Replies to This Discussion

Permalink Reply by marc.kulhavy on January 17, 2012 at 10:19am

Hi to all, today during some video flight, I hat this yaw-lock multiple times. I could recover the ever-yawing/not responding copter from this by applying full throttle up and down.

Ok, so here ist the video:

http://www.youtube.com/watch?v=ZsqXQYG25tY

After backwards flying, I wanted to turn around but the copter didn't stop yawing. You can hear when I apply full throttle to get control back and the copter "releases" yaw and can be controlled again.

I have uploaded the complete logfiles from this flight and the current settings from the Mission Planner, too. What I can see in the *.kmz file is that the orientation of the copter is not right. I don't know if this is normal behavior for copter-firmware, I know with arduplane the orientation was right.

I hope someone can help!

Regards

Marc

Attachments:

• 2012-01-17 05-56 10.zip, 227 KB
• Ardu3DR_current.param, 2 KB

• ▶ Reply

Permalink Reply by marc.kulhavy on January 18, 2012 at 1:24am

Good morning,

this morning I attached a servo with a long rod on the servo arm through an y-cable to the receiver output so I could see if the inputs from my transmitter are passed correctly to the APM.

After a few seconds of flight yaw did not respond, I could give full yaw left and right input and one could clearly see the servo going from full left to right without any reaction from the copter. I held down the yaw stick for a few seconds to full left and suddenly the copter reacted again correctly and yawed full left.

I took a look into the mavlog data and one thing seems to be very suspect:

The value for yaw in the ATT message varies from about 32500 in the beginning to somewhat 2900 to the end of flight, peaking back to 32500 over the entire flight.

Can anybody help please?

Marc

• ▶ Reply

DeveloperPermalink Reply by Marco Robustini on January 18, 2012 at 2:16am

Have you the same yaw issue In Acro mode?
Have calibrated the throttle limit of the esc?
Your magnetometer responds correctly in the planner?
Have tried to make some corrections on the yaw pid?

• ▶ Reply

Permalink Reply by marc.kulhavy on January 18, 2012 at 5:17am

Hey Marco,

The issue persists in Acro mode.

The ESC's are all calibrated.

The magnetometer responds correctly in tne Mission Planner. It shows the correct direction when connected and reacts correctly when I turn the copter around by hand.

The PID's of yaw are untouched, as I have a stock 3DR frame with 880kv motors. Anyway, the yaw should respond SOMEHOW if I apply FULL yaw left/right, but it seems to be stuck for some seconds or until the next bigger throttle change.

Regards

Marc

• ▶ Reply

DeveloperPermalink Reply by Marco Robustini on January 18, 2012 at 6:01am

I know, since I start with ArduCopter I immediately noticed mismanagement of the yaw compared to other systems not necessarily more expensive, just for example the little KK (25$) have a nice yaw control, unfortunately there is still much to be placed in this direction, it all depends on a non-optimal matrix table of motors and it is something sadly lacking in AC.
Anyway i think you have a bad yaw or throttle pid, or some bad solder around... i hope you do not have bullet connector between esc and motors.
If the problem persists even in Acro the magnetometer has nothing to do.

• ▶ Reply

Permalink Reply by marc.kulhavy on January 18, 2012 at 6:34am

Thanks for reply, Marco.

As I have relative good soldering skills and already double-checked EVERY soldering joint on the board after this happened, i can be pretty shure that there are no issues with this.

It seems to me that the yaw-routine (if there is one) lags or hangs, while the rest of the control loops keep on working. I will try to change the PID-values of yaw and throttle and see if the behaviour changes.

But I still would like to know if the very high value for yaw in the ATT message of the log files is correct or not. Could someone compare this with his own logs?

Regards

Marc

• ▶ Reply

DeveloperPermalink Reply by Marco Robustini on January 18, 2012 at 8:20am

Marc, i created a short video with AeroSim that shows very well sometimes anomalous behavior of the yaw with throttle and altitude climbing.
I'm in Acro mode, without stabilization, magnetometer is disable.
I hope that in future it will be enabled better control of the yaw, perhaps with the potential for the planner to manage the matrix table of the engines due correct the yaw control with different configuration.

• ▶ Reply

Permalink Reply by marc.kulhavy on February 17, 2012 at 4:16pm

Marco,

it seems like I have a defect board. Last weekend I nearly lost my copter because it was not responding on CH5 and did not RTL. The yaw-issue is still there, I just repeated the beaviour with the HIL-firmware and X-Plane. Yaw does not work while all other commands (roll, pitch, throttle) behave normal. I even attached a servo between tx and apm to filter out the possibility of a defective receiver.

For some reasons, yaw is not recognized sometimes, CH5 commands too, they sometimes dial in after a couple of seconds, sometims never. Something seems to be wrong with the pwm-encoder onboard. Solder joints are all perfect soldered, even carefully resoldered, cables and tx/rx changed a couple of times.

I bought my kit on December 2011 by a german reseller (lipoly.de), shall I contact the reseller or 3dr directly?

Thanks in advance and regards

Marc

• ▶ Reply