New full scale unmanned winged tiltrotor VTOL

Hi, I am developing a new type of winged tilt rotor vtol, similar in functionality to the Convergence RC model. It is a flying car with 4 large wheels, that uses the body as a wing, So there are no wings to unfold and the vehicle flies just as it is. The vehicle is capable of horizontal and vertical flight. In addition it is amphibious. The vehicle weighs a 150 pounds it is 15’ long 8’ wide and 4.5’ high. It uses 2 counter rotating tilting motors in the front, and 2 coaxial coaxial counter rotating motors in the back behind the cockpit, that blow the thrust through a hole in the body. Thrust tests show that the prop motors generate 64 pounds of thrust each.

Initial tests showed that the vehicle will not take off in VTOL mode. The front motors can easily lift the front of the vehicle, but the back coaxial motors can not lift the back end. But when we lifted the back end manually, the back end suddenly lifted. We think that the back motor thrust has increased the pressure under the vehicle liimitng the motors thrust. Perhaps the motors need to also be tilted to help direct the thrust out the back. If pressure build up is the reason, then we need to lower the wheels or cut the bottom of the side panels to allow the pressure under the vehicle to escape.

Has anyone seen this effect before? Do you think it is a pressure build up problem?

I have never used a coaxial setup before and do not trust it yet. I have a 1/6 scale model of the vehicle that flies very well, but it uses just one back motor placed above the hole in the back, and it is tilted a little sideways to counter the motor torque. Tests with the small model show that the back end lifts even if the hole is extended via a tube all the way to the ground. Extending the tube above the motor (simulating the lower motor in the coaxial setup), shows the thrust is reduced. My guess now is that the pressure problem is due to the bottom motor and not the top motor. So another solution could be to place both motors above the hole.

Here are some pictures of the full scale unmanned vehicle:






Picture of 1/6 model:
https://drive.google.com/file/d/1innPCGBoCvsGY7UdGcp_YOSvnSYgCdQt/view?usp=drivesdk

Here link to see the 1/6 vtol test flight:
https://drive.google.com/file/d/1jZfHPpZte_6V5OYBiKk4o7pi2t-CtVx6/view?usp=drivesdk

Here is a link to see the full scale vehicle test flight in Realflight 7.5:
https://drive.google.com/file/d/1AAvC3BLvut7xjlhY81S9EqD8OW0Qw-kS/view?usp=drivesdk

This is awesome! Wish you the best of luck…

We finally found the problem. The thrust from the motors developed a vacuum under the vehicles side panels, which prevented the vehicle from taking off vertically. We verified this by blocking the thrust using foam sheeting taped to the inside of the side panels. The vehicle now lifts without any problem and hovers at about 60% throttle. The 1st test flight had only 2 of the 4 batteries in the front of the side panels to make it easier on the motors and the vehicle. It hovered very well and was very stable:

https://drive.google.com/file/d/1FBNGn9vmKVE7DIyU-81XFSWxTqEReTWX/view?usp=drivesdk

The 2nd flight had all 4 batteries in the front of the side panels. It hovered well, but it is not as stable. The 1/6 scale model had the same problem, and increasing the Roll P to its max value of 0…8 fixed the problem. But since the full scale vehicle needs a lot more heavy batteries 60 lbs, then it sees to need even more Roll P. But we are at the max P value available in Mission Planner so we need to find a solution. Either we can get more P value, or perhaps there is a way to increase the motor differential control range. If this is not possible, then perhaps we can get more P range from the ESCs. Does anyone have any suggestions?

Hi Andrew Tridgell I am looking at your code for the Comnvergence and there a few things I don’t understand. I was wondering if you could help me understand some of the details, so I can apply it to my vehicle which should function in a similar way?

  • I don’t understand the difference between (QA RAT_RLL and RLL_RATE) PID controls?

  • What do Q_VXY PI controls do?

  • What does SCHED_LOOP_RATE do?

  • What does FF in a PID do?

Thanks

Q_A_RAT_RLL is for VTOL control whereas RLL_RATE is for fixed wing control

they are for horizontal velocity control in VTOL position modes like QLOITER

that controls the speed of the main update loop

it is a feed-forward, which means the output is proportional to the desired value (eg. desired roll rate) and doesn’t depend on actual measure value (eg. does not depend on actual rate measured by gyro)
A FF value is needed on a control axis that is drag limited, such as fixed wing roll and pitch. VTOL thrust driven axes normally do not have a FF set

Thank you very much!

Hi, I would like to know if there is a way to control the tilt of the front motors to move the vehicle forward during hover flight mode, instead of tilting the entire vehicle. The Convergence RC model can do do this and uses the elevator control stick to control forward movement. The reason I want to do this is that because I have a lot of surface area, tilting the entire vehicle to move forward significantly increases the drag of my vehicle, slowing it down, and making it difficult to penetrate winds. Tilting the motors instead will solve this problem.

Thanks, William

yes, two methods depending on whether you want to manually control the tilt in QHOVER or if you want to automatically control the tilt in velocity controlled modes like QLOITER.
For QHOVER you need to set an RCn_OPTION to 209 for “forward throttle”. Then when you move that channel it will tilt the motors.
For QLOITER (or other velocity controlled modes like AUTO, QLAND etc) you need to set Q_VFWD_GAIN to a non-zero value. Try 0.1 to start with.

Hi Andrew,

Thank you for the info. I tried both of these methods and can’t get it to work. I am using a Pixhawk cube orange flight controller, and a Taranis X9D Plus 2019 transmitter, and have assigned my right finger controller (RS) to Control Ch 12. On the transmitter channels monitor screen I can see movement on channel 12 as I actuate the right finger controller (RS). In Mission Planner I set RC12_Option = 209. I can see channel 12 move in Mission Planner radio calibration as I operate the right finger controller (RS). But I do not see the front motors tilt when I am in QLoiter flight mode (by the way according to Ardupilot it should also work in Q_Stabalize which I prefer [ref Ardu link below]). I have tested it with the the flight controller armed and the transmitter throttle up to 50%. By the way I am currently using RC9_Option=41 for my transmitter Disarm Switch and this works very well.

The only way I can get the front motors to tilt is to start in Q_Stabalize, arm the vehicle, throttle up to 50%, and switch to Stabalize. This immediately tilts both front motors to about 45 deg, then I reduce the throttle to 0% and I can then tilt the motors up to 0 to 45 deg as I like using the throttle. This method works but it can be very dangerous when the vehicle is near the ground since I have no throttle control and the altitude pressure sensor is not very accurate. I would really like to get the method you mentioned to work. Can you please advise?

Thank you, William

  • Arduplane Quadplane RCx_Option reference link:

https://ardupilot.org/plane/docs/quadplane-flying.html

  • Link to see video of tilt working using Stabalize and throttle control:

https://drive.google.com/file/d/1YX-dXCo1lJCPWodLEGGjMCI99toN7CMY/view?usp=drivesdk

logs?
I’d also suggest you get RealFlight 9 setup for testing, you could use something like the griffin model which has the same basic setup

Hi Andrew, I have gotten something to work in QLoiter mode, but it is not very well behaved. For it to tilt the motors in unison, the thrust must 50% or more, so the elevator control can tilt the motors in unison. But a little elevator control in one direction sends the motor to its limit, either 45 deg or 0 deg. One can try to stop it by reversing the elevator control, and in some spots it will stop, but it generally tends to drift to the limit. This is not what I want. I need more positive control like I get with Stabalize mode, but I don’t want the throttle stick to control the tilt, since I then lose my throttle control. Also I need Yaw control, which does not work in Stabalize mode. Below is link to a log file of me operating the front motor tilt in QLoiter: 181.bin

https://drive.google.com/file/d/1qRStztokmbi9bxF9_7TtsbjbgF37mtWO/view?usp=drivesdk

RCin:
-ch2=Transmitter Elevator stick
-ch4=Transmitter Rudder stick
RCou:
-ch3=Left front motor tilt
-ch4=Right front motor tilt

Picture of signals:
https://drive.google.com/file/d/1ZEiY-j0vwngjWOe_UbvHBliEPjDF5j2Y/view?usp=drivesdk


The QHover method using RC12_Option= 209 for forward throttle seems more promising, but I can not get it to work. Below is a link to a log file of me trying to get the front motors to tilt in QHover. I get yaw control but no motor tilt with me operating the Right finger slider control (RS) Ch 12, where Mission Planner shows Ch 12 is clearly being operated: 183.bin

https://drive.google.com/file/d/1dkZpQWrkHCqJIkIc7jk-9jZCQ7Ycp_kV/view?usp=drivesdk

RCin:

  • ch4=Transmitter Rudder stick
  • ch12=Transmitter right finger control (RS)
    RCou:
  • ch3=Left front motor tilt
  • ch4=Right front motor tilt

Picture of signals:
https://drive.google.com/file/d/1YzWqsWLWF77u_yZVcei9OvXaorjhxHW8/view?usp=drivesdk


Again what does sort of work is using Stabalize mode. But the throttle controls the front motor tilt, and I lose the ability to control the throttle (also elev control does not control it either), as well as asymmetric tilt yaw control. Below is a link to a log file of me tilting the motors in Stabalize mode, as well as attempting yaw control and throttle control using elev: 185.bin

https://drive.google.com/file/d/11eXzv2n6uMvhhxLTiaC24VEGSHTmfPZx/view?usp=drivesdk

RCin:

  • ch2=Transmitter Elevator stick
  • ch3=Transmitter Throttle stick
    RCou
    -ch3=Left front motor tilt
    -ch4=Right front motor tilt
    -ch5=Back motor signal
    -ch6=Right front motor signal
    -ch7=Left front motor signal

Picture of signals:
https://drive.google.com/file/d/1Z0Lsb6HBIgxB2aponmDy77wGkdehbIgI/view?usp=drivesdk

Thanks, William

P.S. What exactly should be testing in the Realflight simulator?

We were able to successfully tune the flight controller by pivoting the vehicle at is CG while in the ground, and giving it impulses to study the stabalization. It worked great. Yesterday we tested it indoors and it hovered great and was quite stable:

https://drive.google.com/file/d/1csnLLtug1sz9msU4BM_flKp9pHXlquIa/view?usp=drivesdk

We only used 90% charged batteries (112V) because we are not sure the new APD 28S 300A ESC can handle the voltage spikes from a fully charged battery (117V).
The ESC is spected to handle our 3m battery cable lengths. But we are being causous since the ESC that came with the 18kW cont/40kW back motor burst into flames when we tried it at 100% charge. The ESC manufacturer is currently viewing our ESC battery scope traces.

Interesting unmanned flying car concept. What kind of usage scenarios do you envision for the various modes of operation?

Hi Mark,

I am developing a new type of flying car. This is just an unmanned full scale prototype. The next prototype will be manned. It will drive like a car on the roads and fly both vertically like a drone and horizontally like a plane. In addition it is amphibious. So it can navigate on the water, as well as take off and land on the water. It uses the body as a wing, so no wings need to unfold to fly. Even though the wing is very short, it is long and very efficient. The oversized side panels act a wiglets and reflect the air back onto the wing, and the two front motors provide constant thrust over the wing giving it more lift. The vehicle has many applications: transportation, recreation, search and rescue, police, military, cargo drone, taxi, and flying car racing.

I have been working on this project for about 6.5 years. I am a physicist from San Diego California, but a year and a half ago I came to the middle of Sweden to develop the current full scale unmanned prototype with my friend and business partner Dag Stranneby. We hope to start a company to develop and produce this vehicle, which we call the Sky Chaser.

Can you tell me a little about yourself? Where are you located, and what are your interests in this hobbie? Again, I want to thank you for all your help. I really appreciate it.

William

I’m just a retired engineer/scientist with an expensive hobby :slight_smile:
I live near Denver, CO, and fly RC at the AAM club field: Location - Arvada Associated Modelers
My focus has been mainly on tailsitters and tiltrotors for the past several years.

This would be a fun amphibious VTOL at model scale. I don’t suppose you would need power to the wheels either, but they would definitely need to be steerable.

Hi Mark,

Nice to meet a kindrid soul. Here is a link to see more details of the 1/6 scale model. It was made from simple foam board and hot glue and flies really well as a airplane:

https://drive.google.com/drive/folders/1j68z1Qh_AZvRhpnr3g798ylBluGNtuC0

William

The model has free wheeling steerable front wheels with shocks made from surgical tubing. There is no power to the wheels and uses the front props to pull the vehicle on the ground and water

My father was from Dever and we have family still there. He was a great aerodynamicist and we use to design, build , and test all kinds of crazy rc models. One really cool model was a plane that took off and landed using the Magnus effect, where the wings of the aircraft pinwheeled about its center axis. Ken Willard wrote an article about it in RC modeler magazine in Sept 1977:

https://rclibrary.co.uk/title_details.asp?ID=2209

Hi All,
I am having trouble getting the vehicle to hover after changing firmware from Plane ver 4.0.9 to Beta 4.2. The vehicle seems table in roll and yaw, but seems to have problems stabilizing in pitch. I tried hovering it a few days ago and the video and log file of the test are below. I need to know if the oscillation in pitch observed is operator error or an oscillation in the pitch PID controller that needs to be fixed?. I should mention that I tested the PID controllers for pitch, roll, and yaw by supporting the vehicle at its CG and manually giving it an impulse. At these low thrust levels, all the PID controllers seemed to work well except with a minor drift due to motor noise. I am currently using a FFT notch filter which lowered the noise by a factor of 10. Unfortunately I changed that reference level during the testing and forgot to recalibrate the reference level before the test flight. So this may have caused a slight nose down attitude during the test hover.

Looking at the log files I have noted several things:

  1. The vehicle tilted nose down at lift off and I gave a series of up elevator pulses to correct the level of the vehicle to slow it down, since I have limited testing runway indoors. I can see the motors correctly responding to my commands and trying to correct the level of the vehicle. What concerns me is that the oscillation is growing as I give up elevator pulses.

  2. the front motors are getting too close to their 2000 max limit in the test. I only charged the batteries to 80% and these high numbers tell me I need to charge the batteries to 100% charge to give the motors more headroom, especially in the front.

  3. The motor signals show that the motors cut out briefly as the vehicle was at its max oscillation in pitch where then, and I reduced throttle and gave up elevator. The motor signals come back as I land the vehicle in a nose down attitude. I would like to know if the motor cutting out like this is normal?

Link to video of test:
https://drive.google.com/file/d/1MTyVETTbGtH7NB2D-LtlW4wA6e7s2Wyk/view?usp=sharing

Link to log file of test:
https://drive.google.com/file/d/1W6SLKKTcl_u3rqhiTGUTYHWhxm388UjZ/view?usp=sharing

Do you think that if the vehicle’s level is recalibrated correctly, and if the batteries are charged to 100% that the vehicle is ready for another test, or do you thing there is a large oscillation in the Pitch PID controller that needs to be fixed?

Thank you, William Walker