Hack a Day

What i see here is a flying lawnmower. And that's exactly how I'd use it.

I'm glad that you put a disclaimer "Think before you start" at the beginning of your instructions, but you go on to say "it's only as dangerous as the user". That's exactly the problem! There's plenty of fools out there, and I'd hate to imagine one of these machines in their hands. And what happens when you have engine problems in mid-flight, or a belt breaks?

Yes there are people out there who could do bad things with Goliath, but there are many other things that they can get there hands on as well that'd be dangerous.
Once I start flying more than a few feet off the ground I intend to install a ballistic recovery chute. In an emergency the parachute is shot out using compressed gas or a rocket. There are some sized for ultralight aircraft. So that would be ideal. It'd be nice to have redundant belts, but I don't think there's enough weight margin for that.

I am interested in designing/ building a recovery system for this monster! I have emailed you.

I look forward to hearing about your ideas. I'll contact you soon!

This is by far one of the most compelling quadcopters I've seen to date.

Thanks Mike!

Im designing my own heavy copter drone. Several points just looking cursorarily at your build out:

1. Placing your heavy motor on the top makes it inherently unstable. Far superior would be placing it underneath the center of gravity, not above it. The difference in stability is between either balancing a basketball on your finger tip (your design) or suspending a basketball from your finger on a string (like a helicopter with the aircraft body below the main rotor).

2. Hexa would be more failsafe than quad. On a hexa if two (or even 3) props go out, assuming the software can detect and compensate for the loss, the thing can remain stable and keep flying (long enough to return to a safe landing).

3. Your frame and engine look too heavy. You may be over building a lot of it, and galvanized steel and a heavy cast iron engine block may kill your efforts to get airborne. Look into ultalight aircraft engines or even mor advanced stuff. Yours looks like a generator or lawnmower engine which is designed for an application where engine weight is irrelevant. But for you, engine weight is VERY relevant.

4. You need a way to individually control rotation to each prop. Consider a fluidics type transmission. Basically, your engine is attached to a pump and creates continuous water pressure. If no torque is needed to any props, the water (or oil) recirculatesgoes around in a continous loop. If a prop needs torque, a valve shunts water into it in varying degrees to a reverse pump that drives the prop. In this way you could control all props variably instantly... and the motor would run at its optimal constant speed.

5. I myself want to buildone large enough to be controlled by and carry a pilot underneath. So I'm thinking even larger than your design...

CG

You should post your design to HaD.io! Definitely interested to see your project.

I'd counter point (4); pumps are heavy. Hydraulic motors are heavy. Control valves are heavy. I could see using this for control surfaces (like an airplane), but the flow rates and pressure requirements needed for the propellers would make such a system difficult to integrate into a flying platform.

I agree with Zakqwy, you should post your project to HaD.io. I'd like to go bigger as well, but this is step 1. In regards to your comments:
1. Yes it does make it unstable, but I haven't found a light weight way to do sling the engine underneath. The electronics will at least address the stability. Goliath should at least be slightly more stable than Gimbalbot (http://hackaday.io/project/996-GimbalBot) :p
2. Going to hexa is an intriguing idea, but since Goliath theoretically has a extra thrust margin of about 10%, even going to hexa wouldn't help much. It still just fall fast.
3. Yes mass is the number one issue in making sure Goliath works. I've traded heavier mass for reduced cost and effort for Goliath, because I know there is going to be a learning curve and I'd rather wreck a this design and learn a few lessons than a higher performance and higher cost vehicle.
4. I actually did research doing a hydraulic design and as Zakqwy pointed out, it does get heavy. The pressures involved lead to very heavy motors and at least at the scales I looked at, it didn't work out. I also looked at the same concept using pneumatic design which was also interesting, but there's a lot of energy losses with the expanding air and there would need to be some sort of thermal recovery system.
5. I'd love to see your design! I've gotten great feedback from the community here at Hackaday Projects.

66
Placing your heavy motor on the top makes it inherently unstable.
99

This is not true: https://en.wikipedia.org/wiki/Pendulum_rocket_fallacy

That is quite a software thicket there sir. Have you considered abandoning the 'Arduino Way' and going for Atmel studio with JTAG/ISP through an inexpensive programmer like JTAG/ICE. Source level debugging and more available memory (no bootloader) and you get to use the USB port on the board. As you may have discovered the core of the Arduino platform is thin and they do stupid things like hogging timers for beeps and unnecessary delay functions. Unless you really want the IDE backward compatibility, but then it seems you had to fork that as well. I ended up doing it this way so I can speak to the benefits. Good luck.

I do some programming, but haven't done any on a microcontroller yet, so almost all that went over my head in the first read. After googling most of what you wrote, that sounds pretty intimidating. (This coming from the guy who's testing a giant gas powered quadcopter in his driveway).
Now that I've done a bit of research, I agree with you that it would certainly make sense to go the Atmel studio route, especially for follow on versions of Goliath, and have an optimized bit of software. I do like the ideas behind the Ardupilot software and would like to contribute to their community. Also the singlecopter (http://copter.ardupilot.com/wiki/singlecopter-and-coaxcopter/)
has demonstrated the control system route that I'd like to do, so I can possibly leverage the Ardupilot software already written for that.

Sorry. I guess a distilled version of that would be: you have so much code there that you might consider doing it the way the professional embedded systems developers do it with single step debug and viewing memory/registers etc instead of the way hobbyists do it Arduino style with printf and such ;] You can still use all the Arduino libraries this way, which is the bulk of what 'Wiring' really is. I think their platform is great for little one-offs but at a certain point you need real tools.

Highly inefficient, but for simple control (until you can design something better) you can place a flat baffle under each prop that slides on the frame from the center outward. It could not only reduce the effective lift of a prop, but also shift the CG toward that prop, inducing roll in that direction. You probably don't even have to cover more than 1/4 of a prop to effectively reduce lift. Also, your actuator only has to overcome sliding friction since the aerodynamic force is perpendicular to the actuation direction.

It's a good idea. My only concern is reducing the prop thrust. I don't have a lot of excess thrust currently so I might not be able to implement it. My hope with the vanes is that since airfoils provide more lift to drag, they'll able to produce a large amount of side force with a minimum of thrust reduction.

I say baby steps first; get your thrust:weight ratio over 1 and sustain a constrained hover. I posted a link to the Project Morpheus video archive which has a lot of good test setups. Definitely worth a look!

Can't get this project out of my thinking. Now I believe the reason to be your rotors. Your rotors are not opposite of each other, that is going to throw your torque off.

Should it not be front left and right rear same rotation? Opposite corners rotate the same?
not parallel.

John

Electric quadcopters have the same rotation on opposite corners to do yaw control. They speed up the propellers that spin one direction and slow down the propellers that spin the other direction. This allows a torque difference that spins the vehicle, but the total thrust is balanced across the corners. Having them parallel still cancels out the torque, but if you tried to do yaw control, one side would drop.
Goliath has the props spinning parallel to allow the the belts to wrap around the drive pulley more and increase the torque transfered to the belt. Since I won't be using differential thrust, I can get away with doing parallel.

Peter,

Had a thought, for the purpose of testing only, your details show the support under the main body of the unit. If you shift your support to under the propellers, you would have a better view of the stress caused when it is try to lift using them.

If your struts do not support the motor how can it lift it under load?

John

Goliath is capable of supporting itself at the propellers. The saw horse are under the center to take up less room in the garage and allow the shop crane in and out.

Peter,
using the Metal framing for your structure, do you have the ability to put an extra bend into one face of it for a second angle. It should make the overall structure much more rigid, without adding weight?

John

I have a sheet metal brake, that's in pieces. I could use it to add the bends. The current frame is intended for prototyping.

wow, this is incredible. why did you choose to go with a belt drive over shaft drive?

Thanks! After researching belt, chains and drive shafts, I chose a belt system because it was light, could handle the RPM/torque I was targeting and able to handle shock loads well. If I use drive shafts I'd have to have a gear box at each propeller to change direction as well as a a larger gearbox at the engine to attach multiple shafts.

Peter,
I like the idea, but not the approach. You should consider slightly different approach. At least research drive system stress details, Torque details, and give some planning for rotation speeds required to gain lift.
How do you plan to change altitude?
Could you use some sort of flywheel to aid in rotation control?
Consider using a less direct drive to change the stress points to a better location, while increasing rotation speeds.
There are multiple thoughts that could be helpful depending on some of the requirements you are looking for.
Bicycle or motor cycle drive systems or indirect drive systems would be a good place to start.

John

Thanks for the feedback John. I did start this project looking at motor cycle, bike and aircraft drive systems. I have sized the components to the loads, as well as done the calculations for lift, I just have documented the details. I'll have to get those added at some point.

Peter,
I noticed your framing flexed and that is with no load, you might need to add either tubing or reverse angle framing to reinforce your struts. have you considered using chains instead of belts as they tend to flex less. Also was wondering about your shaft sizes, your shafts should all be close to the same size for the torque they are receiving which should be similar.
I don't know if any of this is helpful but hope it works out for you either way.

John

I do need to fix some of the flex. I'm still debating on the best method that won't add too much more weight. The shafts for all the props are 3/4" all thread.

Hi Peter,
What are the props made of? I look at this and see a 810cc motor that must be 30+ Kg and then I look at the pitch and surface area of the props and wonder how they don't tear apart from centripetal force at the prop RPM you will need to lift 30+ Kg.

The props are made from a foam core with wood stiffeners and then covered with 3 layers of 9 oz fiberglass. I have a few project logs detailing the progress, the last is: http://hackaday.io/project/1230/log/6507-not-so-rapid-prototyping

You might want to take a look at how other belt powered propeller jigs have worked.... https://www.youtube.com/watch?v=rG9clKE6268 shows a universal hovercraft UH-14P.... There is added weight, but these belts really whip around.... constraining them like shown is a good way to keep them out of the prop

Hey Peter, Sorry you're having trouble starting the engine. I'm no small engine expert, but I've fought with a few of them in my day.
You know the old saying - gas engines need air, fuel, and spark. You know you're getting fuel to the carb, but is it letting that fuel in to the intake. (closed needle valve?)

Spark - the easiest way to do this one is to place the spark plug wire somewhere near the engine, and look for a spark while cranking. (You want to disconnect your props for this)
You could also disconnect the entire spark plug, touch the threaded portion of the plug to the block, and check for spark. If you're not getting spark, check your ignition system - sometimes these engines have a low oil cutout, which could be causing you problems. (You did put oil in it, right?)

Finally air - check for a clogged air filter, (could be packed in a plastic bag from the factory).

Thanks Adam. I did put oil in it, but perhaps it needs more now that it's been circulated around a bit. The air cleaner wasn't wrapped, but I did remove it to get at the carburetor and left it off for the last few tests. I had been leaving testing the spark until last since the gas setup is sketchy, but if the oil doesn't work I'll try those spark tests.

Hello Peter, you need to add a few more bits of documentation on Hackaday Projects to give your project the best chance of going through to the next round of The Hackaday Prize.

By August 20th you must have the following:
- A video. It should be less than 2 minutes long describing your project. Put it on YouTube (or Youku), and add a link to it on your project page. This is done by editing your project (edit link is at the top of your project page) and adding it as an "External Link"
- At least 4 Project Logs (you have this covered)
- A system design document (I can't see one. You should highlight it in the Details)
- Links to code repositories, and remember to mention any licenses or permissions needed for your project. For example, if you are using software libraries you need to document that information.

You should also try to highlight how your project is 'Connected' and 'Open' in the details and video.

There are a couple of tutorial video's with more info here: http://hackaday.com/2014/07/26/4-minutes-to-entry/

Good luck!

Wow... This reminds me of Terminator 4...

Forgive me if I am stating something obvious or dumb, cuz I'm a complete newbie to this area , but have you considered making this as a tri-copter? You wouldn't need custom blades since all the blades are the same spin direction, the yaw control and flight stability is better, plus three blades instead of four would mean lighter design?

If nothing else, it'd be nice to understand the reasons why you chose a quad-copter over other designs.

I wanted to do a quad-copter because by having two clockwise and two counter clockwise propellers, the torque from the propellers will cancel out. To be honest I hadn't considered a tri-copter. Tri-copters tilt the rear rotor to offset the torque like a helicopter tail-rotor. The belt system doesn't allow the blades to tilt. I guess Goliath could be built as a Tri, but the the control surfaces would have to be bigger to compensate for the torque.

I just discovered this electric clutch and figured it might interest you, at least you must agree it is interesting: http://www.ebay.com/itm/Electric-PTO-Clutch-for-Scag-61-72-Hydraulic-Drive-3-Wheel-Riding-Lawn-Mowers-/121060668200?pt=LH_DefaultDomain_0&hash=item1c2fc73328

This is interesting. WillyMacD had suggested electric clutches, but I didn't find these when I was looking around. I'll have to look into these.

Invent a pulley that can expand and contract by turning something. This will give you prop speed control, and thus allow simple control by all the normal RC control baords.

That would certainly make things easier. CVTs use a conical pulley to adjust the radius to do what you're talking about, but I don't think it'll work with a toothed belt. An expanding and contracting pulley would have to allow the teeth to slip in some manner. Perhaps a Derailleur setup might work, but I don't think it'd be responsive enough.

I agree it's hard. The more I read the others comments the more it really sounds like you should spring for 1/5th scale helicopter assemblies and blades, then simple servos can adjust pitch and therefore thrust just like an electric quadcopter. I just created a guided rocket powered by the 4hp edf from Dr. Mad Thrust, with vectored thrust by flaps, flaps suck, couldn't get stabilization good enough with the multiwii controller. Just a heads up. I've given up with flaps for good thrust vectoring and am changing the design of the craft and giving it swivel nozzle styled vectored thrust.

Steve, I'd like to see your project if you do a gimbaled thruster. In my experience, it's not easy.

For the propellors; when I built my composite uav wings I used the scraps from the hotwire cutting as a support for vacuum bagging. Since you mill your cores you may want to mill shells that fit around the props when bagging? Also some UD carbon on both sides of the prop will increase the bending strength :)

Thanks! Milling support shells sounds like a good idea. I do plan on eventually switching to carbon to increase the strength, but I'd rather get all the kinks worked out with cheaper fiberglass.

Awesome project! Would love to see it fly.

Thanks! I'm looking forward to seeing it fly too!

Whats your current plan for thrust control ? You could use some kind of wing, controlled with a (big) servo which can increase its produced air resistance under a rotor. (like flaps on a airplane)

Thanks for the inputs! The current plan is to use vanes similar to how hovercraft steer. That'd be a good way to control pitch and roll, but I'm not sure how well that'd work for yaw.

Peter, I'm planning a similar albeit, smaller project. For thrust control, I was bouncing the idea of using electrically controlled clutches on the shafts of the props. Perhaps this is something that may work for you as well as with minimal changes to current control software. cheers and good luck

Thanks for the input! I'd really be interested in hearing more about your project if you feel like sharing. I don't know much about electrically controlled clutches but I'm definitely intrigued. The control system isn't set yet so it's a possibility. Do the clutches act simply in an On/Off manner or can you vary much much they engage?

Ever thought about using the fuel engine as an electic generator? I understand that's not really helpfull at this point, but in my opinion thats a better way to go. Im curious about the flight dynamics of a mechanical sytem like yours. Maybe you add some serious nonlinearities to the system regarding control theory. ... Or maybe it'll just fly as heck. Have fun.

I had thought about coupling the gas engine with an electric generator and then using electric motors (hybrid system), but it's just too heavy at this scale. For aircraft the engines run near the optimum RPM most of the time, so the advantage of a hybrid system isn't as great.
I'm really curious as to how the controls are going to behave as well! I'm not expecting Goliath to be very nimble, but we'll see. Thanks for the feedback!

You might consider emulating what you see in an ordinary A/C outlet in a car....3 or 4 vanes controlled by a single horn sitting outside the duct. Air can flow down the duct and then down and the vanes can direct the air straight down, or horizontally perpendicular to the quad's arm/duct. That gives you pitch/roll and yaw. the degree the vanes move controls the vectored force at the edge of the arm and so altitude and direction. You can also control airflow through the duct with a butterfly valve inside the duct also controlled by a horn (someone mentioned that as well, I think). What I think is going to be super interesting is the flight control software...

Did someone already suggest this....I know you are using fixed pitch props, but there have been developments with constant speed props that don't use a governor and use small electric motors to vary pitch. Might make this too complicated and heavy. If you consider ducts you are going to need quite the concentrated blast out of each tip. The Harrier uses a rotating valves with vanes to vary the thrust direction, you might get some value out of just nozzles that open and close out the end....at any rate this is so very interesting I'm looking forward to seeing it fly.

Thanks for the feedback! I've been really hesitant to look into variable pitch props. Maybe I'm just biased towards the fixed pitch for simplicity, but one big concern is the loading. Each blade will be supporting 30 lbs in hover, which will provide a large moment at the blade root and make any variable pitch hardware heavy.
I really wanted to have the ducts be part of control system and opening and closing the nozzle at the end of the ducts would be a great way to do that. I'm just not sure how to implement it. I had thought about using nitinal wire to expand and contract the nozzle, but I'm not sure how responsive it would be.

Wow! that's one big quadcopter - thanks for submitting Goliath to The Hackaday Prize! I'm just curious how you'll get the props spinning in opposite directions to avoid torque issues?
Keep up the good work, and don't forget to document your flight control system - I'm curious to see what sort of system you go with - vanes and servos or something completely different!

Thanks! I'm planning on posting about the drive system later today and that'll show how the props spin in opposite directions.
I think that vanes and servos might be the best choice. All the other methods I've considered wouldn't handle yawing the vehicle well since all of the propellers rotate at the same speed.

Did you considered Continuous Variable Transmission http://en.wikipedia.org/wiki/Continuously_variable_transmission instead of a gearbox?

It might help with the thrust distribution too.

I had not considered a CVT. Goliath doesn't have a gearbox, but it does use belts. Maybe a CVT could be used to vary the propeller speeds. Thanks for the suggestion!

This sounds so dangerous/awesome, those two words do go hand in hand you know. :D

Im gonna pay attention to this, I may have to recreate your results when you finish.

Thanks, I'm glad to hear someone else is thinking about building their own!

How to you plan on adjusting the thrust vector? Considering the belt drive I would think adjustable pitch props would be WAY easier mechanically, and provide more controllable result.
Also check out fiberglass poles for windsocks and such, using ones as a driveshaft inside a larger one as the support might be a lot lighter than a frame rigid enough to handle all that belt tension and let you space the props out further.
Or possibly even just using the engine as a generator and plopping electrics on the corners. ;)
Looks like you're already well underway but I hope the ideas help.

Thanks for the inputs! I had looked into driveshafts and while the fiberglass would be light, I'd need gearboxes at the center and at the propellers. I was really hoping to build a gas/electric hybrid since control would be simpler with electrics at the propellers, but I think the quadcopter would have to be even bigger before you can carry a generator and motors to go with it.
As for the thrust vectoring I haven't decided for sure yet, but I'm currently thinking servo actuated surfaces or something integrated with the ducting.

I've been doing a lot of design work trying to optimize thrust:weight ratio; while I plan to do some real-world tests to validate prop and motor selection, this site has given me a good quick starting point:

http://personal.osi.hu/fuzesisz/strc_eng/

It allows you to input prop size, pitch, type, air density, amd RPM and spits out a thrust calculation and motor power requirement. I haven't dug in to the tool's calculations, but it may be a good starting point for you. I'm guessing your belt-driven design will give you plenty of alignment flexibility but will also limit your maximum RPM, requiring a fairly good size set of props.

It's worth mentioning: 30 HP is A LOT of power. I suggest doing some research on typical maximun safe belt RPMs and using this as a primary design constraint to minimize safety concerns; in addition to whipping about and damaging stuff, a broken drive belt would likely result in a catastrophic and unrecoverable loss of flight control. Might be worth taking the weight penalty to equip the chassis with some kind of belt guard.

So I'm looking at this project log of the motor with it's cover off and I got an interesting idea. This is a little off the wall and mostly half-baked brainstorm kinda thing but maybe you can think about it.
instead of a frame, belts, pulleys, etc.. mount the biggest blower fan 30HP can support directly to the motor. Then construct the frame as a big hollow X shape where the air gets ducted out the corners. maybe some in the center too but the corners will have your airflow vane directional vectoring control going on. Like a big hovercraft thing. :D
Just wrapped fiberglass or something would be pretty lightweight.

Cool idea, Dave. Isn't that how the Harrier and the VTOL version kf the F35 work? Definitely easier to actuate louvers then props; in addition to maintaining belt tension during vectoring you'll need to consider gyroscopic effects of the spinning props, which might drive up actuator torque requirements.

Harrier turns the turbine exhaust straight downwards so yes kind of similar to that but we wouldn't be having a forward facing intake.. F35 I think has a separate turbine for VTOL? Not sure.
It occurred to me that in this x shape ductwork thing you could also have some throttle-body style valves to proportion the outputs.

Yup, butterfly valves of some kind would work. Turndown isn't great but they would only need enough rangability for angular velocity changes; overall ascent rate could be controlled using engine speed and maybe trimmed for response using the dampers. I'll bet you could actually salvage the throttle plate mechanisms out of a few old engines, the basic design requirements are likely similar.

Thanks for the link to the information, I'll have to check that against what I've calculated so far. Yes 30HP is ALOT, but the frame and ducting will encompass the belting. Once I go beyond tethered flying, I plan to incorporate a Ballistic Recovery System to help with any failures.
Also interesting idea with the blower and jetting the exhaust. I'm not sure if that method would work with this engine's weight and power though.

Are you using some kind of swashplate system? How are you varying prop pitch? This is quite awesome.

The plan is to use fixed pitch props and to vary the thrust direction for control. The controls haven't built yet, and there are a few methods I'm considering using. Thanks for the interest in the project!

Cool! We should compare notes, I'm putzing about with gimbaled fans albeit on a far smaller scale. Looking forward to seeing your design.