New Defect: P2 and H3-3D

ianwood

Taco Wrangler
Joined
Jan 7, 2014
Messages
5,107
Reaction score
2,043
Location
Lost Angeles
Symptoms: horizon drift (gimbal not-level).
Cause: significant location changes (e.g. between Los Angeles and Chicago).
Steps to recreate:

  • Take a brand new P2, update to 3.06, calibrate and test fly it in Los Angeles.
  • Note the perfectly level horizon in all shots.
  • Travel with same P2 to Chicago (in hard case as carry-on luggage).
  • Recalibrate everything.
  • Test fly it in Chicago and notice drifting (not level) horizon in video (up to 1.5 degrees off).
I've done this with two different P2s now. Both exhibit exactly the same behavior. One P2 is brand new. I am going to guess there is actually something in the 3.06 declination fix that influences the IMU output to the H3-3D. The P2 is remembering something between the two locations and causing the IMU output to be off.

Looking for others who have travelled with their P2 / H3-3D using 3.06 firmware to test for similar symptoms.

An example of horizon drift:

navy-pier-tilted.jpg
 
I have travelled 500kms by car and not experienced this. In South Africa. Have also not done any compass or IMU calibrations between locations. All on 3.06.
 
I live here in NY and I just took delivery of my drone, it came from California. I have not been able to go outside to fly it due to rain, but with it sitting level on my desk, i notice exactly what you are explaining. I have not re-calibrated the compass yet.

Curious if anyone has some insight as well.
 
Pacific Barbarian said:
I have travelled 500kms by car and not experienced this. In South Africa. Have also not done any compass or IMU calibrations between locations. All on 3.06.

You should most certainly recalibrate the compass in this situation. But it is curious you haven't seen it and didn't recalibrate.

Horizon drift is sneaky and a lot of people don't even notice it in their shots. In the shot above, you can clearly see the Lake Michigan is not level in the shot. Typically, the gimbal needs to be fully up to notice it.
 
I have a Quanum Q-2D gimbal from hobby king i bought a few months ago and it has manual adjustments for the pitches. Am i missing it, or does the Zenmuse not have manual adjustment?

I can see what you mean with the horizon shots being the worst. Only way to fix that is to edit the photo and manually rotate a few degrees. That gets annoying...
 
ianwood said:
Pacific Barbarian said:
I have travelled 500kms by car and not experienced this. In South Africa. Have also not done any compass or IMU calibrations between locations. All on 3.06.

You should most certainly recalibrate the compass in this situation. But it is curious you haven't seen it and didn't recalibrate.

Horizon drift is sneaky and a lot of people don't even notice it in their shots. In the shot above, you can clearly see the Lake Michigan is not level in the shot. Typically, the gimbal needs to be fully up to notice it.

I purposefully did not do the compass calibration to see how it would respond at a location that far away. And I was quite pleased with the results.

On the horison tilt issue, I used to notice it as bad as in your picture on the previous firmware all the time (together with serious TBE). After few minutes in the air, both improved gradually and went away after a while.

Since upgrading to 3.06, both issues dissiapated. Now I never get the TBE and maybe a slight horison tilt upon start up but you can hardly notice it.

Now, this picture was taken 500kms away from my house, with no compass calibration at new location, but with new firmware 3.06, shortly after take off.





Here is what all my pictures looked like with previous firmware, first picture at the end of the flight, second picture at the beginnng of the flight.






The new 3.06 I am really pleased with, the hovering and just everything seems more stable and it has gained my trust.
 

Attachments

  • uploadfromtaptalk1414121119812.jpg
    uploadfromtaptalk1414121119812.jpg
    141.7 KB · Views: 698
  • uploadfromtaptalk1414121148998.jpg
    uploadfromtaptalk1414121148998.jpg
    133.2 KB · Views: 699
  • uploadfromtaptalk1414121244841.jpg
    uploadfromtaptalk1414121244841.jpg
    112.4 KB · Views: 718
I have been having constant horizon drift with my P2 and H3-3D, no travel involved. I have been flying in the same area the whole time. I've done multiple advanced IMU calibrations and it will fix it briefly, but then I go fly and it returns to 1-2 degrees off. Getting quite frustrating.
 
I have flown in many locations without performing any type of calibration without any problem.
 
OK. Some replies:

  • There is no calibration for the H3-3D other than setting limits in the P2 Assistant. You can reset the gimbal and relaod the firmware. I haven't tried that yet.
  • If you travel over a significant distance, you should recalibrate your compass. Period. If you don't, you're pushing your luck. It's like tightrope walking without a safety net. You make it across until you don't. Let's not debate this here as it is not the topic.
So, back on topic, does anyone know how to hard reset the Phantom back to factory defaults? I'm not talking about hitting the default button on each page in the P2 Assistant. I'm talking about making sure the P2 has no memory of any previous flights or settings. Like a factory reset.

Also does anyone know if the H3-3D IMU listed in the upgrade page is a separate IMU on on the mainboard or is it just a proxy for the IMU in the Naza? It must be something as it has its own firmware. When you put in the anti-interference board, it too has its own firmware level.

Interestingly enough, with respect to this issue, I have one P2 with the anti-interference board and another without. It doesn't seem to make a difference.
 
I cannot answer anynof your questions, and yes, I run the AI board. Researched it and its to reduce jello?
 
So Ian, i have noticed this for awhile now. I did do some traveling to Ohio, but i saw the issue before that. The issue i am having with mine is when i yaw the craft in anyway ( fast or slow) the horizon slightly shifts down into the direction of the yaw and when i stop it levels out. I actually did really notice this after the 3.06 update and i even went back to footage before the update and didn't see it. I have already done the gimbal reset and that did not fix it. Unfortunately with my P2 deciding it wants to go all kamakazi it is a DJI again, so i can't dig into it more. The video posted does show how to reset the naza, so maybe you can try that. If you try it let us know if it solved it.
 
ianwood said:
Also does anyone know if the H3-3D IMU listed in the upgrade page is a separate IMU on on the mainboard or is it just a proxy for the IMU in the Naza? It must be something as it has its own firmware. When you put in the anti-interference board, it too has its own firmware level.

I don't have a P2 handy to connect and see what the upgrade list there looks like, but I do know that the H3-3D does have an on-board IMU that can sense movement independent of the NAZA's IMU...the input from both is used in the positioning computations.

In the NAZA v2 Assistant I believe the upgrade page labels it as "gimbal CMU" or something like that.
 
Yes, my original P2 shows both a CMU and IMU. I think the CMU is the controller for the gimbal.

I don't know if the gimbal IMU is actually another IMU (probably on the P2 mainboard) or if it simply the Naza IMU signals being duplicated (and possibly altered) and sent to the gimbal. But, there's something in there that takes firmware and calls itself the gimbal IMU. Maybe it's internal.
 
There's no IMU in the Phantom mainboard, just the built in GCU (data bridge between NAZA and gimbal).

You can confirm that the gimbal has its own IMU (functionally if not by name) the same way I stumbled across that fact: connect the gimbal's 8-pin cable to the mainboard on your Phantom, but mount the gimbal to a second Phantom or other frame. You'll find that when you power on the Phantom (and wait till the gimbal initiates), the gimbal not only adjusts for movement of the Phantom, but also for movements of the frame it's physically mounted on.
 
OK. For some reason, I thought it wasn't. Makes sense. CMU is probably the "anti-interference board" as my new P2 doesn't have one yet and it doesn't show up in the list. I'm adding it tonight, so we'll see.

I am going to try reseting the gimbal first, and if that doesn't work, then the Naza itself. Hopefully, one will work to keep the gimbal level.

If I have to reset the Naza every time I travel to get a level shot, I'll do it but I'll probably extend the reset pins to the body of the P2 so I don't have to crack the shell every time I get on a plane (on one right now as a matter of fact).
 
OI Photography said:
Cool, but...are you talking about the reset process of jumping the two pins on the gimbal's board in the camera cradle?

That's the only one I know of.
 
ianwood said:
OI Photography said:
Cool, but...are you talking about the reset process of jumping the two pins on the gimbal's board in the camera cradle?

That's the only one I know of.

Like others I learned the hard way that doing that reset on a functional (even though flawed) gimbal can brick it. It seems it should only be used on a gimbal that's completely limp from a failed firmware upgrade.
 
I thought you can only brick it if you remove power before re-applying the firmware. Without a bootloader, you can't upload firmware. I may start with the Naza but I think the same rule applies. Once you wipe the NV ram, you can't kill the power before applying the new firmware.
 

Recent Posts

Members online

No members online now.

Forum statistics

Threads
143,086
Messages
1,467,525
Members
104,965
Latest member
cokersean20