Redundant CC3D orientations help clarify please?
« on: August 25, 2018, 07:09:54 pm »
So I had a new build flying with the default cc3d orientation (top up, arrow forward).   I want to move the position of my board and pitch the FC 90 deg forward (Arrow pointing down top of FC facing forward).   Ordinarily I would think you just change the "Virtual attitude relative to board" to 90 deg in configuration, however that hasn't seemed to work.   
Two things that confuse me here are:
1. Doesn't it automatically set that when you go to level the craft?   It should already know its orientation then, right?
2. Supporting the point above- the "flight data" tab indicates the proper orientation and a level horizon view out the "window".

So adding a value to that virtual orientation sure seems redundant and I'm not sure why it's not.  It still flips when trying to take off & I've tried it both ways.  Any help?

f5soh

  • *****
  • 4465
    • LibrePilot
Re: Redundant CC3D orientations help clarify please?
« Reply #1 on: August 25, 2018, 08:00:53 pm »
Its better set the initial orientation using the Virtual attitude rotation, in your case Roll:0 Pitch:90 Yaw:0
and do the leveling.

About the instant flip, be sure the motors are correctly assigned according to the drawing in Vehicle tab and test each motor looking the names (VtolMotorNW, NE...) in Output tab.

Double check if the "Zero the integral when Throttle is low" is checked in Stabilization > Basic (at bottom)

Re: Redundant CC3D orientations help clarify please?
« Reply #2 on: August 25, 2018, 09:00:40 pm »
>Its better set the initial orientation using the Virtual attitude rotation, in your case Roll:0 Pitch:90 Yaw:0
>and do the leveling.

Ok, I thought so too- that's what I did initially, recently undid it, and now I'm going to redo it.  I'll check the motors- I do have the "reverse all motors" box checked right now but that correctly inverts all the props on the diagram which matches the rotation on the quad.  I will double check that output tab & zero integral thing though- thanks!  All the motors spin properly and in proper order when I calibrate them also.   I may take out the cc3d (which sucks 'cause its glued in place now) and orient it default again to see if it resumes flying correctly.

Re: Redundant CC3D orientations help clarify please?
« Reply #3 on: August 25, 2018, 11:34:00 pm »
Does the PFD display in the upper left corner of the Flight Data page look correct and respond correctly?  Starting with it level in your hand, when you pitch it forward, you should see more brown dirt since the view from the cockpit sees more earth.  When you roll left the screen should roll RIGHT.

You can leave the board mounted as it is, and set Rotate Virtual back to 0,0,0 and then hold the quad so the BOARD is level and see if the PFD looks and works OK.

Setting "reverse all motors" incorrectly will cause instant flip.

Long ago I recall a problem with CC3D.  I also wanted to rotate a CC3D forward 90 degrees, but that does not work.  That may or may not still be the case.  It is not a problem with Revo using INS13.

Re: Redundant CC3D orientations help clarify please?
« Reply #4 on: August 26, 2018, 03:24:17 am »
I'm not sure what PFD is but I assume you mean the d is the "Display" that shows the brown dirt and blue sky.  Assuming yes, then yet it all reacts as you describe.  I will say that the bit about rolling left/right is counter intuitive but I guess that's why you used all caps! :)     The 3d rendered quad shown on the same flight data page also reacts appropriately.

Doing the test you suggested (board level- but not running a new "level" routine results in an unexpectedly level horizon still but left right tilts are kind of muted, thought they react somewhat.  Strange- I will have to experiment some more.  Maybe it Is that CC3D problem but I'm surprised I don't see more discussion about it.  I guess it's just an unusual orientation.

Re: Redundant CC3D orientations help clarify please?
« Reply #5 on: August 26, 2018, 03:44:13 am »
WIth RotateVirtual pitch=90 and quad level, the PFD looks good.  That means there is no problem with rotating the board like you have done.

Holding the board level (quad pitched back 90 degrees) with RotateVirtual set to 0,0 and PFD should act perfectly normal.  If you roll the board left 90 degrees, but you don't see the PFD rotate right exactly 90 degrees (and for instance 45 right should show PFD=45 left), you have a problem.

Have you seen this thread?
https://forum.librepilot.org/index.php?topic=4408.0

Re: Redundant CC3D orientations help clarify please?
« Reply #6 on: September 01, 2018, 05:59:53 pm »
Hi,
I have read that now- and in spite of that have been struggling until just now... well, ok, I'm still struggling but I tried using the new 16.09+r711 version and maybe that helped?
Here's whats going on now:   For some reason on this attempt my motors seem to have rotated 180deg yaw before I even set them to do that, that is to say that during motor testing motor #1 is now #3... but whatever if that's how it wants to work- no problem.  So as a result I changed my pitch to -90 and yaw to 180.  Now the PFD operates correctly again, and more importantly it manages to not insta-flip.  It's still very wobbly and not really flyable but I can fly it enough to determine that the orientations are correctly defined.  That is- pitch forward/back and Roll left/right are making the correct adjustments.  I've since dialed in the motors very carefully and leveled it as accurately as I can without a proper level handy where I am.  So it's close, but I know from the default orientation/configuration how its supposed to fly (MUCH nicer!).   I'd been attempting to run Cleanflight instead but so far haven't succeed in that regard- lots of weird driver configurations.  I also have a Naze32 to play with but haven't succeeded with either board yet.  I'm happy to be so close with Librepilot so plan on working out some of these final details if i can.

Re: Redundant CC3D orientations help clarify please?
« Reply #7 on: September 02, 2018, 06:37:53 am »
On Output page, do you have at least PWMSync or [email protected] for all banks that are attached to a motor?  [email protected] will cause oscillations with stock PIDs.