told you that controlling a fixed pitch quadcopter by motor RPM alone was damn near impossible. Thank goodness no one listened to them.

  12. Some constructive comments in here already but I’ll give it a shot of my personal list.

    Ditch the four stroke engine, for weight find an equivalent two stroke motor then ditch its oiling system and other un-needed parts and run it on premix in a lightweight nylon fuel tank. Some of the mini moto engines are amazingly light, and wont care what orientation they are ran in so you don’t have to specifically seek out a vertical crank variant. If you repurpose somethng like this it will probably need an outrigger bearing to support the outside of the pulley on if you mount the belt sprocket right on the crank.

    There is a LOT of belt wrap but small diameter pulleys, small diameters induce more bending to the belt and more frictional losses as a result. If they are solid hubs, a larger sprocket with a lightweight flange style hub would improve things at no extra weight penalty. It should also help you with the belts contacting each other on shutdown. Bigger diameter does mean longer belts which means more weight in the belt itself however, some thought and weight calculations would have to be done here…

    The belt itself is a veebelt, you will see 12% more efficiency from using a syncronous belt with say curvilinear tooth profile. The sprockets dont HAVE to be cast iron or aluminium, you could use nylon sucessfully and loose weight in the process for the amount of wear likely to occur in short flight test times.

    If its shedding a belt, arm whip is contributing to this, there is not a lot of bracing at the ends of the arms and they are waving round with shutdown harmonics, I think if you added two smaller diameter tensioned wires between each leg and its neighbor, it would stop the tips whipping about with very little weight penalty. You could expand this with a central spigot and guy’s off it to brace in the Z also.

    If you are thinking about switching to gear drive, bear in mind that a gas engine produces power pulses, so it whips the transmission system about with each large pulse per firing cycle, and drags the train for the other cycles. A belt cusions the drivetrain from this cyclic nature so a gear system would require similar dampers somewhere. Motorcycles with gear primary drive use a rubber drive cushion usually inside the clutch hub.

    I really hate the dexion, I would like to assist in knocking something up in lightweight aluminium but I live in europe so its a bit of a blocker, but you must be able to find some other home shop fiend who would be interested in playing nearer.

    You really need some simple nylon slipper guides halfway down the arm lengths, they would be lightweight and suitable for taming the belt a little. Again, go look at a motocross bike chainguard setup, it guides the drivechain inside a nylon window. Or cotton reel style spools for it to run on.

    Also is there no way to make this one continuous long belt? four belts will stretch unevenly and give unpredictable results, probably your control can cope with this but its not great practice. You can get double sided belts to ease this if its possible.

    Having said that, good effort for taking it on!

  13. Im still in the “it’ll never leave the ground camp. aside from the crazy flex in every possible component of this thing that is going to eat any power that might have went to thrust, I don’t think a lawn mower engine is going to overcome the thrust to weight ratio problem anyway. a welded aluminum frame with a proper ultralight rotax engine would be where I’d start, distrubuting all that power is another issue entirely.

  14. It is my understanding that Quad coppers are controlled by the difference in RPS of the individual motors/rotors
    IE Left motors power up 10% to move the Copper right, Back motors power up to move the it forward. ext

    Also The props that move in the same direction are kiddy corner not side by side, this is so when you drop One pair down 10% and the other pair up 10% the Lift stays the same but the toque will rotate the copter left or right,

    How are you getting any control other than Throttle with all your rotors spinning at the same RPMS???

    other issue
    Top heavy??
    Props too small?

Leave a Reply to RichCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.