Fixing DJI's Compass Problem

[ianwood]

UPDATE:

As of April 13, 2014, there are 87 P2x users in the forums reporting the same issue. If you are affected, make sure you say so here! What we know so far:

• Affects all P2x in areas of high magnetic declination (+/-10 or higher).
• Curved flight and hooking is worst when first flying. It improves during flight but may take several minutes.
• Once power is removed, the P2 forgets any adaptions it made during the previous flight.
• It appears to not affect the P1.
• Areas with very high declination will experience TBE during initial hover.
• Several people have returned their P2x thinking it was defective only to get a new one with the same problem.
• This is a firmware defect. Possibly a signing bug in the code.
• This defect is still present in versions 2.0 and 3.0 of the firmware.
• Measuring the angle between initial true heading and initial course lock heading will give you the real margin of error.
• Rotating the compass is not a recommended workaround. Some further testing is being done here.
• Using a very rough bottom up model, it is expected that at least 1 in 4 P2x users experience this problem.
• DJI is looking at the problem and has selected 5 of us to do some testing to help diagnose the problem. We will be evaluating the same raw data to see if we can further identify the source of the problem.

Fixing DJI's Compass Problem

Some of you know about this and some of you aren't affected. If you live in the following areas, you are subject to this problem even if you don't know it:

• Western US and Canada
• Northeastern US and Canada
• Eastern OZ and all of NZ
• Northeastern South America
• Southern Africa

Your compass and GPS are disagreeing over which way is forward. This results in curving (hooking), tracking to the left or right in GPS and course lock, and/or not hovering in place. This is a caused by a defect in DJI's firmware which they need to fix.

[youtube]http://www.youtube.com/watch?v=9BUilGeH74Q[/youtube]

I am working through some contacts to get in touch with DJI's LA office to present the issue in a clear fashion along with the backing of the user communities like this one. Most importantly, I want to show them how many other people have this issue.

If you are in one of the areas above, confirm in this thread that you have experienced this issue. I have written a complete summary of the issue here: DJI Phantom and Magnetic Declination which I will be sharing with DJI as well. Please share this so that we can collect as many of the affected users as possible.

And before anyone says "Just rotate your compass", that is not the answer. DJI has said so themselves.

 

[nhoover]

In N. Calif at 37, -122 I have the problem. It's not consistent, but I have to calibrate every flight. If I fly around quite a bit and switch in and out of ATTI mode the problem seems to be reduced. In one flight recently, near the end, it was actually flying straight and not drifting left in GPS mode. That was the first time in over 40 flights since Jan 25 I've seen that.

I never had this problem for the first 20 flights until I installed a Flytrex core on Jan 25. Flying with or without it now doesn't matter but out of the box the P2V was straight as an arrow and never drifted - I only calibrated it once before the first of those 20 flights.

I've tried stuff like taking my laptop with a long usb cable to the middle of a remote field and doing full IMU calibration even though it didn't need it. Calibrated the compass all directions a million times. Have not crashed other than a harmless tipover.

Would love a fix.

 

[dcozad]

Southern California Redlands, 34.04 -117.19I did not have the issue out of the box, however in the first 15-20 flights. I get straight flight in ATT mode, but in GPS I get the left hook forward and right hook back along with the toilet bowl as it comes to a stop. I have done all the calibration IMU and Compass, I just learned to fly in att mode for video. I would love to have a real fix too.

 

[SkyView]

I believe your analysis is correct. My experience with my P2V matches your description of the problem.

I bought a P2V in December and on those few occasions when I could fly it – due to the poor weather – it deviated to the right in flight and then went into circles of ever increasing radius during hover. I later learned that this is called the "toilet bowl effect."

When I informed my dealer that this was happening he wrote back with instructions for me to recalibrate the compass and perform an IMU calibration using the assistant software. I did this and the software indicated that the calibrations numbers were in the correct range. I informed my dealer of this result but never heard back from him on the issue.

Frustrated and without a solution I returned my P2V. I received a full refund.

Going further, last week I sent an e-mail to DJI corporate customer service. I got a useless marketing reply!

I believe your analysis of the problem is correct. I live in halifax, Nova Scotia, Canada, which is in the northeastern arm of North America. Thank you for posting this thread – it gives me some closure on the problem.

In the end I would still like to get a P2V that functions properly. Hopefully, DJI can update the firmware to fix this problem.

Halifax
Latitude: 44° 38' 60" N
Longitude: 63° 36' 0" W
Magnetic declination: -17° 49' WEST
Declination is NEGATIVE
Inclination: 67° 26'

 

[Shrimpfarmer]

Here in the UK its all flying fine. I had heard of problems but this looks like a good thread to find a solution. Can you give more information on what you have discovered and also report your explanation of what your sending to DJI.

Good luck in your quest. It must suck if the quad does not fly as intended.

 

[horsepowerphoto]

I am near SkyView, and I have the same issue. Rotating the compass fixed the symptoms for me. Though it should be fixed in the firmware!

 

[xgeek]

Can you give more information on what you have discovered and also report your explanation of what your sending to DJI.

Document link in first post

 

[Shrimpfarmer]

Can you give more information on what you have discovered and also report your explanation of what your sending to DJI.

Document link in first post

Its my age you know .......

 

[Shrimpfarmer]

I have read it now. What an excellent analysis and well presented report. Lets hope they act upon it.

 

[RemE]

I have read it now. What an excellent analysis and well presented report. Lets hope they act upon it.

Agreed very nice write up! As far as I know the P2 and P2V GPS is a receiver only so its orientation (besides pointing up) is not relevant. I think that it's the relationship between the compass and the NAZA box itself as that is where the multi-axis accelerometers are located and they must be oriented correctly (arrow forward) and to the compass as well.

I wonder if having the compass out on a leg, instead of closer to the center axis of the model makes any difference. I'm in Southern Calif (+12 Dec) and have been flying my P2V and P2 most weekends. I haven't noticed this issue but will be looking much closer next time out!

I have a 450 with the GPS and compass in the puck which I believe would be more centered. I have not built it up yet to do a comparison test

 

[ianwood]

So far, I have 21 people confirming this issue with the P2/P2V. I want 50 or more. This is a bug in the Naza and it's time to get DJI to fix it (or at least tell us how to).

If you are in western or northeastern US and Canada, you need to test for this and chime in here! Same for NZ, ZA, eastern OZ. Also, northeast of South America. All you Europeans and American mid-westerners consider yourself lucky.

(Appreciate the compliments on the doc. I do a lot of technical briefing in my day job...)

 

[flight-of-eye]

I am in Northern California, received my unit in February of 2014 and found that this is an intermittent issue. I have done the so called "dance", but still remains the toliet bowl effect. I have done some research on this matter and found that there may be a fix in the axis mounting range values. However, at this time I have not found a way to change the values in the assistant software. When linked to the software my values are as follows; X=0, Y=0, Z=0. Most of my research points to this being a possible reason as at least one value, Z, should be a negative.

I found this video for reference.

http://www.youtube.com/watch?v=9WgEcoDRL3o

 

[BenDronePilot]

In N. Calif at 37, -122 I have the problem. It's not consistent, but I have to calibrate every flight. If I fly around quite a bit and switch in and out of ATTI mode the problem seems to be reduced. In one flight recently, near the end, it was actually flying straight and not drifting left in GPS mode. That was the first time in over 40 flights since Jan 25 I've seen that.

I never had this problem for the first 20 flights until I installed a Flytrex core on Jan 25. Flying with or without it now doesn't matter but out of the box the P2V was straight as an arrow and never drifted - I only calibrated it once before the first of those 20 flights.

I've tried stuff like taking my laptop with a long usb cable to the middle of a remote field and doing full IMU calibration even though it didn't need it. Calibrated the compass all directions a million times. Have not crashed other than a harmless tipover.

Would love a fix.

Similar issue to you my p2v worked fine initially then later began drifting to the right, on some flights after running through most of my battery it actually flew straight without the drift but change battery and same thing. Repeated compas advanced and basic calibrations did not help. I contacted dronefly and they sent me a free shipping label to get a new replacement unit as I'm also have other issues too.

 

[DesertFlyer53]

There seems to be two threads with the same subject. One in the P2 section and one in the P2V section. Do I need to post in both!?

 

[ianwood]

UPDATE: There's 30 of us P2 and P2V users so far who have reported that they have this issue. Every single one of us is in an area where the magnetic declination is greater than +/-10. This supports our belief that this is a compass related bug. Further supporting this, we have 2 (or is it 3) people who have both P1 and P2 and only see the problem on the P2/P2V.

If you have this problem and haven't yet reported it here, please do so. I want 50 or more before I take it to DJI.

Also, can someone tell me what the latest firmware version is for the Vision? I am assuming everyone who has reported here is on it.

 

[SkyView]

DJI Phantom 2 Vision
Hardware --------- Firmware
Main Controller --- 1.00
GPS ----------------- 6.0.1.3
Receiver-5.8G ---- 1.0.1.3
P330CB ------------ 1.0.1.16
Battery ------------ 1.6.0.0

 

[Tripnman]

Count me in... Northern California here. My P2V dog-legs to the left after any significant forward movement. No amount of recalibration seems to fix it. All firmware is up to date and I have had no uncontrolled landings (yet).

 

[layertone]

Count me in as well. I'm in Northern California. I'd love to take my P2V out, but now so nervous about the drifting & crashing, just left it sitting there waiting for a firmware update / fix. Was going to send it in, but looks like others are having the same issues. It crashed the first time, costed a bundle of $$$ for repair parts. Updated all firmware & software / calibration.

 

[cougar]

Hi There,
A new pilot -to-be.
My new Phantom is due to arrive next week and these problems make me shiver, live in Nova Scotia (Halifax)
Hope a fix will be very soon or it has to go back.
:?:

 

[ianwood]

You guys should be able to fly despite this issue. It just means you have to be alert and compensate for it. Cougar, please come back and confirm you see the issues after your first flight. Good to get a couple out-of-the-box reports. Thx.