GPS Mode and the Compass J Hook

[GMANNZ]

Well replaced the bad motor I had and also the compass .... seems didnt make a huge difference so as some of you say look system/hardware related.

I tried the GPS->> ATTI ->> GPS switching made no difference that I could tell.

 

[ianwood]

Well replaced the bad motor I had and also the compass .... seems didnt make a huge difference so as some of you say look system/hardware related.

I tried the GPS->> ATTI ->> GPS switching made no difference that I could tell.

Congrats! According to the magnetic map, you have the largest declination shift we've seen yet! Wellington has a 22 degree difference between magnetic and true North. So, in theory you should be seeing the hook more than any of us and if we can fix it for you, we can fix it for everyone. Lucky you! Southeastern OZ and NZ are not very far from the magnetic South pole which also means declination changes fairly rapidly. From North to South, the declination in New Zealand changes by a whopping 8 degrees.

 

[Xrover]

Everyone having this problem.
PLEASE ANSWER THIS QUESTION

Are you running a GoPro 3 as part of your setup? (I'm not talking about the wifi being turned on issue. I know we all keep that turned off)

I have been reading reports about the GoPros emitting interference. Just another possible culprit.

 

[shartlza]

Everyone having this problem.
PLEASE ANSWER THIS QUESTION

Are you running a GoPro 3 as part of your setup? (I'm not talking about the wifi being turned on issue. I know we all keep that turned off)

I have been reading reports about the GoPros emitting interference. Just another possible culprit.

I am having the issues with no GoPro or anything else attached.

 

[Xrover]

Shartlza,

Thanks..

 

[Xrover]

"Almmohd" has the same issue but his hooks to the Right.
Guess where he lives?
Trinadad.. (Minus 12 degree Magnetic Deviation)

This hooking issue sounds like confusion between the GPS (Probably programmed to True North) and the Compass (Magnetic North)

China has a -2 Deviation. Probably no hooking or drifting when they do their testing.

I think if DJI includes a setting in the P2 Assistant software to input the Magnetic Compass offset, it will solve the drifting/Hooking issue.

 

[ianwood]

Well, the Naza has a built-in adaption function that will compensate for what it believes are compass errors up to +/-5 degrees. This is why some people see the hook more when they fist fly. A lat/lon look up table that fully compensates for declination automatically wherever you may be would be far better. Are you listening DJI?

I will be trying some shims tomorrow on the compass.

 

[Xrover]

I think the shim idea might be a good solution. So are you going to rotate the compass 12 degrees clockwise?

 

[Madkeen]

Everyone having this problem.
PLEASE ANSWER THIS QUESTION

Are you running a GoPro 3 as part of your setup? (I'm not talking about the wifi being turned on issue. I know we all keep that turned off)

I have been reading reports about the GoPros emitting interference. Just another possible culprit.

Have tried it with the gopro On, off and completely removed. Makes no difference. Still have the issue.

 

[ianwood]

Testing for a fix

I was pretty excited to verify that shimming the compass would fix the J hook and course lock issues. Alas, it didn't work out that way. In short, shimming helps but it doesn't seem to be the fix. I tried a few different angles aligning it to a compass rose printed on a piece of paper.

The tests:

• No shim (control): Calibrated compass. Straight out in GPS mode, observed left hook and right hook in reverse. Course lock was -15 degrees from forward. Switched to ATTI and flew for a minute, perfectly straight. Back to GPS and noticed the hook was nearly gone. Course lock after ATTI was unchanged and still about -15 degrees off. Switched off Phantom. Retested without calibration to see if correction from previous flight would remain. It didn't.
• +12 degree shim: This is the one that should have worked as it corrects for the local declination. Calibrated compass. Stright out in GPS mode, observed left hook still there (WTF?) as with right in reverse (double WTF?). Course lock was still about -15 degrees off. GPS flight after ATTI use, the hook was finally gone. Course lock was still wrong.

I was not pleased at this point. My theory was unraveling. But, I tested further with some wider angles:

• -15 degree shim: Just to see if I'm dyslexic! No. This one failed. Compass calibrated but the Phantom did not move predictably. It was a short flight. I had to wrestle it to a landing in the wind.
• +26 degree shim: Calibrated compass. Hook was gone right from take off! Straight out and straight back! Course lock closer to intended direction of travel but still a little off. Switched to ATTI mode and then back to GPS and observed a slight right hook! Uggh! Course lock remained close but still off to the left.

So, I am at a loss. This is still a declination issue as there's a group of us all with the same issue and we all are in areas with greater than +/- 10 degrees of declination. But, the DJI firmware is doing something funky and unlike the Naza users with the integrated compass and GPS who turn the puck to correct for declination, shimming the compass doesn't provide the correction as expected. And no matter what you do, course lock is broken. Seems the Phantom version of the Naza may have some subtle but important differences with the regular Naza. Dare I say, some bugs! Bugs that need to be exterminated!

I think it's about time I contact DJI and ask them why their birds fly crooked. Fortunately, they have a location in Burbank which is only a 15 minute drive from my home. I could drop in on them and give them a real-time demonstration so they can relay back to their colleagues in Shenzhen where the problem doesn't exist. I need to find their actual address in Burbank.

 

[Madkeen]

Thanks for all your effort to fix this.

Hopefully DJI will listen, but will not hold my breath.

I never noticed this problem when I was flying in Nha Trang, Vietnam. Not that I was looking for it either. But I will be back there in 2 weeks so will check it out

 

[Almmohd]

I've had to return my P2V because the compass was so messed up it kept asking for calibration and nothing I did worked. I'm just glad my P2 flies [SMILING FACE WITH OPEN MOUTH AND COLD SWEAT]


Sent from my iPhone using Tapatalk

 

[Xrover]

The more i thought about repositioning the compass, the more I came to the conclusion that it wouldn't make a difference. ianwood did an extensive test and pretty much put any further questions on that theory to rest.

I grabbed a couple magnetic compasses i have and stood and pointed the compass facing North. I then kept my body pointed North but turned the compass to the right 12 degrees. I walked fwd. (North) and the compass of course still reads my heading as North. The only way to physically correct deviation on the magnetic compass is to use a tiny magnet and place it on the compass so that it has a 12 degree positive correction. ( Los Angeles deviation)

I didn't try it on my Phantom and would tell everyone Don't even think about doing this to yours.

After reading so many posts on different forums, I have come to the conclusion that DJI has already included corrections in past firmware updates for the controller. Unfortunately it takes the controller a little while to build up a correction table based on flight time data acquired during your current flight. After a while, It determines your Lat/Long and eventually ads in the deviation correction to the compass input for your location. I think you would need to fly a consistent course line until the GPS has a solid fix on your position and heading.

I think a more optimized algorithm by DJI would probably speed up the process. Hopefully another firmware update fixing this is in the near future. I haven't done a lot of flying in ATTi mode but that mode supposedly has much less drift.

 

[ianwood]

Well, you might be right about the automatic adjustment: http://www.rcgroups.com/forums/showpost ... ount=15039

But, there's a world of misinformation out there if it's true. And it doesn't work (which might explain the misinformation). It shouldn't need to adapt over several flights. DIYDrones has a lookup table that takes up 1.6K and is accurate to 0.5 degrees. What I don't get is the relative simplicity of dealing with declination. In a Cessna, you switch it all on, note the compass heading, compensate for declination and deviation and dial it into your gyroscopic compass and away you go. Every once in a while you check to make sure the gyro hasn't drifted and that's it. The Naza should be able to do all that and more.

Even more concerning is simple math seems to change as you change modes (e.g. GPS > ATTI > GPS). And I think they forgot to add it at all to course lock. Going to email DJI.

BTW, I saw your PM, I am doing all this testing out in Elysian park near Dodger Stadium if you want to meet up out there. It's the best spot I've found so far to fly without 85 people asking you how much it costs and where to buy it.

 

[Xrover]

I agree ianwood.

PM sent re: Testing.

 

[lightingbryan]

Any updates or new thoughts about this?

 

[alankkwong]

Just update your Phantom to 3.06 and this will solve the problem. I did it on mine and its settled!

 

[marksrig]

Well, Naza users that have the GPS and compass in one unit have to rotate the puck to adjust for declination. It seems we have been given no such provision. Why there isn't a magnetic declination lookup table in the Naza that automatically compensates is beyond me. I am going to hunt down a non-magnetic Phillips head driver in my workshop tomorrow and shim the compass so that it is pointing 12 degrees to the right. Will report back tomorrow.

Hello Ian,
I got into the Phantom 1.1.1 and adjusted the puck under the lid to offset my 15 degrees here in Oregon. It was a little unconventional. I used 3m tape and only one screw to hold it in place. I covered it back up with the shielding. I also adjusted the compass by the 15 degrees as well. This resulted in no toilet bowl effect which was pretty severe at start up. Now TBE is completely gone at take off. Unfortunately, The J-Hooking to the left is still there. I was certain it would have disappeared but only goes away after flying about 4 minutes. I'm going to double the compass to 30 degrees to see if I can hook it to the right. I've read somewhere that it supposed to get this straight.
UPDATE: Just got back from trying out the 30 degree rotation on the Phantom with the 15 degrees still on the puck. Although it did reduce the length of the J, it didnt fix it or make it go to the right. Fresh out of ideas.

 

[ianwood]

Marksrig, The Phantom 2 got a firmware update with a fix a few months back. Not sure what if anything is planned for early Phantoms.

 

[marksrig]

Marksrig, The Phantom 2 got a firmware update with a fix a few months back. Not sure what if anything is planned for early Phantoms.

Ianwood,
I'm running 4.02 in the Phantom 1.1.1 I'm not sure if this is the upgrade you're referring to.
UPDATE: I was reading in another forum that to properly calibrate the compass, you need to switch it 6 times from GPS to MANUAL and have the controller in manual mode when calibrating. I tried this and it immediately got rid of my J-hook when i took off in GPS mode. No need to fly all around wasting time. I always had the MANUAL selection set to FAILSAFE from day 1 and never thought twice about it. Apparantly operator error again! I still have the compass shimmed to 15 degrees and still no TBE. DJI said they are no longer supporting the Phanom 1...firmware updates? Goodbye. So much for your customers