Fixing DJI's Compass Problem

[landonkk]

Ian - just curious if Terry and the guys at DJI are being responsive to your inquiries. I understand and agree that these things take time and they need to collect data, parse data and troubleshoot, but it has been some time now since they agreed to look at this. It doesn't seem like this is of any priority for them.

Anyway - keep up the good work. I think maybe we should start thinking of a 'deadline date' that we aim for before we collectively start making some 'noise'.

 

[onfourblades]

Ian - just curious if Terry and the guys at DJI are being responsive to your inquiries. I understand and agree that these things take time and they need to collect data, parse data and troubleshoot, but it has been some time now since they agreed to look at this. It doesn't seem like this is of any priority for them.

Anyway - keep up the good work. I think maybe we should start thinking of a 'deadline date' that we aim for before we collectively start making some 'noise'.

i second this.
Check Plus to Ian and the rest of the beta testers

 

[mcmax20]

Ian - just curious if Terry and the guys at DJI are being responsive to your inquiries. I understand and agree that these things take time and they need to collect data, parse data and troubleshoot, but it has been some time now since they agreed to look at this. It doesn't seem like this is of any priority for them.

Anyway - keep up the good work. I think maybe we should start thinking of a 'deadline date' that we aim for before we collectively start making some 'noise'.

i second this.
Check Plus to Ian and the rest of the beta testers

+1 to the deadline as well. Anyone trying to create quality video in any areas of high declination are having a very hard time. DJI should look at this as an opportunity to not only keep current customers but also attract new ones. I am not at all able to recommend DJI products with this ongoing problem and quite honestly would not buy again if this doesn't get fixed. DJI needs to switch gears and get this resolved, put whatever new products they're looking to release on hold and spend the time needed to get their current product line to function like it should. Simple economics really, else they wont have a market to sell to; in my geographic area I have enough pull that I'll make certain of that! Lot of other brands out there...

Great job Ian and other testers! The rest of us however are sick of DJI's "buying time" tactics and need some real results and useful firmware. Not just "more capable of larger payloads" and slower decent rate updates, what kind of crap is that? I'm stuck using firmware, while at least reliable, isn't even able to run my GS. Nevermind the fact that when I did use it with firmware 2.0 it would fly straight ahead but in a cockeyed angle resulting in totally useless video. How in the world did that get passed through DJI's "testing"? The longer they take to resolve this issue the more bad press they'll get which will be hard to recover from. I'd hate to see that happen to a company that has been so innovative but if they refuse to address these problems in a timely manner they're asking for it.

JM(and several others)O

 

[ianwood]

I understand everyone's frustration. I share it. But we gave them the last set of test data last week. 3 batches of tests have been done so far. All things considered, this might not be their top priority but from what I can tell, they're taking it seriously.

If I get the idea that they aren't taking it seriously, I am going to raise a stink. And should all else fail, there's the nuclear option where we all go to social media in a series of coordinated and concentrated tactical media blasts. But, as long as I can see that they are making progress, I am going to do my best to be patient.

 

[tom3holer]

I understand everyone's frustration. I share it. But we gave them the last set of test data last week. 3 batches of tests have been done so far. All things considered, this might not be their top priority but from what I can tell, they're taking it seriously.

If I get the idea that they aren't taking it seriously, I am going to raise a stink. And should all else fail, there's the nuclear option where we all go to social media in a series of coordinated and concentrated tactical media blasts. But, as long as I can see that they are making progress, I am going to do my best to be patient.

I live south of Boston and the Dec is around 14.5deg and have reported I have TB big time.

Now LA's dec is around 12.5deg. and many have reported they have the issue also.

What I don't understand is why, with an LA base for DJI they can't do their own data gathering.
It is pretty pathetic that accompany that has had as many problems as they have had that they can't find the time and rescourses to diagnose the issue inhouse, but rely on us as users to do their work.

 

[NYRangers99]

I have to agree tom…I downgraded from 3.0 to 1.08 because of jhook and TB…now with 1.08 I am getting drifting to the point that if your trying to hover to get good pics…was trying to get some pics of a gator in Florida…my phantom 2 was drifting backwards to the point where it almost crashed into rocks, I am constantly fighting to keep it steady especially towards the beginning of the flight…compass was calibrated multiple times and controller sticks were calibrated also…I love my Phantom 2 but something seriously needs to be done because this drifting problem, jhook and TB is bad!
Its much appreciated that Ianwood and all of you other guys are taking the time to send data in to get this problem resolved but DJI needs to step it up big time!

 

[Sanababit]

I understand everyone's frustration. I share it. But we gave them the last set of test data last week. 3 batches of tests have been done so far. All things considered, this might not be their top priority but from what I can tell, they're taking it seriously.

If I get the idea that they aren't taking it seriously, I am going to raise a stink. And should all else fail, there's the nuclear option where we all go to social media in a series of coordinated and concentrated tactical media blasts. But, as long as I can see that they are making progress, I am going to do my best to be patient.

I live south of Boston and the Dec is around 14.5deg and have reported I have TB big time.

Now LA's dec is around 12.5deg. and many have reported they have the issue also.

What I don't understand is why, with an LA base for DJI they can't do their own data gathering.
It is pretty pathetic that accompany that has had as many problems as they have had that they can't find the time and rescourses to diagnose the issue inhouse, but rely on us as users to do their work.

Well just a tought, maybe they are gattering info from many users from different parts so they can come up with a universal fix, if they only fix the DEC in LA then it might not work in iowa or new york, it could happen, lol, i hope they can come up with a fix soon, i want to fly but without 100% of control i am affraid to fly it high or far, come on DJI, you can do it

Sana

 

[onfourblades]

Well just a tought, maybe they are gattering info from many users from different parts so they can come up with a universal fix, if they only fix the DEC in LA then it might not work in iowa or new york, it could happen, lol, i hope they can come up with a fix soon, i want to fly but without 100% of control i am affraid to fly it high or far, come on DJI, you can do it

Sana

Im afraid to fly as well... Safe flight can not be achieved without a fix...DJI PHANTOM is grounded

 

[d4ddyo]

Well just a tought, maybe they are gattering info from many users from different parts so they can come up with a universal fix, if they only fix the DEC in LA then it might not work in iowa or new york, it could happen, lol, i hope they can come up with a fix soon, i want to fly but without 100% of control i am affraid to fly it high or far, come on DJI, you can do it

Sana

Im afraid to fly as well... Safe flight can not be achieved without a fix...DJI PHANTOM is grounded

Failsafe still works. Return to Home still works. ATT mode still works. GPS corrects itself half way into flight. From Launch... I hover with GPS mode for a minute and observe behavior. Switch to ATT mode. If I want to use home lock/cl lock it's usually already autocorrected.

Safe flight can be achieved. I DO want a fix ASAP... but I am still able to fly safely.

 

[landonkk]

What I don't understand is why, with an LA base for DJI they can't do their own data gathering.
It is pretty pathetic that accompany that has had as many problems as they have had that they can't find the time and rescourses to diagnose the issue inhouse, but rely on us as users to do their work.

The impression that I get is that LA is nothing more than a phone center where people have access to a database and a script and try to resolve low-level issues like "How do I calibrate the compass?"

 

[obiwan_pierogi]

Some of you may have seen my video, displaying the major toilet bowl effect I was experiencing, with the compass in the factory set position.. The declination here is about 17.5 degrees positive.. I've got my compass turned about 30 degrees clockwise, as viewed from the top.. I still experience an odd hook, usually left if I travel South, and right, if I turn the craft around, and head front first, in a Northerly direction.. However, I can live with that, until DJI comes up with a fix.. However, turning the compass has eliminated the toilet bowl effect.. Check out my video.. I did run it through youtubes stabilization, and added music, but as you can see: no toilet bowl effect.. And let me add: with the compass in the normal position, the craft is almost unflyable.. https://www.youtube.com/watch?v=rdBE5-WbVjA

By the way... Here is my video taken a month ago, with the compass in the factory set position.. The only control I put into the copter was turning it on its axis.. it then went into wider and faster circles.. **** near wiped out.. https://www.youtube.com/watch?v=VJBcAhd ... e=youtu.be

 

[ianwood]

Let's put an end to suggestions of rotating the compass. We all know and understand it might alleviate some of the more extreme cases but it is a bad workaround. The more I learn about how the compass works, the more I realize it should be in a known orientation. Moving it on any axis throws off the math regardless of calibration. There is only one fix for this problem and that involves DJI and their firmware. We don't want them thinking we'll accept a bogus workaround because we won't.

How and why this is acceptable on the more expensive DJI platforms is beyond me!

 

[Sanababit]

Let's put an end to suggestions of rotating the compass. We all know and understand it might alleviate some of the more extreme cases but it is a bad workaround. The more I learn about how the compass works, the more I realize it should be in a known orientation. Moving it on any axis throws off the math regardless of calibration. There is only one fix for this problem and that involves DJI and their firmware. We don't want them thinking we'll accept a bogus workaround because we won't.

How and why this is acceptable on the more expensive DJI platforms is beyond me!

Well said sir, come on DJI

Sana

 

[nzvideoguys]

Let's put an end to suggestions of rotating the compass. We all know and understand it might alleviate some of the more extreme cases but it is a bad workaround. The more I learn about how the compass works, the more I realize it should be in a known orientation. Moving it on any axis throws off the math regardless of calibration. There is only one fix for this problem and that involves DJI and their firmware. We don't want them thinking we'll accept a bogus workaround because we won't.

How and why this is acceptable on the more expensive DJI platforms is beyond me!

There are two issue at hand. One is that DJI have messed up the firmware causing all sorts of unpredictable flight characteristics. The other issue is, that many unsuspecting customers have bought these products, and finding them to be, in some cases, unflyable, there is a need to find a temporary solution, so that the product can be used, and not gather dust on a shelf in the garage. If rotating the compass is what it takes for me to go out and produce films like this: https://vimeo.com/94486996 then that is far better for me, than sitting and twiddling my thumbs while DJI get their **** together. Naturally, I am expecting that DJI will get their **** together and fix this problem, and then we can all get on with doing what we bought these products for in the first place.

 

[Bigbells]

Let's put an end to suggestions of rotating the compass. We all know and understand it might alleviate some of the more extreme cases but it is a bad workaround. The more I learn about how the compass works, the more I realize it should be in a known orientation. Moving it on any axis throws off the math regardless of calibration. There is only one fix for this problem and that involves DJI and their firmware. We don't want them thinking we'll accept a bogus workaround because we won't.

How and why this is acceptable on the more expensive DJI platforms is beyond me!

Where is the "like" button, lol? My personal solution is to stick with Assistant and firmware version 1.08, but it doesn't make much sense that the newer versions create more problems.

 

[xtonex]

I'm not sure why people are claiming the newer firmware makes this problem any worse... Buffalo NY and with 1.08, 2.0, 3.0, and 3.02 and on 2 different phantoms it all acts the same for me. TBE and Jhooks... no more or no less on any of the firmwares!

 

[tom3holer]

Let's put an end to suggestions of rotating the compass. We all know and understand it might alleviate some of the more extreme cases but it is a bad workaround. The more I learn about how the compass works, the more I realize it should be in a known orientation. Moving it on any axis throws off the math regardless of calibration. There is only one fix for this problem and that involves DJI and their firmware. We don't want them thinking we'll accept a bogus workaround because we won't.

How and why this is acceptable on the more expensive DJI platforms is beyond me!

I agree, it needs to be fixed.
What confuses me is the fact that the Naza-M requires you rotate the compass to compensate for dec and it works fine. I have done it on my Discovery pro and it eliminated TB and there was no change during the flight.

Tom

 

[obiwan_pierogi]

In my case.. It's either have the copter sit, because I'm afraid to fly it because of its unpredictable handling abilities, or rotate the compass, and fly it with confidence-- until DJI comes up with a fix.. Since I've rotated my compass, I've done numerous flights with nothing happening out of the ordinary.. I've also discovered that at this point in time, neither my dealer or DJI have been any help.. In fact, the video I just posted, showing the major toilet bowl effect, was after my copter was returned to me, from multirotorheli, in Calgary, where the owner claimed it was fixed, and working fine.. I'm just lucky it didn't end up in a million pieces, during my initial test flight.. I want to buy the vision 2+, in the worst way... But not until they get this thing fixed.. Once bitten, twice shy..

 

[Sanababit]

Well in my case i do have TBE and jhooking to the left, i live in tijuana mexico, but after 2-3 minutes it corrects itself and i fly normal, the problem are the firts couple of minutes, because i loose confidence in the craft, and the remaining time i fly with a lot of caution making my flight experience dull, lol, rotating the compass might be a temporary fix, but in my case if i move the compass in 2-3 minutes the craft might compensate or autocorrect itself and do the TBE or jhooking to the right, who knows, i am new at this, i dont do aerial footage i am just a rookie flyier, so for now i rather wait for a fix than telling DJI that rotating the compass works, because that takes pressure off of them and they will take their sweet time on fixing this..

Sana

 

[sar104]

Let's put an end to suggestions of rotating the compass. We all know and understand it might alleviate some of the more extreme cases but it is a bad workaround. The more I learn about how the compass works, the more I realize it should be in a known orientation. Moving it on any axis throws off the math regardless of calibration. There is only one fix for this problem and that involves DJI and their firmware. We don't want them thinking we'll accept a bogus workaround because we won't.

How and why this is acceptable on the more expensive DJI platforms is beyond me!

I agree, it needs to be fixed.
What confuses me is the fact that the Naza-M requires you rotate the compass to compensate for dec and it works fine. I have done it on my Discovery pro and it eliminated TB and there was no change during the flight.

Tom

There could be a number of reasons for the discrepancy. Firstly - if there is no software correction for declination and the FC doesn't expect the compass to be in a particular vertical orientation, then compass rotation is obviously a perfectly reasonable solution.

However, if, in the case of the Phantom, the FC does expect a particular orientation and is supposed to be doing a software correction (but getting it wrong), rotation could cause other problems.

For example, the calibration process involves rotating the compass through the magnetic field around two orthogonal axes, which I would guess are expected to be two of the principal axes of the compass sensor. If the sensor is displaced around its vertical axis then the second calibration step no longer uses a principal axis. One could tilt the Phantom (as has been suggested previously) to make the calibration step 2 around a principal axis, but if that is detected by the IMU (which it will be) then it may lead to other issues.

That is quite separate from the possibility (likelihood?) that the software correction is only partly correct and, as we discussed much earlier in this thread, that in part of the code the sign of the declination is reversed, leading to the observed condition that initial CL is corrected by rotating the compass by twice the local declination. But either way - if only part of the correction is wrong then a compass rotation will never correct the problem entirely.