ArduCopter-3.2 beta testing

Replies

- Steve Cain > Pomaroli October 5, 2014 at 3:01pm
  
  Presuming the first step in the auto route was set to "takeoff" you activate the route by arming, setting the mode to auto and then raise the throttle above 50%
  
  Best of luck
  - Pomaroli > Steve Cain October 5, 2014 at 3:23pm
    
    yes i've done the mission with droidplanner, then i send it to drone. Followed by arming and raised the throttle. i'm not sure if it was 50% or a little bit lower (but should be enough) and switched to auto.
Phill Simpson October 5, 2014 at 11:07am

I’ve come across a problem with the way that the desired yaw angle (DesYaw) is handled in gusty wind conditions after issuing a Condition-Yaw command with WP_YAW_BEHAVIOR SET TO “0”. Anytime the flight controller cannot maintain the yaw heading to within 10 degrees of the DesYaw value for more than a second or two it changes the DesYaw value to exactly 10 degrees from the actual yaw heading. It seems to be doing it at the same rate that the ATT log entries are being written to the flash log (about 50 times/second).

My use case is an aerial panorama where I position the hex manually then invoke a mission that points the camera North (0 degrees), takes 5 photos at different gimbal pitch elevations, points the camera 32 degrees clockwise, takes another 5 photos, and continues this process every 32 degrees until 10 columns are completed. It then does an RTL.

In one recent panorama shoot I was at high altitude in the mountains with gusty winds when the yaw angles did not come out right. I’ve attached a log. At time 444780Ms the DesYaw angle starts to change from the Condition Yaw command value. It’s real apparent starting at 461916Ms, between 509316 and 515564Ms, and between 543924 and 545603Ms. There was no pilot yaw input and the RC values were well within the deadband. I’ve attached a spreadsheet where I calculated the difference between the Yaw and DesYaw angles to illustrate that the DesYaw angles is getting changed anytime the difference is 10 degrees for more than a second or two. These changes are highlited in color on the spreadsheet. This excludes the initial yaw heading change time (about 4 to 5 seconds) where the hex rotates to the new heading, overshoots a bit, then rapidly seeks out the correct yaw heading.

I’ve looked through the code on Github to see where this might be happening but I was unable to find the correct module. It seems unlikely that this is happening by chance, it there some reason that the flight controller is giving up on maintaining the DesYaw heading when it gets past 10 degrees from the actual yaw? Is this configurable? I see a PARM called RATE_YAW_IMAX that has a default of 1000 Centidegrees (10 Degrees) , is this the configuration PARM that is causing the problem?
- Developer
  
  Randy > Phill Simpson October 5, 2014 at 6:22pm
  
  Phil,
  
  You're right that the desired yaw is dragged around with the actual yaw if it goes off by more than 10degrees. This was added to try and make the platform safer by reducing the chance of sudden yaw movements happening when the environment changes. those "environmental changes" could be any thing from a take-off to the vehicle hitting tree.
  
  We could probably make an exception to always correct the yaw back to the initial target in Auto. I've created a to-do item here. I don't think we will be able to correct this for AC3.2 though probably...
  - Phill Simpson > Randy October 6, 2014 at 10:45am
    
    @Randy
    
    Thank you. When the yaw gets dragged around while doing a panorama it creates gaps in the columns of pictures which can't be discovered until one tries to stitch them together back home.
    
    I agree that this should be an enhancement for after 3.2 is released. At this point the 3.2 codeline should be closed to everything except fixes for critical bugs and fixes for major bugs that don't have a workaround.
- Phill Simpson > Phill Simpson October 5, 2014 at 11:11am
  
  The log file was too big, here's the spreadsheet. I'll find another place it upload the log file if needed.
  
  Log 58 ATT Extract.xls
  
  https://storage.ning.com/topology/rest/1.0/file/get/3702523364?profile=original
Martin Ouellette October 5, 2014 at 10:29am

Hi,

I'm running rc10 on my APM 2.5.

I switched my RX (FrSky) to use CPPM mode yesterday. Now i'm trying to configure my flight mode. I was using 6 flight by mixing my 3-pos switch and gear switch (turnigy 9x remote).

Now my 3-pos switch respond on Channel 6...and APM "wait" for Flight mode on Channel 5.

One of my friend, that run version 3.2 rc2, told me that he recompiled its version after changed int16_t pulsewidth = g.rc_5.radio_in; for int16_t pulsewidth = g.rc_6.radio_in; in control_modes.pde file.

In the solution of 3.2 rc 10 i'm not able to see the file control_modes.pde...and I searched through the other files I didn't find the line "int16_t pulsewidth = g.rc_5.radio_in;".

Can someone can point me where this line is now located ? Or point me of what I need to change to be able to correctly use my CPPM rx with arducopter 3.2 rc 10 ?

thanks in advance
- Brandon Clifford > Martin Ouellette October 5, 2014 at 12:04pm
  
  I swear there used to be a parameter that let you change was ch. your flight mode input was on. I just spent 10 minutes looking for it and couldn't. I am wondering if I reload 3.1.5 if its still there... Was this option removed?
  
  - Nevermind that param. is on APM Plane... I wonder if it can be added to copter.
  - Martin Ouellette > Brandon Clifford October 5, 2014 at 7:11pm
    
    Well, I found the parameter ...
    
    in switches.pde file.
    
    just change int16_t pulsewidth = g.rc_5.radio_in; for int16_t pulsewidth = g.rc_6.radio_in;
  - Richard Kennedy > Brandon Clifford October 5, 2014 at 1:28pm
    
    I wish plane had the ch 7 & 8 options like copter. Not sure why they aren't the same.

of 373

Discussions

ArduCopter-3.2 beta testing

Replies

Activity