After using flat washers, business cards, folded paper, I realized there was a better way of getting a level surface for calibration of the IMU on my Phantom.
Piece of 5/16 lexan, 1/2 inch holes drilled in each corner, four 7/16" nuts and bolts with the nuts epoxied over each corner hole. This results in a sturdy flat surface and able to be leveled no matter how unlevel the tabletop is.

Good idea,  I like it

Yes certainly takes the guesswork out of ensuring a level surface.  I assume this is for IMU calibration- but heres a thought :what do we have to get level- the motors or the IMU box? and if it is the box, how do we know if the box is parallel to any surface we may be measuring from?  more sleepless nights!

rod, i would think of it like a tare situation.  you level the motors and if the IMU is slightly out of level but the motors are, that is what the IMUs new tare level is (even if not pure).  so it wants to keep itself exactly as it was at calibration (which even if the IMU isnt level the craft will constantly want to put the IMU in that exact state of unlevelness resulting in perfectly level motors).  the diff between imu level and motor level is, imho, what causes the yaw drift to begin with.  if both were the exact same levelness, you wouldnt need to calibrate, i think the calibration is just a way to compensate for that difference and make the craft want to keep the imu exactly to a point that keeps teh motors level.  
.
that is at least how i interpret all of the banter we've been having on the subject.  just my interpretation though.


.
and great idea with the bolts...

gnixon2015 Posted at 2015-2-4 12:39
rod, i would think of it like a tare situation.  you level the motors and if the IMU is slightly out ...

Hi G, yes that makes sense. So the null point is set with the motors level- so the magic box could be a few degrees off  parallelism with the motors, but it now knows where level is for the motors-- got it!.

yes, rod, that seems to be the right deduction from all the 'results' we've seen.  now the true test of that would be to level the IMU itself (and confirm that the motors were slightly out of level at the same time) and see if a yaw presented itself.  however, i thought that someone posted something similar to that (where they did it not necessarily to prove this point but it was the result they got nonetheless) which is what i thought made me think about it this way.  

i love this idea but ill probably just get a spirit level and do the plexiglass or glass on the motor tips.  honestly, the only time mine doesnt yaw right is in decent wind.  so for now im leaving it alone.

Its a good idea, but I would not use a phone to do it, rather an old fashion spirit level for me.

The phone has to be calibrated to level first. How did you calibrate the phone?

What irritates me is why won't DJI who is supposed to be monitoring this forum won't provide some input to what really needs to be done to get a good calibration. It just boggles my mind.

The app has that feature in the settings for calibration. I have a two axis spirit level I use to double check as well. This allows me to pop the top shell off the a/c to level the imu itself

Rexter! How's the gimbal gaurd, any issues ?

tonyphantom147. Posted at 2015-2-6 06:35
Rexter! How's the gimbal gaurd, any issues ?

Haven't really had an occasion for it to come into play, but it's a very sturdy piece of kit and I feel safer flying with it as well as the c/f protector attached to the bottom of the legs

rod Posted at 2015-2-4 12:35
Yes certainly takes the guesswork out of ensuring a level surface.  I assume this is for IM ...

To answer the question, that's another reason I made this thing to make it quick and easy to level either the motors or the IMU. I've done both now since making the levelling plate and can say that I get a more steady flat flight from levelling the IMU than I did the motors.

I'm curious if leveling the IMU and not the motors will cause the craft to lean if the IMU and motors aren't parallel and therefore cause the video/pictures to be tilted. Can anyone who has leveled the IMU and recalibrated verify if their images still have a level horizon?

Yes, the horizon is level as can be after calibrating IMU using my system

Hi.

Waiting for my drone, I've also prepared such a plate.
But I'm using a .1mm/m spirit level instead of an android or Iphone  app.

I also have a .01mm/m spirit level to level laser beams. There's a plate to calibrate such a level, and I must rotate it 180°, record both measurements, do an average. Stupid tool for such a drone, but yes, I do not really know the level tolerances for calibrating !

My point of view : the level should be done on the compass itself !

I'm also thinking of a central pivot that could rotate the drone for the compass calibration. Do not know if it can improve it !
Like a cheese plate or a micro oven plate ?

Do not know if a steady speed for this calibration matters or not !

...
And 3 points is enough and easier to level a plate : one in the middle of an edge, 2 in each corners of the opposite edge.
You could easily check this with your plate : just drill one extra hole.