Fixing DJI's Compass Problem

[landonkk]

For those of you who are members at RCGroups (or would like to join), please see our thread here and add support. We can get more eyeballs there. http://www.rcgroups.com/forums/showthread.php?t=2175384

Ian, I will shoot some video of the issue this weekend. For some reason even though I am only at 12 degrees here the j-hook is very noticeable with my P2's. I will shoot identical video's with my 2 P2's & P2V to prove the point.

 

[obiwan_pierogi]

My new vision 2+ is arriving in the mail tomorrow.. (I couldn't wait for them to get the **** declination fixed).. I'm just wondering if I'll have to rotate the compass 30 degrees with this guy, as I had to with my vision 2..

 

[damoncooper]

Avatar on

 

[kfallsbob]

Today I just activated the IOC Course lock and home lock via the Phantom Assistant software. As soon as I started to fly, I noticed immediately what you are describing here on this forum. I did not notice it while using the standard Phantom 2 settings or at best, I did not notice it until I converted and then it was very noticeable.

I had never noticed the TBE until converting to NAZA mode.

After reading this, I will test again tomorrow morning with calm winds.

I am located in Southern Oregon and I believe our declination is 12 degrees.

Bob

 

[varmint]

T.

I am located in Southern Oregon and I believe our declination is 12 degrees.

Bob

First post in this topic. I'm in the same zone up here in Portland and can confirm that flying in NAZA mode with IOC off in my case, that I get a consistent left-hand flightpath (I was flying to the North). It occurs both on my Vision and Vision+ exactly the same, every time. They simply will not fly in a straight line, and it's pretty aggravating.

 

[Sky Surv]

Hi guys,

First post but pretty experienced with DJI from the P1 through to the S800's. Been operating Phantom 1's as part of my business now for over a year, not one issue. We use them on survey, and also for filming. We have also been operating one of our P1's minus a GPS module for 6 months after it failed, and never bothered to replace it as it was not worth it, still flys perfect in ATTI.

Just added 3x P2's to our mini UAS fleet, where we require the CL lock for flying grid surveys. Waste of time currently. Appear to work fine on ATTI mode. I am assuming everyone else is not having any issues with ATTI mode either? Seems to be the best option for flying down a road or something where you want to ensure it wont hook off.

We only discovered this issue yesterday being put onto the issue by Ian. Why on earth they can make the P1's perform perfectly (never had one issue, no fly away - running futaba TX) All our machines are calibrated on the naza assistant weekly, compass rotation before take off and never any funny business, until 'upgrading' which I use loosely, to the P2. Will be great when the sort it, the extra flight time is a massive help after our loaded up P1's are in the 4-8min region, but they ALWAYS fly straight.

Always been a big advocate of DJI, yet always weary of software updates and changes. As a commercial aeroplane pilot its always hard to put so much faith in a piece of software that could go haywire and hurt someone. Risk mitigation is key in the aviation world, and more so in the unmanned world.

We are also based in NZ at 22deg East. I am already looking to replace with 3DR gear, as its a business asset not a toy I can't risk something I don't have 100% faith in.

Add another 3 here, and a mates P2 did the same thing into a tree a few days back so that's 4.

 

[Captain Salty]

I'm in Queenstown, New Zealand @ 24 degrees.

I flew mine in GPS for the first time today (always fly ATTI) and sure enough TBE, J-Hook & Left drift. So +1 to the list please.

 

[gecarey]

Avatar installed !

 

[nzvideoguys]

Christchurch, New Zealand. Testing.

I just took the P2 out for some testing.
Running 3.02 firmware. Default gains.

Compass calibration done before each flight.

First flight, compass rotated 40deg clockwise.
Took off and hovered 30secs. No TBE. Stable.
Lined up forward direction on the top edge of a tennis court fence. Very little, if any drift
and very little hook.
Did several runs all with similar results.

Second flight, compass turned back to default position.
Took off, immediate TBE, moving anti clockwise in bigger and bigger circles.
Had to regain control.
Lined up forward direction on the top edge of a tennis court fence. Aprox drift
corresponding to 11am if forward is 12 o’clock, or 15degrees. Huge hook to the left at the end.
Ran up and down for about 6mins. Towards the end it was getting better.
Less drifting, less hook. No noticeable TBE.

I have all the video footage but is it worth uploading it? I mean, who’s going to look at it?

Had a strange thing happen on the first flight. I was viewing through the goggles
and the video switched off after about 11mins. Bought it down and discovered GoPro
had switched off. Wouldn’t turn on, and red light not coming on when powered by
gimbal. Took the GoPro out of the gimbal, removed and replaced battery.
Everything back to normal.
I think I’m going to make a point of charging the GoPro via the USB port before I do any flying.

 

[Sky Surv]

Christchurch, New Zealand. Testing.

I just took the P2 out for some testing.
Running 3.02 firmware. Default gains.

Compass calibration done before each flight.

First flight, compass rotated 40deg clockwise.
Took off and hovered 30secs. No TBE. Stable.
Lined up forward direction on the top edge of a tennis court fence. Very little, if any drift
and very little hook.
Did several runs all with similar results.

Second flight, compass turned back to default position.
Took off, immediate TBE, moving anti clockwise in bigger and bigger circles.
Had to regain control.
Lined up forward direction on the top edge of a tennis court fence. Aprox drift
corresponding to 11am if forward is 12 o’clock, or 15degrees. Huge hook to the left at the end.
Ran up and down for about 6mins. Towards the end it was getting better.
Less drifting, less hook. No noticeable TBE.

I have all the video footage but is it worth uploading it? I mean, who’s going to look at it?

Had a strange thing happen on the first flight. I was viewing through the goggles
and the video switched off after about 11mins. Bought it down and discovered GoPro
had switched off. Wouldn’t turn on, and red light not coming on when powered by
gimbal. Took the GoPro out of the gimbal, removed and replaced battery.
Everything back to normal.
I think I’m going to make a point of charging the GoPro via the USB port before I do any flying.

I am going to try this as well, and have three to test so certainly easy to identify a trend. You didn't get any hook in the opposite direction eventually like Ian mentioned?

One thing I have also done is turned the YAW gains down slightly as well, stops the thing turning so erratic. I think the P2 gains are too high out the box and seem to exacerbate any issues.

Also funny you say about the GP, mine did the same thing on a zenmuse. Maybe a power surge? Always charge normally prior I have found. Also I don't like how hot the GP gets on there, but that's a different topic

 

[SilentAV8R]

They simply will not fly in a straight line, and it's pretty aggravating.

You mean without any other input.

I have noticed that the faster you are moving when you let go of the right stick the more pronounced this effect is. Move slower and it is not as apparent. I did not notice this until I stopped flying mine and tried to simply push the one stick to go forward/backward. I was automatically correcting because that was what I do with any RC aircraft I am flying.

 

[ianwood]

Has anyone else noticed a difference between Phantom and Naza mode? I've used Phantom mode all of maybe 3 flights before I switched. Curious if anyone can confirm a difference or lack thereof.

I have noticed that the faster you are moving when you let go of the right stick the more pronounced this effect is. Move slower and it is not as apparent. I did not notice this until I stopped flying mine and tried to simply push the one stick to go forward/backward. I was automatically correcting because that was what I do with any RC aircraft I am flying.

The slower flight allows the FC more time to correct for course drift. In faster flight, the correction builds up and then unloads at the end. That's a gross over simplification but it kind of works like that.

 

[Andy T]

Hi guys,

First post but pretty experienced with DJI from the P1 through to the S800's. Been operating Phantom 1's as part of my business now for over a year, not one issue. We use them on survey, and also for filming. We have also been operating one of our P1's minus a GPS module for 6 months after it failed, and never bothered to replace it as it was not worth it, still flys perfect in ATTI.

Just added 3x P2's to our mini UAS fleet, where we require the CL lock for flying grid surveys. Waste of time currently. Appear to work fine on ATTI mode. I am assuming everyone else is not having any issues with ATTI mode either? Seems to be the best option for flying down a road or something where you want to ensure it wont hook off.

We only discovered this issue yesterday being put onto the issue by Ian. Why on earth they can make the P1's perform perfectly (never had one issue, no fly away - running futaba TX) All our machines are calibrated on the naza assistant weekly, compass rotation before take off and never any funny business, until 'upgrading' which I use loosely, to the P2. Will be great when the sort it, the extra flight time is a massive help after our loaded up P1's are in the 4-8min region, but they ALWAYS fly straight.

Always been a big advocate of DJI, yet always weary of software updates and changes. As a commercial aeroplane pilot its always hard to put so much faith in a piece of software that could go haywire and hurt someone. Risk mitigation is key in the aviation world, and more so in the unmanned world.

We are also based in NZ at 22deg East. I am already looking to replace with 3DR gear, as its a business asset not a toy I can't risk something I don't have 100% faith in.

Add another 3 here, and a mates P2 did the same thing into a tree a few days back so that's 4.

Mine still flies to the left in atti mode but but a big improvement over gps mode. Home lock and course lock are useless to me. Maybe there is something else wrong with my phantom but I'm not sending it back because their service is totally useless too. Cant imagine using my Phantom for any kind of commercial work.. too dangerous

 

[Andy T]

Has anyone else noticed a difference between Phantom and Naza mode? I've used Phantom mode all of maybe 3 flights before I switched. Curious if anyone can confirm a difference or lack thereof.

I have noticed that the faster you are moving when you let go of the right stick the more pronounced this effect is. Move slower and it is not as apparent. I did not notice this until I stopped flying mine and tried to simply push the one stick to go forward/backward. I was automatically correcting because that was what I do with any RC aircraft I am flying.

The slower flight allows the FC more time to correct for course drift. In faster flight, the correction builds up and then unloads at the end. That's a gross over simplification but it kind of works like that.

I had the same problems in NAZA and Phantom mode. no difference

 

[kfallsbob]

Update from yesterday's post:

Flew this morning with no wind (perfect conditions).

In NAZA, 10 seconds into liftoff, good hover no TBE. Advanced forward at 3/4 speed and got pronounced hook to the left while in GPS mode. Experimented with first battery and as stated, it got better with flight time but never a true straight line. Switched to ATTI and it flew straight as an arrow every time.

Second battery.....slight TBE on lift off, pronounced hover drift to the left.Distinct left hover drift diminished within 4 to 5 min.
All other tests identical to first battery. Always a left hook on forward acceleration.
Location, Southern Oregon

Yesterday I mentioned that I didn't notice it in Phantom mode, well I think the hook and drift was there but I was unconsciously correcting as I flew
Bob

 

[Dadcat]

I tried Phantom mode this weekend. J-hook and everything else are exactly the same as in NAZA.

I noticed a couple of other things:

1) The effect happens relative to the direction of travel. For example, if you apply full left aileron so that it's traveling sideways, it still hooks left (backward relative to the direction it's nose is pointed) when you release the stick. When I tried this, I automatically used right aileron to try and stop the drift instead of forward elevator and it went kind of wacky. I need to try it again someplace where a wrong control input won't put it into a tree. I guess this is consistant with what happens with reverse elevator.

2) The calibration and correction it figures out over the first 7 minutes or so works even if you fly in Atti mode the whole time. Apparently, it's still comparing the direction the GPS knows it's traveling with what the direction the compass says it's pointed even though it's not using the GPS to make course corrections. I suspect the accelerometer enters into the equation as well. When I tried this I still had about 10% hook remaining (upon switching to GPS mode) after 8 minutes. If I fly that time in GPS mode there is no hook at all by then. I think this was because there was about a 3 mile per hour wind. My P2V+ flies the direction it's pointed straight as an arrow in Atti, except for wind drift.

 

[d4ddyo]

I tried Phantom mode this weekend. J-hook and everything else are exactly the same as in NAZA.

I noticed a couple of other things:

1) The effect happens relative to the direction of travel. For example, if you apply full left aileron so that it's traveling sideways, it still hooks left (backward relative to the direction it's nose is pointed) when you release the stick. When I tried this, I automatically used right aileron to try and stop the drift instead of forward elevator and it went kind of wacky. I need to try it again someplace where a wrong control input won't put it into a tree. I guess this is consistant with what happens with reverse elevator.

2) The calibration and correction it figures out over the first 7 minutes or so works even if you fly in Atti mode the whole time. Apparently, it's still comparing the direction the GPS knows it's traveling with what the direction the compass says it's pointed even though it's not using the GPS to make course corrections. I suspect the accelerometer enters into the equation as well. When I tried this I still had about 10% hook remaining (upon switching to GPS mode) after 8 minutes. If I fly that time in GPS mode there is no hook at all by then. I think this was because there was about a 3 mile per hour wind. My P2V+ flies the direction it's pointed straight as arrow in Atti, except for wind drift.

I have a video where I was flying FPV ... I flew the phantom OUT... turned around and faced me... I can't remember if I was in ATT or GPS. Either way I thought I was more than half way into the flight and the declination was corrected.... low and behold i'm in the branches of a tree.. holy cow... I actually by sheer luck maneuvered through the branches unscathed.

 

[Sky Surv]

I think for the short term ours will just be operated at low speed, easier to maintain heading and less likely to fling off into a tree. I don't seem to be getting any major TBE which helps. Lack of IOC certainly is real pain in the but however.

I am also going to see if my gain adjustments help slow the pronounced hook turn when centering sticks. I am going to do some testing today with my P1 as a control, and the P2 with various settings. I'll let people know how I get on, buuut more than likely will be same old story.

SS

 

[d4ddyo]

I think for the short term ours will just be operated at low speed, easier to maintain heading and less likely to fling off into a tree. I don't seem to be getting any major TBE which helps. Lack of IOC certainly is real pain in the but however.

I am also going to see if my gain adjustments help slow the pronounced hook turn when centering sticks. I am going to do some testing today with my P1 as a control, and the P2 with various settings. I'll let people know how I get on, buuut more than likely will be same old story.

SS

Never mind testing. we are in Project Drunken Sparrow Now.

 

[N017RW]

Testing?... we don't need no stinkin' testing.

We're with Project Drunken Sparrow!