Beta Testers requested to fix Compass Issue

[Xrover]

Yeah sar, i've wondered that also. Is anyone who is getting the uncommanded yaw running a 2D gimbal?
I notice that when you apply yaw stick input the 3D gimbal will also yaw when it detects your input. I think the rubber band reaction is it trying to re center when you ease up on the yaw stick. With the P2 just sitting on the ground and battery powered on, you can move the yaw stick and the gimbal will follow it. Or maybe it counters it, i don't remember. But i do know the gimbal will yaw in conjunction with your stick input. It makes shooting video very difficult due to the fact that if you are trying to do a straight in track shot and you are applying yaw input to keep your subject center frame, you can't determine if the camera or the P2 is actually yawing.

 

[mangoboom]

Okay I need to know if this issue is the whole compass problem.

Here's the scenario with video.

Headed down to the river yeserday morning at 6am as with zero winds (5mph) and bright sun thought ideal conditions.

Go through normal procedure, full GPS lock, launch and fly no problem. Running 3.04. After about 11 minutes in all fine.Good flight.

Running a Phantom 2 with an H3-3D gimbal - this is a new gimbal - I've had hours of flight time with the H3-2D this was about my 10th flight with the H3-3D.

Have also got mini-iosd and Fatshark FPV running.

So start to bring the Phantom back in. As I’m doing this there’s a large commercial ship passing along the river.

I figure it was maybe 600m away (see pic)..

so as I bring it in to me it doesn’t respond and just starts to wander off, you can see from the video this is the activity of it from when it starts to not take the input from the control. It sporadically lets me get control then wanders off again

I’m using a Futaba 14sg fully charged.

I’m in GPS mode the whole time , battery about 35%. You can see I almost get the chance to grab it then it just gets pushed away from me. I should have switched to Atti but I was just too worried about all the water (there’s a moat you can’t see) so I ended up ditching it in the brambles to not take any chances (broken prop only damage), plugged in post crash and all okay, no calibration needed. Compass MOD settings all normal.

My initial theory - some interference from the passing ship ???? Radar possibly affecting the TX ? As they can be in the same frequency ?

Magnetic field issue from the metal fence as the problem became pronounced after it flew close to the fence.

I'm going to head back down and take some magnetic readings to see if that is a potential issue.

https://www.youtube.com/watch?v=gwrhTan ... e=youtu.be

The more I read on this thread the more I'm thinking it is the overall 3.04 firmware /compass problem issue..

What do you guys think ?

 

[xtonex]

i think too many things get labelled a "fly away"... you sure you were in gps and not atti? seems like the wind was pushing it to me. was hardly out of control in my opinion

 

[mangoboom]

i think too many things get labelled a "fly away"... you sure you were in gps and not atti? seems like the wind was pushing it to me. was hardly out of control in my opinion

No, I checked my controls and was definitely not in ATTI. Wind was negligible even in ATTI . It may look controlled as I was trying to get a handle on what was happening so taking very slight movements of the control. I can tell you though it was definitely not taking my inputs. like I said with the right stick forward and right it should have been heading over towards me instead it was heading in the opposite direction I've done lots of flying to know it wasn't pilot error - this was definitely out of my control. Someone else mentioned this looks like the toilet bowl effect which points to a compass issue.

 

[Andy T]

Okay I need to know if this issue is the whole compass problem.

Here's the scenario with video.

Headed down to the river yeserday morning at 6am as with zero winds (5mph) and bright sun thought ideal conditions.

Go through normal procedure, full GPS lock, launch and fly no problem. Running 3.04. After about 11 minutes in all fine.Good flight.

Running a Phantom 2 with an H3-3D gimbal - this is a new gimbal - I've had hours of flight time with the H3-2D this was about my 10th flight with the H3-3D.

Have also got mini-iosd and Fatshark FPV running.

So start to bring the Phantom back in. As I’m doing this there’s a large commercial ship passing along the river.

I figure it was maybe 600m away (see pic)..

so as I bring it in to me it doesn’t respond and just starts to wander off, you can see from the video this is the activity of it from when it starts to not take the input from the control. It sporadically lets me get control then wanders off again

I’m using a Futaba 14sg fully charged.

I’m in GPS mode the whole time , battery about 35%. You can see I almost get the chance to grab it then it just gets pushed away from me. I should have switched to Atti but I was just too worried about all the water (there’s a moat you can’t see) so I ended up ditching it in the brambles to not take any chances (broken prop only damage), plugged in post crash and all okay, no calibration needed. Compass MOD settings all normal.

My initial theory - some interference from the passing ship ???? Radar possibly affecting the TX ? As they can be in the same frequency ?

Magnetic field issue from the metal fence as the problem became pronounced after it flew close to the fence.

I'm going to head back down and take some magnetic readings to see if that is a potential issue.

https://www.youtube.com/watch?v=gwrhTan ... e=youtu.be

The more I read on this thread the more I'm thinking it is the overall 3.04 firmware /compass problem issue..

What do you guys think ?

I once had mine in home lock not realizing it and that sure can mess with your mind! lol

 

[mangoboom]

I once had mine in home lock not realizing it and that sure can mess with your mind! lol

Nope definitely not that - I checked all the switches as I thought that was possibly the issue as well.

 

[Navman]

When I upgraded to 3.04 I didn't hit the default button, guess I thought the upgrade would automatically do it :shock:

Any link to what "default" settings Should be for the various firmwares?

I get gimbal yaw with the yaw stick on 3.04 too.

+ 16 degrees mag declination near Portland Oregon

 

[sar104]

I once had mine in home lock not realizing it and that sure can mess with your mind! lol

Nope definitely not that - I checked all the switches as I thought that was possibly the issue as well.

I'm pretty sure that it's not the declination problem - it doesn't manifest itself like that. It does look like wind-induced drift, but if the winds really were light then that suggests interference of some kind (magnetic or GPS) or just poor GPS coverage. Did you check that time and location for satellite coverage?

I also doubt that it was TX interference since that would presumably just lead to failsafe and hover / RTH. On that subject - you were clearly moving, so where was the home point?

 

[ianwood]

Okay I need to know if this issue is the whole compass problem.

Unrelated issues. See the first post so you can see for yourself. You simply got a taste of what it's like to live in NZ. You already have a thread open for it here: viewtopic.php?f=19&t=18491

So, let's continue to troubleshoot it there and leave this thread for the the beta firmware discussion.

 

[Andy T]

I once had mine in home lock not realizing it and that sure can mess with your mind! lol

Nope definitely not that - I checked all the switches as I thought that was possibly the issue as well.

I watched your video a couple of times and it sure does not seem to like that railing at all. There is probably a lot of steel in that concrete too.. just odd that it was fine for the first 11 minutes. Were you near those things then too?

 

[mangoboom]

I watched your video a couple of times and it sure does not seem to like that railing at all. There is probably a lot of steel in that concrete too.. just odd that it was fine for the first 11 minutes. Were you near those things then too?

No, knowhere near the railings for the main flight, as soon as I got near the railings that is when it started to play up. I'm going to return with a magnetic field detector app and see if that may be the issue.

 

[xtonex]

i think too many things get labelled a "fly away"... you sure you were in gps and not atti? seems like the wind was pushing it to me. was hardly out of control in my opinion

No, I checked my controls and was definitely not in ATTI. Wind was negligible even in ATTI . It may look controlled as I was trying to get a handle on what was happening so taking very slight movements of the control. I can tell you though it was definitely not taking my inputs. like I said with the right stick forward and right it should have been heading over towards me instead it was heading in the opposite direction I've done lots of flying to know it wasn't pilot error - this was definitely out of my control. Someone else mentioned this looks like the toilet bowl effect which points to a compass issue.

Is there any chance that it was going the opposite direction than you wanted because it was turned toward you?

 

[mangoboom]

Is there any chance that it was going the opposite direction than you wanted because it was turned toward you?

Nope definitely not.I've done a fair amount of FPV but yes I can see how it does look like that, the rest of the flight I did lots of flyarounds where the ship was facing me and was working fine but this was definitely not responding to my inputs.

 

[ianwood]

Is there any chance that it was going the opposite direction than you wanted because it was turned toward you?

Nope definitely not.I've done a fair amount of FPV but yes I can see how it does look like that, the rest of the flight I did lots of flyarounds where the ship was facing me and was working fine but this was definitely not responding to my inputs.

Take this discussion over here please: viewtopic.php?f=19&t=18491

 

[Andy T]

I watched your video a couple of times and it sure does not seem to like that railing at all. There is probably a lot of steel in that concrete too.. just odd that it was fine for the first 11 minutes. Were you near those things then too?

No, knowhere near the railings for the main flight, as soon as I got near the railings that is when it started to play up. I'm going to return with a magnetic field detector app and see if that may be the issue.

With my compass app if i walk within a couple of feet of my car you can see how much of an effect it has on a compass.. its huge

 

[Xrover]

I am wondering if the old gains are making my un-commanded yaw worse. And I am not sure, but I think the 3.05 beta has worse un-commanded yaw than 3.04. I don't remember it simply popping out like that so much. Really frustrating. Did a bunch of flights today and I would correct for the un-commanded yaw and it would keep going back like it was on a rubber band. It might be my imagination, but it seems worse even since I've put 3.05 on there. Probably not but I am noticing it a lot.

I will try updated gains next weekend before I head out of town. Hopefully, the new beta will drop by then.

The uncommanded yaw was probably there in 3.04 but since we had the major drifting issue on that one we might have been on the yaw stick try to keep the P2 flying straight. I do notice it more on 3.05! If 3.05 was tweaked by DJI, most likely it did not include anything that might now be confusing the 3D's correction performance. I guess a good test might be to adjust the tilt up gain setting for the 3D so it is always showing the booms. Then we might see if the booms yaw and what the gimbal does in relation.
I will try to test that today.

 

[sar104]

I took the opportunity of some very poor GPS satellite coverage this morning to try to investigate the effect on this issue. PDOP peaked around 3, with only 5 - 7 satellites available.

First runs, with 7, produced results not inconsistent with last time - reasonably straight flight immediately with just a little drift at the end. After power cycling, with 6 satellites, the track errors were significant, with noticeable J-hook. As the satellite count started to drop down to 5 the test turned into an observation of the hazards of a P2 transitioning in and out of GPS mode.

I hadn't ever tried flying in those conditions before. The aircraft was very twitchy, one moment flying fast in ATTI mode, and then apparently suddenly acquiring a position fix and making very abrupt corrections. Several times I resorted to switching to ATTI just to stabilize the flight. It's easy to see how people could completely lose it if trying to fly GPS mode with insufficient satellites. Remarkable how well the H3-3D coped, however.

I also did some brief tests on the effect of startup orientation, and saw no systematic effect, but that really needs to be repeated with more consistent GPS coverage. More later.

 

[justin]

Hi is there still beta tester place, i live in south america, suriname
and i have the same problem TBE and j hook
The setup i own is P2V+
Ow if you need footage i can make some ASAP.
Thanks.

Hi, Everyone,
This is Terry from DJI HQ. I've read our phantom users have compass issues like TBE, not flying in straight line in certain areas. This is very strange to us and engineers are very concerned about it. We believe it is related to magnetic interference locally. To find out the cause and solve the problem, I'd like to request for beta testers in this forum who:

1. own Phantom 2 or Phantom 2 Vision.
2. suffer compass issue and compass calibration does not help with latest 2.0 0 firmware.
3. have iOSD Mark II installed and provide flight data for analyse.
4. experienced in flying.

We will try our best to fix this issue and very appreciate your help as testers. If you are interested, pls reply the thread and I will contact you shortly. Thanks.

 

[sar104]

I'm still getting very random results. Another 3 test flights this evening with an average of 8 satellites, calm winds.

First battery, first couple of passes cleared up some initial track errors (j-hook left, maybe 20 - 30 ft), and CL (tried both GPS and ATTI mode) was less than 2 declinations off - probably less than one. Then it started to j-hook and drift to the right - a bit like has been observed with the compass rotation fix.

Second and third batteries produced very similar results. The lack of consistent behavior is really starting to annoy me - I just cannot see much of a pattern to this.

 

[ianwood]

This is really weird. I'm wondering what it could be and why it's different on yours. You're at 8,500ft, right? Maybe that's a factor somehow.

In other news: After not hearing anything for the past several days, I have word that the new beta is almost ready. They're doing final checks on it now. It apparently has "more intelligent function". Can't argue with that!

So beta testers, you know the drill. We all need to update, recalibrate, test, make a video, post results, debate, etc., etc. Let's be sure to include gains, other firmware versions, installed equipment, etc. in our observations.