JSBSim + ArduPlane SITL: No Thrust from Right Engine on Custom Blended Wing Body

[rspandolfi]

At Kashmir World Foundation, we design aircraft to support the protection of endangered species. Nearly all of our aircraft are of a blended wing body design, optimized for efficiency and stability, and operating on ArduPlane. Currently, we’re developing a new aircraft intended for patrol missions over the Himalayas, where it will face high and variable winds. This design is being optimized not only for efficiency but also for agility.

As part of our development process, we are stepping through a detailed analysis of efficiency, stability, and agility using ArduPilot SITL. Unfortunately, we’ve run into a snag.

We’re working on integrating our latest Blended Wing Body aircraft into ArduPlane SITL using JSBSim as the Flight Dynamics Model. The aircraft has two pusher engines mounted symmetrically on either side of the centerline. Currently, only the left engine is producing thrust during simulation.

The FDM was created in JSBSim using aerodynamic parameters derived from OpenVSP. Our ArduPlane .parm file is based on the default JSBSim SITL configuration. Engines are configured on Servo 3 (Left Engine) and Servo 5 (Right Engine)—please confirm if this assignment is correct.

We compiled ArduPlane SITL from the latest source. We’ve tried disabling and reversing the outputs on Servo 3 and Servo 5 to verify thrust behavior. The Left Engine responds as expected. However, the Right Engine does not produce any thrust regardless of configuration.

Questions:

Has anyone encountered a similar problem when integrating JSBSim with ArduPlane in a dual-engine configuration?
How does ArduPlane recognize engines from a JSBSim FDM XML file?
What are the required steps—starting from defining the FDM and ending in simulation—for ArduPlane to properly simulate two engines?
We’ve attached both the .parm parameter file and the JSBSim FDM in .txt format for reference.

Any advice or experience shared would be greatly appreciated.

Thanks, Ron / Team Pegasus / Kashmir Robotics / Kashmir World Foundation
ER_W2_FDM.txt
ER_W2.txt

[bcoconni]

You did not join the engine/propellers definition files so we cannot run your FDM ourselves. However, while reviewing the file ER_W2_FDM.txt I saw that you had to define a dummy tank. I guess you are using electric engines so you should be able to remove the lines <feed> 0 </feed> from both of your engines and as a result be able to remove the tank #0.

Engines are configured on Servo 3 (Left Engine) and Servo 5 (Right Engine)—please confirm if this assignment is correct.

I guess you are talking about ArduPlane features ? For questions related to ArduPilot/ArduPlane, we suggest that you ask them to the ArduPlane team. Unless @Jaeyoung-Lim is able to answer your questions ?

[rspandolfi]

Hi Bcoconni .. Thanks for the quick response .. Someone from our Team will post the XML file so you can run the FDM. If it runs with both engines started, it should run the same within ArduPilot. I checked my notes from a prior project, and we had trouble getting both engines started then as well. Unfortunately I failed to make a note of the fix!

[seanmcleod70]

@rspandolfi can you confirm whether you can start both engines and get thrust from both if you run your FDM stand-alone with JSBSim without any Ardupilot SITL integration?

[rspandolfi]

I can verify that we can start both engines and get thrust if we run JSBSim via a Python Script. In order to attach the Python, I had to change the format to .txt.
EngineTest.txt
We are also going to verify that we can start both engines through Flight Gear.

[seanmcleod70]

Then it's most likely an issue with the ArduPilot SITL setup. Maybe even though you've connected 2 servos it thinks there is only 1 engine?

[rspandolfi]

It is nice when your own team solves the issue. We assigned two of our best computer scientists to our aerospace team, and collectively they identified the issue and derived a solution. Attached is a very general solution that everyone can use:
JSBSIM_Two_Engine_Solution.pdf

[seanmcleod70]

Thanks for the feedback on how you resolved the issue with Ardupilot SITL.

My suggestion would be that the changes in /ardupilot/libraries/SITL/SIM_JSBSim.cpp for starting the engines and for sending through throttle commands shouldn't be hard-coded. There should be an input parameter specifying the number of engines and then the code can dynamically build both the string for starting the engines and for sending through throttle commands. So someone with an aircraft with N engines doesn't have to edit C++ source code and rebuild the source code.

[rspandolfi]

Thanks for the suggestion. Sammy created an adequate solution which we will post within the next few days. https://www.linkedin.com/posts/kwf_intern-star-of-the-week-sammy-gray-activity-7349482621468307457-kPZT?utm_source=share&utm_medium=member_desktop&rcm=ACoAADtxaHgB25zRtF83U_lfOT1SMmmxvxGxij4 Ronald Pandolfi, Ph.D. Director, TACP-Network CTO, Kashmir-Robotics, a division of Kashmir World Foundation www.KashmirWorldFoundation.org <http://www.kashmirworldfoundation.org/> "*Extinguish your spot fires **before the winds of chaos **ignite a wildfire in your path."* — *from the reflections of Princess Aliyah Pandolfi*

…

On Mon, Jul 7, 2025 at 4:18 PM Sean McLeod ***@***.***> wrote: Thanks for the feedback on how you resolved the issue with Ardupilot SITL. My suggestion would be that the changes in /ardupilot/libraries/SITL/SIM_JSBSim.cpp for starting the engines and for sending through throttle commands shouldn't be hard-coded. There should be an input parameter specifying the number of engines and then the code can dynamically build both the string for starting the engines and for sending through throttle commands. So someone with an aircraft with N engines doesn't have to edit C++ source code and rebuild the source code. —