Servers by jDrones

Petrol Boosted Tricopter

The “Trinity” Tricopter with petrol supplement drive.


At CanberraUAV we have recently been mucking about with all sorts of VTOL aircraft with the general aim of finding something for the UAV Outback Challenge that is perhaps a little more elegant than the quadplane we used in 2016.

Jack had often mentioned the idea of putting a petrol engine in the centre of some form of multirotor. This concept is not new, it has been done before:

The concept is fairly simple, build a multirotor with a fixed internal combustion engine to provide the majority of the thrust, the electric motors provide mainly control and stability. All in the hope the endurance of the combination is greater than electric alone.

The Build

So I thought I would have a crack at it.

The obvious problem of the internal combustion engine is the torque produced by it needs to be counteracted. I chose to use a tricopter as the tilting tail out of the box provides the flexibility of providing the stabilisation in all situations whether the engine is running or not. So I started hunting for a tricopter/engine/motor combination.

I wanted the electric motors to be powerful enough to fly the machine in the case of an engine outage. I also did not want to go overboard and kept it modest in size. In the end I came up with the following set of main hardware components:

Turnigy Talon (350g)
An RCTimer combo (300g)
RCGF 10cc engine (623g)

Batteries, fuel, tanks and other hardware it all came up a little over 2kg. The 3 electric motors @ 4S suggest about 4Kg of thrust, so I felt my goal of the electric safety net was possible. The electric motors I eventually got enough to run all clockwise opposite to the engine so as to naturally counteract as much engine torque as possible.

For the flight controller I chose a PixRacer as the 3 Rotors, tail and throttle tail servo and ignition kill switch added up to 6 channels. I really did not have the space for anything larger.

The build came together fairly quickly over a few evenings the tricopter was largely unmodified with some aluminium plate and some polyethylene blocks improving rigidity due to the extra weight. It needed boots on its legs to accommodate the belly fuel tank. Add some anti vibe mounts for the GPS and flight controller.

Google Photos

CanberraUAV had a build day and the extra help allowed the build to be completed and some electric only flights performed:

The next day we fired up the engine, with the throttle servo being directly mapped to throttle on the transmitter sticks. In this sense it was pilot controlled and the autopilot had no say in the throttle position.

It quickly became apparent after these flights we needed a better way to control the ratio of electric to petrol thrust and let the autopilot have some say in proceedings. Tridge quickly made some changes:

And a fortnight later we tried out these changes for real:

As you can hear in the dialogue, for stable flight the electric motors are running at about 6A with the engine running as opposed to about 30A when on electric only with the copter just barely staying in the air. The small 2.4Ah battery was not able to deliver sufficient current nowhere near the 17A/motor the spec sheet suggests.

All in all things are looking promising thus far.


  • Tune the engine. This is proving to be troublesome as is often the case on small engines.
  • Move to a 5Ah main battery to improve the peak electric thrust and overall endurance.
  • Experiment with prop sizes to provide a more responsive thrust profile from the engine. The flights shown had a 13x6, we will try 13x5 and 13x4 next.
  • Is the gyroscopic precession effect of the heavier engine/propeller combination require consideration in the controller ?
  • Optimise the fuel tank and battery size for endurance and get some real numbers.
  • If the numbers stack up, build a bigger one!


I have to say this build has been a lot of fun, and came together very quickly when compared to helicopters of similar size. The ease at which Arducopter was adapted to this application is testament to the hard work of many.

Many thanks go to Tridge and Greg for wiring and footage and to the test pilots Stephen and Justin. As well as everyone else who helped even those who held it down while we attempted to tune the engine.


Really great advance and thanks for the excellent write up. I’m going to take sections of this and add to our wiki!

Nice. Just a little bit scary… :fearful:

Using a petrol motor will add hover time but will it also add range? Would some tilt rotors or pusher prop help to get extra range? What is the intended use?

I actually worry about the stability of that without cyclic pitch. I already tried that same concept, although with an electric lift assist. The stabilizing motors were never running at enough power level to keep it from crashing once it got really moving. I removed the electric assist motor and tried a gyrocopter version of it that was a total failure because it needed cyclic pitch to make it work right.

I eventually went to a winged TriCopter that was much better and had better flight time. But it has serious problems in the wind.

In the end I came to the realization that rotary wing aircraft were not meant to fly without cyclic pitch, and multi’s are basically a proof-of-concept that if you strap enough horsepower to a cement block you can make it fly :astonished:

Yes I agree, It would not greatly extend range when compared to anything fixed wing. But would hopefully provide a general increase in endurance for typical multirotor applications (where noise was not a big issue) The intended use is mainly to have some fun :slight_smile: I can’t honestly say it is a good fit for the OBC.

Yes, The last video clearly shows some stability issues. Whether that was a gyroscopic effect or that the stabilising motors simply stalled and lacked effect is not clear yet.

If I went to the next size up I would probably put variable pitch propellers on the stablising motors and keep their RPM up high (but current low) so they can be more responsive but have low current draw.

The moment of inertia of long weighted tipped helicopter blades would be way above what is going on here, so I am not sure it is quite the same problem. (not that the gyroscopic effects should be ignored altogether… )

I’m guessing, based on what happened to mine when I tried this (albeit with electric assist) that the ratio of lift between the assist and stabilization was not allowing the stabilization to do its job. Using tail rotor assemblies off a helicopter on the stabilization motors, running those motors at constant speed and using servos to change the pitch was one idea I had. But then you may as well just go with a helicopter that is way more power efficient and has less blade tip losses.

When I tried it as a gyrocopter/TriCopter hybrid, the drag from the gyro rotor killed it. Without cyclic pitch that just didn’t work.

I completely agree that the stabilizing motors need to have a certain rotor loading to be useful as attitude control. Reverse thrust could be an option but collective pitch control would be better for response times, unless you use really small control motors and props that in turn wouldn’t be able to fly it without the ICE on… All this is getting very close to a helicopter design, in fact a coaxial heli setup might work as well to hybridize battery and ICE power.

I understand it’s for fun and to gain some experience with that type of configuration, but if the intention is long hover times then a lager disc area is key to efficiency. In this case the power density of fuel is giving you longer hover times (and sacrificing range) but at the same time it seems to be introducing control issues. The point of the ICE is to reduce current draw on the control motors, but that in turn is reducing their loading and their ability to control the aircraft. I’m thinking one easy way to fix this might be to increase the tricopter arm length, so that a small amount of thrust will produce the desired control over a longer lever arm. It might also help reduce any downwash affects from the ICE rotor on the control props.

A variation might be to use a asymmetric tricopter instead. So have differential tilt front motors (for roll/pitch/yaw control) and a single rear larger rotor running of the ICE. This would require a reliable ICE propulsion system however, and would leave you without electric only recovery. (even with a larger battery! :slight_smile: )


I think a tilt rotor tricopter has merit and will operate in wind just fine provided there is some dynamic control of attitude to better counteract wind interactions with the wing areas. This is to allow the wing to have a positive angle of attack in hover and slow forward operations into the wind (along with windvaning) and will in turn improve in wind hover efficiency and operation overall.This is something that needs to be added to Ardupilot I think, I haven’t seen it yet, and we’re currently doing this mechanically by tilting the quad prop mount on a QP.

With cyclic pitch do you happen to mean collective or CoG/disc pitch control for the gyrocopter? The control stick input on a gyro is typically neither cyclic or collective, rather a form of CoG control by changing the the angle of attack of the rotor disc in relation to forward airflow. On a balanced tricopter airframe I’d have expected the gyro to work, provided the gyro rotor was driven under partial load by a motor for aircraft pitch control. What I mean with that, is that although a “real gyrocopter” has a completely unpowered rotor, in the case of these tricopter tilt hybrids, that does not need be the case. Having a powered main rotor for takeoff and landing that then “unloads” the motor in forward flight because of the gyrocopter effects of forward airflow that produces lift, is still a gyrocopter in my opinion. The easiest way to test this is to strap a pusher prop to any quad and watch how hover current is only moderately lower than at full forward velocity.

Any “rotor” (being something I define as operating mostly parallel to the earth surface on a aircraft) is not very efficient in producing a sideways force for forward flight. Any prop (for me being a propeller that can operate in all orientations) oriented in the direction of the required thrust will always be significantly more efficient.There’s also a key difference between a gyro and a heli, in that the heli rotor is imparting a force on the air to produce lift, whereas a gyro is essentially absorbing a force (forward airflow produced by the props) to produce lift. The improvement in efficiency is more than one would at first think, but becomes obvious if one considers the efficiency of motor/props in comparison to freewheeling rotors (wind turbines) that also operate in and benefit from less turbulent airflow overall.

Broadly, there is no substantial difference aerodynamically between a wing and a spinning wing called a rotor or propeller. Both produce lift by moving a surface through the air, the only time that spinning a wing is beneficial is if you want the aircraft to not move as fast laterally as is required to produce enough lift from the wing to stay airborne. So the slower you want to move the aircraft laterally the faster the wing will have to spin. In the case of a gyrocopter there is a lower limit to airspeed at which point it will not produce enough rotor RPM and enough lift and so it will descend. This lower limit can be “artificially” assisted by using a motor to the point stationary hover is possible. The beauty of a brushless electric motor is that it can operate at partial loads and therefore at lower current draw, regardless of their RPM. This is ideal for an motor “assisted” gyrocopter and takes advantage of several positive attributes of the various platforms and components used, and at the same time facilitates attitude control.

There is one more item worth mentioning here in regards to hybridization and that is prop/rotor/motor optimization. For hover, a large disc area leads to a lower disc loading and better efficiency, along with a low rotor pitch angle. However, for forward flight the prop should have the pitch optimized for the desired forward cruise speed (in combination with motor efficiency/RPM/kV etc) and the rotor disc (prop diameter) should only produce enough thrust to overcome airframe drag at that cruise speed. So if you have a low drag airframe in forward flight at a high cruise for best range, this will mean it’s highly unlikely the same propeller configuration will work well in hover as well. These are two mostly contradictory design parameters that are not easily overcome in conventional symmetrical type designs, and are also a consideration for tilt rotors and tailsitters alike.

I decided that what I needed for the gyrocopter hybrid was both collective and disc tilt control. Which I do not have the coding skills to modify the code for.

The configuration that was/is successful is my winged Tri

New photo by Chris & Kristin Olson

It is a 1 meter wingspan and I get 50 minutes of flight time from a 10A 4S battery with it. It’s limitation, without being able to automatically adjust the AoA of the wing is the flight speed, which is fixed at 10m/s. I did briefly try a tilt rotor concept with it that manually tilted the wing rotors ahead by 30 degrees and still flew it with the stock TriCopter code. But it’s limitation there was not being able to create enough lift with the wing to achieve any higher range (in distance traveled on the same battery). I could get more flight speed at the expense of higher power requirement to fly it.

The other downside to the winged design is wind. Flying at a forward frame angle of 15 degrees the wingtip motors are throttled down so far that it has somewhat the same problem as the engine assisted Tri - not good enough stabilization so it is easily buffeted by turbulence.

I still fly the Tri on still days, but eventually went to a highly modified Align Trex 700 nitro helicopter that has a cruise speed of 22 m/s and flight time of one hour on 15% nitro fuel. The helicopter is quite expensive to fly compared to electric with the fuel costing $30 US/gal. But it has the range, speed and payload capacity that I could not get from the electric multi experiments.

Nice! Thanks for sharing.

I’ve followed your builds with interest, and I think we’ve touched on some of these subjects before, including you offering the differential tilt yaw control code to test. I’m assuming that you are using the same prop motor combinations on all three positions? Where roughly is the CoG on those aircraft? Is it close to or on the front wing? I think by having a larger rear rotor and smaller front motor/props you might be able to have both better hover and control performance.

From what I can tell the position of the wings need to be further back so both the CoG and CoL are close to eachother in both hover and forward flight. Have you done a rough weight/lift distribution calculation? The front motors could also be mounted on tilt canards which could double as control surfaces for forward flight, and the main wing could be mounted more centrally, closer to the main rear rotor center of lift. It might be worth a try, and is quite similar to what we’re experimenting with.

On the subject of fuel power density: I heard of one guy that was using a solar PV modules to convert light from a burning wick to produce electricity for electric propelled flight. (So fuel >light>PV>electric) At first it sounds a bit crazy, but given that the energy density is so high and general ICE motor’s in that size are very low efficiency, are loud and have excessive vibrations, the idea tends to have some merit provided you can get the required PV surface area and light intensity together on the aircraft. Just roughly for smaller 50W systems using methylated spirits you need about 8.5% overall conversion efficiency to have the same energy density as 18650 li-ion cells.

The CG is just slightly in back of the wing on both. The wing motors are Tiger 3510-630’s spinning 16" T props. The tail motor is a Turnigy Elite 3508-640 spinning a 15" T prop.

My thinking, for hover, was that I need more power on the wing to sort of balance the CG location properly, and that actually works very well in hover in the wind. The wing motors start to throttle down with as little as 5 m/s forward speed. And it reaches its peak efficiency at 10 m/s.

I’ve pondered several ideas like using control surfaces on the wing. Coming up with a scheme to vary the AoA of the wing. I played with the tilt rotor concept for a couple weeks. I never intended to fly it on the wing alone as I knew it required enough load on the wingtip motors to maintain control with the stock Tri code. But I still haven’t come up with what I’d like to do with it to get more forward flight speed and increase its range. After I went to the piston helicopter, which has a much higher performance level and is relatively impervious to wind, I left the Tri as it is and have just flown it on still days where it really works nice. The helicopter is considerably more efficient considering the fact it has a 6kg takeoff weight full of fuel, and will run on slighly less than 1 US oz of nitro fuel per minute and produce 3.6hp. The downside with the heli has been cost - a new piston, ring and sleeve, and crank bearings every 100 hours (cost ~$250), and $15/hr for fuel. But the helicopter can do anything from hover to 30 m/s, can easily carry a 4kg payload, and is not affected that much by wind so it is much more reliable. I would like to convert the heli to gas but haven’t found a suitable engine that can produce the power of the nitro engine in the same dimensions of weight and size.

@ChrisOlson, If you haven’t seen it already, you might find this aircraft design interesting.

It uses a balanced, hinged airfoil to increase efficiency in forward flight without decreasing efficiency in hover. There’s a really good interview with the designer here,

Bruce, one of the regular guys on the show asks very pointed questions about his doubts that the design is worthwhile. This results in fairly detailed explanation of how it works.

I’ve seen a lot of these concepts. But I think the end result is that when it comes to the efficiency of converting stored energy, of whatever density, to thrust, a single rotor is always going to be the best for a VTOL platform. The more motors and props you put on it, the higher the prop tip losses, ESC and motor losses, etc… It’s why the helicopter performs so well. The single large disc provides the most solidity and lift in hover vs several smaller props at the same power consumption. And in forward flight once it goes into ETL it becomes incredibly efficient even carrying a quite large payload.

The hover power tells this. My winged Tri hovers on 345 watts at 2.9 kg takeoff weight. 118.9 watts/kg. My 600 electric helicopter hovers on 485 watts at 4.6 kg takeoff weight. 105.4 watts/kg. And the size of the aircraft, blade tip to blade tip, is about the same.

In forward flight the helicopter is down to 365 watts at the 4.6kg takeoff weight. The Tri is down to 180 watts at its 2.9kg takeoff weight. But the Tri is only going 10m/s. The helicopter is going 22m/s. That’s comparing electric to electric. My piston 700 is more power efficient than the electric, although the engine is not as energy efficient as electric motors.

I don’t disagree that helis are a good platform if the use is primarily for extended periods of hover, or slow flight with a large load. However for long range forward flight at high cruise and only short VTOL times I don’t think you can’t beat the performance of a quadplane atm. Fixed wings are simply far more efficient than rotating ones. As for wind my experience so far with QP, including flying in 20m/s wind has been quite impressive if using weathervaning and forward assist, the wings really do not detract from it’s hover performance, in fact it positively improves them if configured properly and with the correct AoA.

The NASA platforms cddnflyr linked to are interesting to watch, and well worth it. Tip loses are only secondary to disc load in terms of efficiency in hover, but as I tried to demonstrate in my previous posts, having a large spinning rotor for forward flight is a liability if one can use small, cruise speed only optimized, high aspect wings in forward flight instead. In fast forward flight the main hover lift rotor should completely stop and feather out of the airflow and then only power the minimum of motors required to overcome aircraft drag in cruise.

This is currently the fastest heli:

Note the small stubby wings and forward props, along with a main rotor that slows down.

Helicopters are actually at their worst in hover. In forward flight is where they get more efficient.

I do not disagree that fixed wing is more efficient than rotory. That is obvious. What I needed was a platform that can fit in the back of a shortbed pickup to be hauled to the job site, can lift up to 10kg payload and still be able to fly with it, has VTOL capability, can cruise at 22 m/s for one hour, handles wind speeds of 10m/s in flight while being able to maintain 22m/s ground speed, and autorotation capability, or similar, in event of power loss in flight so the aircraft is not lost. And can be flown with the stock code releases since I am not capable of modifying the code for a more exotic platform.

The helicopter is the only proven and reliable platform I have found, so far, that can do all that. And it does it very, very well. 200 hours of flight time last season and only one autorotation and emergency landing due to a failed clutch in flight. And I had no damage to the aircraft or payload other than a hole in the canopy where a piece of the failed clutch blew thru it when it exploded.

Can’t think of another place where this remark might be appropriate,but, does anyone know if there is experimentation with the idea of boosting range and time via a hybrid motor/electric combination that generates electricity. Kind of like a diesel locomotive or, like a Prius Hybrid.
It seems to work in theory since petrol had more energy density than batteries, and you would, in effect, be carrying POTENTIAL electric power in the onboard fuel. The petrol engine could either be idling, waiting for battery demand below a threshold, at which time the engine would throttle up to generate the needed recharge, or, it might be possible for the engine to be started electronically when demand is needed.
This would probably need a different kind of engine than the stock hobbyist engine, but I don’t doubt there is something like this out there.

There actually is something like that out there with a purpose-built onboard genset. But can’t remember where I saw it right now.

OK, this is where I saw it - a little liquid-cooled single spinning a brushless motor as a generator:

Two other mobs who I came across who are dabbling in the hybrid concept:

I’m working on a similar project, but with coaxial central rotors. What did you use to program the motor patterns to get it to fly properly? Was it based on any existing design, like a Y3 with just a central thrust motor?

Good morning
Can you please help me get an answer to the question
Regarding connecting the ICE servo throttle to the pixhawk

Servers by jDrones