Servers by jDrones

Tiltrotor support for plane


(Chamika Gamage) #222

Frame type is doesn’t matter if it is Tricopter with Q_FRAME_CLASS:7,
I’m using Q_FRAME_CLASS=7 and Frame type 1 , (I has been set from previous test with X copter)


(Antonio) #223

@Chamika Thanks, Yes all work fine!


(Rolf) #224

After the repair of our VTOL, we finally had good weather today and tried the first complete flight:
Back and forward transition worked perfectly. Thanks to Tridge and the other developers for this real masterpiece of aviation software.

https://vimeo.com/207823370

We need to improve the PIDs for yaw at hoovering. With everything else we are fully satisfied

https://vimeo.com/207850525

In the next few weeks we want to provide better films with corrected settings of the yaw-parameters.

Rolf


(tridge) #225

fantastic, I’m absolutely delighted!
I’d love to see the dataflash log if it is available. For the yaw issue, while it may be yaw tuning, it could also be a compass issue. The logs should tell us.
Yipee!


(Rolf) #226

Thank you in advance for the look. Here is a link to the log (96.bin) :
https://www.magentacloud.de/share/40-1.z-i2i


(tridge) #227

@Rolf, thanks for the log.
I had a quick look, and the most noticeable thing is the pitch trim is quite a long way off in fixed wing flight. It needed a elevator PWM of around 1380 to fly level. It took quite a while for the elevator to learn that level of trim. That means either the plane is quite nose heavy in fixed wing flight or you need to adjust your elevator to make it mechanically up quite a lot. I’d initially change SERVO2_TRIM to 1380 and then look at how far off the elevator is. Note the angle of the elevator and then adjust the elevator horn until with a SERVO2_TRIM of 1500 it is in around that same position. Alternatively move some weight further back in the plane. Moving weight could be a problem for hover though, as the plane is actually a bit tail heavy in hover. Getting the balance right in both fixed wing and hover can be quite tricky!
I’d also suggest raising PTCH2SRV_P from 0.6 up to 1.0, and raising PTCH2SRV_I from 0.15 up to 0.3. That should give it much better pitch control in fixed wing flight.
In hover, your motors are wasting a fair bit of power rapidly changing the roll output. I’d suggest halving the Q_A_RAT_RLL_D from 0.02 to 0.01 to reduce the oscillation, and reducing both Q_A_RAT_RLL_P and Q_A_RAT_RLL_I from 1.0 to 0.8. I’d also suggest lowering the gyro filter INS_GYRO_FILTER from 20 to 10.
Next, I’d suggest lowering the Q angular accelerations. Lower Q_A_ACCEL_P_MAX from 110000 to 55000 and Q_A_ACCEL_R_MAX the same.
Regarding yaw, you do have quite large magnetic field variation. The minimum magnetic field strength from the log was 215, and the maximum was 463. The rule of thumb is that the compass total field strength shouldn’t vary by more than 10% in the flight. What that means is that either the compass is badly calibrated or there is significant motor interference. I suspect the primary problem is the compass offsets are a fair way off. You currently have:
COMPASS_OFS_X 134
COMPASS_OFS_Y -60
COMPASS_OFS_Z -395
I’d suggest changing to:
COMPASS_OFS_X 96
COMPASS_OFS_Y -150
COMPASS_OFS_Z 165
that comes from a simple sphere fit to your flight log.
Fixing those compass offsets should help the yaw somewhat. Once that is done try some hovering again and a new log will help with yaw tuning (which is probably also needed, but is hard to judge with the compass errors).
Cheers, Tridge


(Rolf) #228

I’m absolutely amazed at how fast you are able to analyze such logs. Thanks a lot.
We will correct the pitch servo mechanically, recalibrate the compass outside on the field and change the other parameters as you have written.

best regards Rolf


(tridge) #229

I’ve been working more with tricopter tiltrotors today in the simulator and have a number of improvements pending:


these significantly improve transitions in the simulator


(tridge) #230

these changes are now in master after simulation testing


(Greg Covey) #231

Congratulations Rolf and Tridge on continued success with tiltrotor code. I miss the sunny days and green fields as seen in Rolf’s video. Today is a good day for snow planes!


(Rolf) #232

I like to give a few degrees of temperature over the Atlantic and in turn less rain here :wink: . The green fields are result of the many rain and sunny days are unfortunately only rarely - usually when we have no time to fly. The weather forecast for this weekend is windy (up to 40 kmh / 25 mph / 22 knots ) and rainy.

So we use the time to improve the VTOL as recommended by tridge and waiting for better flightconditions next week. We will update to master with improved tiltrotor transition strategy.

Rolf


(Rolf) #233

Finally the weather was good this morning and we had perfect flight transitions. Thanks Tridge.
The battery is a bit weak, so we had no optimal copter characteristics.
But the transitions and plane characteristics are perfect:

Logfile: https://www.magentacloud.de/share/aiubftfo9h#$/

Best Regards Rolf


(tridge) #234

I’m glad you had a good flight!
That log file isn’t the right one I think. It is only 28k zipped. It should be several MByte long. Can you check if you have the full log file?


(tridge) #235

hmm, an odd thing, I notice in the video the left motor is coming up a little bit occasionally in the forward flight. I’d like to understand why that is (if it is a software issue and not a hardware one).
In forward flight the two motors should be fully down, and if they do come up, then they should come up together. See for example 1:16 in the video.
Having the video of the motors is great!
The transition from fixed wing -> multicomputer looked good in the video. The forward transition from MC to FW wasn’t perfect though, I can see a bit of pitch deviation. I’d love to see the full log.


(PG) #236

@tridge a while back someone asked about independent tilting action on the tilt rotors to control yaw in Hover mode on a Tricopter platform, and you had mentioned it could be possible to add but it wasn’t currently supported. Any updates on this? I’ve been closely following this thread waiting for that to arrive…


(tridge) #237

I’m planning on adding it quite soon, as I got an EFite Convergence last week, and plan to get it flying soon. It needs that feature. Just too many things going on right now. I have 4 planes I’m working on building at the moment.


(tridge) #238

what plane are you thinking of trying it on? If I add the feature before I build my convergence would you be willing to test it for me? I would test it in a simulator first of course.


(Rolf) #239

Ouops, the file was probably damaged during the upload.
Unfortunately i can not upload the full file before thursday. I will do.
The left motor is coming up a little bit due to the hardeware.


(Jaspreet Dhanjal) #240

hi
im recently working on a vtol tilt rotor aircraft just like the WINGCOPTER
im really new to this gifhub repository, how it works on cloning the codes and make a px4 file (couldn’t understand a bit) so i need help from you guys in running the codes for my model
its a quad rotor tilt mechanism


now the great issue is that im using a pixhawk mini which has only 8 pwn out

so first four can be used for the quad setup and rest four can be used for (aileron,elevator,rudder,Tilt Servo) is it possible? or an AUX RC pass through can be used (i don’t know)
can any thing be done to make it fly… ?as a vtol

reply awaited
thank you


(tridge) #241

no problem, just like to look at it before your next flight

ok, weak servo? What is pulling it up? The thrust should be pulling it into place