Servers by jDrones

Gas-Powered Hybrid-Electric DJI S1000

Hey Everyone,

Just wanted to share some of the work we have been doing with with LaunchPoint Technologies on their hybrid-electric DJI S1000. They developed the genset using their advanced Dual Halbach Array generator and a custom-designed rectifier/controller. This controller also allows the generator to start the engine remotely via a switch on the transmitter. The genset maintains a nominal 6S voltage and provides enough power for hover while the downsized battery absorbs the power spikes of climb and sporty maneuvers. The battery also allows the vehicle to fly quietly on electric power only for several minutes.

We have been supporting their effort by converting the S1000 to fly with on a Pixhawk controller (AC 3.4), setting up the control channels, tuning, etc. In addition, we are doing some custom Arducopter coding to communicate with the genset over CAN bus. The genset controller sends genset status (cell voltages, temperatures, engine RPM, fuel quantity, etc.) over CAN bus which can be read by the Pixhawk and sent over telemetry to the GCS. Starting and stopping the engine in flight will also be over CAN bus (currently using PWM RC passthru). This community has been very helpful in this endeavor.

Why gas? On a per mass basis, gasoline contains about 240 times more energy than a LiPo battery. Accounting for the energy conversion losses (more losses with the genset than all-electric), gasoline contains about 12 times more usable energy than batteries. And finally, accounting for extra mass of the genset components, this results in about 3-5 times longer flight times. For the S1000, that means we can expect flight times of 45-75 minutes. As we further optimize the system and reduce mass, we will likely see over 2 hours.

Here is a video of the maiden flight from last week. The integration of the system onto the S1000 hasn’t been perfected yet, we are just working on the proof of concept at the moment. We flew without a payload, but there is weight margin to carry one.

This work is being funded under a NASA SBIR contract.


Really great stuff, good work, well done!

This is awesome!

If you’re generating ~25.2v presumably the generator will keep the battery charged up so when you run out of fuel you can just top up the tank and never worry about the battery? Maybe some kind of passive cell balancing on board would be needed for that long term though…

Again, awesome.


Not really anything new.
It is years that hybrid solutions exist for UAV planes.
The problem is to make the gasoline section the lightest or this solution will work only on large scale scenario but since a large scale multirotor has several limitations the final answer IMO is that this solution has no future since on large scale a gasoline helicoper will always win.
The only possibility would be to be able to create an high efficiency generator adopting a micro turbine but I guess that a human mission to Mars is something easier.

BTW two words on the actual flight time of the multirotor I build and sell, quad, 47 minutes with 250 grams payload, hexa , 30 minutes with 1 kg payload, good but standard batteries and props.

Just want to say, many people do not like drone, because of the noise. And they definitely will not like this one.

I wish someone can install a big enough fuel cell battery on the drone, not gas engine…

Hi Justin,

Thanks for the comments! You are correct, one benefit is that you never really need to remove the battery to charge it. The generator along with the battery management system (BMS) will maintain the nominal voltage, which in this case is around 22.2V. It is much simpler, and more importantly safer, to target this voltage than trying to make sure the battery pack is fully charged. The BMS also maintains cell balance and accommodates up to 12S at this time.

You are correct, it is hard to beat a gas-powered helicopter in terms of efficiency/flight time. There are benefits that multirotors provide, however. One of these is the redundancy that you can get with anything more than 4 rotors; if one rotor/motor/ESC fails you can continue to fly or limp back home, if the flight controller and remaining rotors can handle it. When you’re carrying a payload that potentially costs $20k+, redundancy of flight-critical components is vital. Multirotors also tend to be easier to control, at least in my experience, and they don’t rely on collective/cyclic control mechanisms. Replacement parts also less expensive for multirotors, albeit there are more of them.

I share your concern on noise. I will say that this application is not meant for your average hobbyist who flies in the neighborhood or the local park. This is being developed for commercial use, one specific application being agriculture. The farmer can map their entire crop in one go without stopping to recharge several times. Pipeline and power line inspection is another application. If you do need to fly over a noise-sensitive area momentarily, the engine can be shut off in flight and then restarted once clear of the area. The muffler we’re using is an off-the-shelf unit not optimized for noise. There are other companies out there developing mufflers for engines of this size with active valve control that balances noise and power.

Fuel cell-powered drones will be interesting and certainly much quieter! Currently, fuel cells of this size are far too heavy and hydrogen is not easily accessible like gasoline. It will be a decade or more before fuel cells become a viable option for this application.

That’s an interesting concept. But I already do agricultural mapping and NVDI imagery with a piston 700 helicopter flown by ArduPilot. And skip all the losses in generator, wiring, ESC’s, multiple motors and prop tip losses. Have full autorotation capability in the event of power loss. And can carry 22lb payload, fly at 70mph and have 1 hour flight time with 5 minutes reserve fuel. If I want two hour flight time just put saddle tanks on it and add another gallon (6lbs) of fuel.

Multi-rotors are highly inefficient to begin with - they must assume a fairly radical flight attitude to gain any significant flight speed. So their power requirement goes way up in forward flight. A helicopter gets more efficient and takes less power in forward flight due to Effective Translational Lift. While these research projects funded by taxpayer money are interesting, I fail to see the practicality when higher-performance, more efficient and reliable platforms already exist. Especially when you are throwing a system that has high losses (combustion generator) on top of a system that already has high losses (multi-rotor aircraft). Why not just use the first part of the system (the combustion engine) to directly drive the propulsion and lift system and skip all the loss stages?


Hi Mike,
I’m interested in developing something similar to this work and would like to talk to you about it if you have the time.

Servers by jDrones