How would you design a cyclocopter?

[2jujube7]

Here's an update on the state of the cyclorotors. As mentioned in the video, I figured out that my thrust stand is giving me thrust values that are 70% higher than the real values. That's a real problem, but at least it helps explain why the cycloplane and cyclocopter were unable to fly.

 

[CampRobber]

It's out of balance and the rubber band isn't riding right but despite that it made 50g thrust. I think it might fly.

I will buy Jude 3 of these plus a current sensor and an arduino:

https://www.amazon.com/dp/B08KRV8VYP/?tag=lstir-20

...if he wants to collect full data on a rotor.

 

[2jujube7]

It's out of balance and the rubber band isn't riding right but despite that it made 50g thrust. I think it might fly.

That's great to hear! Designing any cyclorotor with over a 1:1 T/W ratio is quite the achievement.

I will buy Jude 3 of these plus a current sensor and an arduino:

https://www.amazon.com/dp/B08KRV8VYP/?tag=lstir-20

...if he wants to collect full data on a rotor.

Thank you for the offer, I appreciate it. I'm (fairly) sure that my thrust stand accurately works now though, so it's not necessary.

 

[IMAN]

Hello everyone ,can someone of you help me to have more informations about cyclogyro ? Because I'm a student and I'm intersted by this technology and I find some difficulties to have informations about it . I hope that my language is clear and that I find someone to help me .

 

[2jujube7]

Hello everyone ,can someone of you help me to have more informations about cyclogyro ? Because I'm a student and I'm intersted by this technology and I find some difficulties to have informations about it . I hope that my language is clear and that I find someone to help me .

Look for papers on cyclocopters on scholars.google.com. There's some good ones out there that are free to access.

 

[IMAN]

Look for papers on cyclocopters on scholars.google.com. There's some good ones out there that are free to access.

Ok , thank you so much

 

[CampRobber]

I just learned about these things:

https://en.wikipedia.org/wiki/Paddle_steamer#Feathering_paddle_wheel

Paddle wheels, of course, are just fundamentally bad... but they came so close to a Voith-Scheinder!

 

[2jujube7]

Not sure if you guys saw my last video but basically

* I burned out an ESC so I'm switching to a tri-rotor. I don't think that anyone has done a tri-rotor cyclocopter yet.
* My flight controller just stopped working (not burned out or anything, it just stopped), so now I'm going to get a Teensy Arduino and figure out dRehmFlight. We'll see how it goes.

 

[FlyerInStyle]

could you use dRehmFlight with an esp32? it is comparable to a teensy

 

[2jujube7]

could you use dRehmFlight with an esp32? it is comparable to a teensy

I probably should've thought of that before buying a $27 teensy lol.

 

[NickRehm]

ESP32 I think will work but needs some pin remapping

 

[FlyerInStyle]

ESP32 I think will work but needs some pin remapping

I might give this a try with it then.

 

[2jujube7]

I finished soldering/crimping up the teensy yesterday. I decided to run all power through the power distribution board instead of the teensy to try to lower the risk of frying it. The light turned on on the MPU6050, so I think everything works.

All I have left to do is hook up the motors/servos, which should be pretty easy considering that the motors only need signal and ground, and there's one servo that needs the 5v power.

We'll see how programming it goes; Nick did a good job documenting it so I think it'll be doable.

Also I think I just soldered the input voltage into D+ and D- on accident. Would it fry the board?

 

[NickRehm]

Those pads are not for power, you'll want 5v to hook into the Vin pin (marked 5v on top of the board next to the micro usb port) and ground right next to it

 

[2jujube7]

I resoldered the power cables, and I'm pleasantly surprised to find out that I did not fry it.

 

[2jujube7]

I started working on the cyclocopter again. I started an engineering internship this summer, so I've been very drained and not wanting to work on it. However, I'm meeting with a professor in about a week so I'm hoping to make some good progress before then.

I had to rebuild one of the rotors due to some slop in the transmission, and I started testing it to make sure that it generates the same amount of thrust as the others. Got it up to 160g (halfish throttle) before I had to stop and and I'm going to have to reprint the driver pulley because it was skipping teeth.

All in all, I'm pretty sure that the rotor will work, and my thrust stand is giving accurate values so it is going to be able to fly. Now, controlling it is another thing. I kinda feel like it's too much work to source and put it on a gimbal for flight controller PID rate tuning, so I'm just going to try the good old fashioned method of holding it up in the middle of the air by a string lol. We'll see how it goes.

Plus, I got dRehm flight working! It's an excellent program, so hats off to @NickRehm for open-sourcing it. It's very easy to understand, modify, and use, so I'd give it 10 stars.

 

[NickRehm]

Glad that you're still pushing forward with this project, I was missing the updates. Congrats on the internship!!

The default gains in angle mode will probably work fine, though you may want to reduce the roll and pitch P gains by about 30-40% as a starting place. The yaw P gain will probably have to go up a little bit from the default. Good luck with flight testing!

 

[2jujube7]

I'm having a little trouble with the PID tuning...

I got the oscillations out, but it's just trying to hover at a high angle of attack and won't level out the pitch.

 

[NickRehm]

The oscillations are from coupling of a few of the axes. For example, giving a pitch output with the cyclos implies the rear cyclos spin up, front spins down. Because of the axis of rotation of the cyclos, this also induces a rolling moment from the single front cyclo. So then you get differential left/right cyclo output as a response to counter the roll disturbance. Rinse, repeat, and you end up in this cyclic oscillation. This can also result simply from your thrust vectors not being perfectly vertical. For example in roll, if your thrust vectors are actually slightly backward, a differential roll output will cause an unwanted yaw. The fix to all of this is less agressive P gain, and more D gain to dampen the self-oscillations out.

As for the pitch angle... I think your rear 2 cyclos are creating a pitch up torque. Then the front cyclo is lifting the nose up as well. This will lead to a natural tendency to pitch up that the controller cannot compensate for fully. Basically, the torque from the rear cyclos is already doing ~half of the work lifting the nose as you'd expect the single front cyclo to be doing. So a few things to try...one or a combination of these might fix the problem:

1. Shift your battery toward the nose to weigh down the nose and put more load on the front cyclo
2. Multiple thro_des by ~0.5 in the control mixer so that it does not spool up as much as the rear 2 cyclos
3. Double the i_limit parameter in the code. This will allow the integrator term in the pitch controller to spool up higher and maybe compensate for the extreme nose up tendency from the rear cyclo torque. Beware that the hover trim throttle value for that nose rotor may very well be 0.. or negative, meaning that to counter the rear cyclos' torque, the nose rotor actually has to thrust up, making the rear cyclos have to do all of the heavy lifting and then some more..
4. Reverse the direction of the rear cyclos so that they are torquing the nose down, not up. This will require more thrust from the nose rotor to prevent a nose-down tendency, which could be compensated by placing the nose cyclo farther forward for more leverage. This will be more similar to my dual cyclo configuration

 

[2jujube7]

The oscillations are from coupling of a few of the axes. For example, giving a pitch output with the cyclos implies the rear cyclos spin up, front spins down. Because of the axis of rotation of the cyclos, this also induces a rolling moment from the single front cyclo. So then you get differential left/right cyclo output as a response to counter the roll disturbance. Rinse, repeat, and you end up in this cyclic oscillation. This can also result simply from your thrust vectors not being perfectly vertical. For example in roll, if your thrust vectors are actually slightly backward, a differential roll output will cause an unwanted yaw. The fix to all of this is less agressive P gain, and more D gain to dampen the self-oscillations out.

As for the pitch angle... I think your rear 2 cyclos are creating a pitch up torque. Then the front cyclo is lifting the nose up as well. This will lead to a natural tendency to pitch up that the controller cannot compensate for fully. Basically, the torque from the rear cyclos is already doing ~half of the work lifting the nose as you'd expect the single front cyclo to be doing. So a few things to try...one or a combination of these might fix the problem:

1. Shift your battery toward the nose to weigh down the nose and put more load on the front cyclo
2. Multiple thro_des by ~0.5 in the control mixer so that it does not spool up as much as the rear 2 cyclos
3. Double the i_limit parameter in the code. This will allow the integrator term in the pitch controller to spool up higher and maybe compensate for the extreme nose up tendency from the rear cyclo torque. Beware that the hover trim throttle value for that nose rotor may very well be 0.. or negative, meaning that to counter the rear cyclos' torque, the nose rotor actually has to thrust up, making the rear cyclos have to do all of the heavy lifting and then some more..
4. Reverse the direction of the rear cyclos so that they are torquing the nose down, not up. This will require more thrust from the nose rotor to prevent a nose-down tendency, which could be compensated by placing the nose cyclo farther forward for more leverage. This will be more similar to my dual cyclo configuration

Thanks for all the great help Nick! I think I discovered why no one has tried this cyclocopter configuration - many different control couplings.
Multiplying the thro_des by 0.5 worked perfectly, and it is now hovering level. The unbalanced torques from the two main rotors are probably yet another reason that people haven't tried this configuration.

The only problem that I'm having now is pitch (just pitch, I got rid of the roll couplings somehow) oscillations. I'm a little stumped by this as I've had to reduce the P value by so much that manually moving the controls seems to have no effect? I'm sure that I can figure this out eventually though as now I'm really just down to the pitch Kp and Kd.

Again, thanks for all the advice.