Hi everyone, I have not seen a fix for this issue on this sight and wanted to share my results. If this has already been posted sorry but it was so helpful to me i wanted to share! When I first got my p2 while hovering it would yaw on a dime and never drift more then a few inches in any one direction. Over time I have added fpv (gimbal, gopro, video tx, mini iosd, tall wide landing gear) and it's ability to do this has slowly gotten worse and worse. After doing an advanced imu callinration the problem still persisted, sometimes drifting in up to a 10 foot circle. I would end up doing the advanced calibration several times to no avail. Finally after searching for a fix for weeks I came across the following info (http://forum.dji.com/forum.php?mod=view ... 343&page=1) and after reading completely through it I decided to give their fix it a try. I bought a small bullseye level (they call it a spirit level in the link) and went to work. The desk that I used while doing the advanced calibration was almost perfectly level like always but the flight controller was not even close to level. I shimmed the legs of my p2 with playing cards until the bullseye level was perfect, then plugged in the GPS unit (leaving the top of the shell off to make sure nothing changed) and did an advanced calibration. I then made a quick test flight and it would yaw PERFECTLY after my first calibration!!!! The whole process took me less then 30 minutes! If you experience yaw drift I highly recommend leveling your flight controller with a bullseye level and then doing an advanced calibration!
Finally fixed my Yaw Drift
- Thread starter sgrim70
- Start date
sgrim70 said:
Thanks for posting. I have added a good bit of stuff too and I am afraid of crashing into stuff when I yaw now. Will read that link.
Well I put a flat board across my motors and used an iphone app called "Bubble Level" (there are about a million of these...I tried to pick out one that gave a numeric reading...) and my desk and therefore my motors were not level during IMU calibration. I re-calibrated with a level quad last night and now have to wait out some weather before I could try it. I am assuming that the iphone level is fairly close, will let you know.
terrylowe said:
I would also get a bullseye level ($2 at a hardware store) and make sure the flight controller inside the phantom is level when calibrating. In my opinion that's the most important component to level! When my motors were level the FC was still not level at all so that's when I decided to only worry about the FC being level. After leveling this way and doing the advanced calibration I have had perfect yaws with no drift.
Got to go out. IT WORKED!!! I can spin on a pole now. When I leg go of the yaw, it sort of sling shots out of position a little like I was in ATT mode rather than GPS mode. It works MUCH better than it did. I can trust it for some tighter squeezes now.
I just got my Phantom 2 H4-3D in this week so I haven't tested this out yet. Will this attempt at leveling and calibrating result in the Phantom spinning in place for both GPS and ATTI? (assuming no wind)
ianwood
Taco Wrangler
Well, this is interesting. Whoever was first to go full nerd and test this, you've got my respect! I bet you DJI doesn't even know this.
Fascinating read last night on the DJI forum by the pioneers of this discovery. Don't think my P2 has yaw drift, but I will watch next flight, and likely check IMU level next time hood is open. Any idea if non-level IMU calibration could influence horizon tilt issue? That's a problem I would love to see understood and solved.
Re circular bubble levels, don't go too cheap, or if you do, verify its accuracy with a known quality stick level. I bought a plastic circular one at local hardware store, for purposes of quickly leveling my case prior to P2 power-on, and discovered plastic level is off. With stick levels, a 180 of the stick should read same as initial read. If not, stick is bad. Verify stick, then use stick to verify circular bubble level.
It would take soooo little for DJI to delight hundreds or thousands of us by being open book on how all this voodoo works, and how users can get their bird back to factory performance (or better) by a few well-documented procedures.
Hats off to the IMU level pioneers.
Kelly
Re circular bubble levels, don't go too cheap, or if you do, verify its accuracy with a known quality stick level. I bought a plastic circular one at local hardware store, for purposes of quickly leveling my case prior to P2 power-on, and discovered plastic level is off. With stick levels, a 180 of the stick should read same as initial read. If not, stick is bad. Verify stick, then use stick to verify circular bubble level.
It would take soooo little for DJI to delight hundreds or thousands of us by being open book on how all this voodoo works, and how users can get their bird back to factory performance (or better) by a few well-documented procedures.
Hats off to the IMU level pioneers.
Kelly
I am glad everyone else is having good results after doing this like I did! I don't remember the original person that thought of FC leveling in the link that I posted at the start of this thread but they deserve all of the credit!!!! After roughly 10 flights my p2 still yaws on a dime (wind or no wind)!
sgrim70 said:
Good to know, thanks for sharing, wondering about your yaw gain, if you don't mind.
Cheers
FedeVision said:
I have not changed any of my gains besides my gimbal tilt gain (I have it set at 4).
I'm having one of those "less than brilliant" weeks. I'll blame it on the cows.
Can someone confirm, the FC is the NAZA module?
Anyone have a picture of a good level sitting on the FC, maybe even a manufacturer or link?
And then it's as simple as performing an Advanced IMU Calibration on a perfectly level FC.
Sounds too logical and simple.
Thanks for sharing sgrim70. Ya done good.
EDIT:
Thank you for your persistence Cory.
Can someone confirm, the FC is the NAZA module?
Anyone have a picture of a good level sitting on the FC, maybe even a manufacturer or link?
And then it's as simple as performing an Advanced IMU Calibration on a perfectly level FC.
Sounds too logical and simple.
Thanks for sharing sgrim70. Ya done good.
EDIT:
Thank you for your persistence Cory.
dirkclod
Premium Pilot
The only place I can see where you could put a level is where I got that one in the 2nd picture .
Now if someone can say what is right ..using level there or with the glass on shafts I will dance at your next wedding !
Last edited:
dirkclod
Premium Pilot
I just spoke with someone that knows his stuff and he did confirm this .
So I guess I will try and level with the little bubble on it across both ways next and see if it gets even better than with the glass .
[ Hey Hunch ]
I noticed an incredible increase in yaw drift while trying out the 9450's on my P2.
Thanks for sharing, most interesting and it does sound very logical.
Going to try this today. Sunny weather with a light breeze. Perfect!
Thanks for sharing, most interesting and it does sound very logical.
Going to try this today. Sunny weather with a light breeze. Perfect!
Try and get a "bullseye level" that looks like this (http://www.preciseshooter.com/blog/Mosi ... 0level.jpg). It was only $1 or $2 at my local Menards (Home Depot or Lowes also) and it will be much easier to get the FC (naza) perfectly level since it levels in all directions at the same time. The top of the naza does have two levels to its top but there was plenty of room to place my bullseye level on the upper part of it.
dirkclod
Premium Pilot
OK,here's what I just tried . Having the table I know is level and been doing the advanced calibration's on since I started I used that little level and redid a advan. cal.
My table is level .
Now with quad on it and level on Naza ya see what I got , not level .
So I shimmed the corners till I got it level and did a advan. cal.
Now I had started doing this by using the glass on the top of the shafts and it had helped
keeping it tighter when yawing .
So I thought while I did have the Naza level I would put the glass on and check and see if it was off or level
there also .
Nope..off just a little .
Now I put it all back together and went outside and tried it .
Was really surprised at how it had made a difference and it did spin
tighter and held in place when I stopped .
I do believe if you are having issues with it being
sloppy when yawing this does fix it .
Thank You to the ones that did figure this out
and I just posted this again I'm sure to show ya what I saw !!
The glass helped a lot but going in did make a difference with mine !
Just thought I would share . :mrgreen:
Last edited:
N017RW
Premium Pilot
Seems to me if your going to this trouble, BOTH the FC and the Thrust Vector (glass across 4 motor shafts) should be coplanar.
Once that is achieved then level using one or the other as the measuring point before IMU cal.
Once that is achieved then level using one or the other as the measuring point before IMU cal.
Similar threads
