Lost Mavic after loss of connection

[deleted member 877]

1. DJI will blame this on pilot error.
2. MP2 or MP, I have had an issue with the RC loosing connection while watching the drone from no less than 20 meters. RTH actually worked, but most definitely not to the settings i had set. The drone flew straight to the location and landed. If there would have been any obstruction the drone would have hit it. RTH in that instance may or may not have worked.
3. If you have line of sight and at 600 feet you lose connection then it is possibly a dud firmware update.

I have one thing to say about DJI support. They will not take accountability, they will blame the pilots. They are not interested in incurring costs by replacing defective drones. I would like to see stats of actual warranty repairs over and above placing the blame on us. As for the missing drone, the evidence is gone, even proof of RC failure will not get them to admit they have quality issues.

Here's hoping we get a decent drone manufacturer to fill the bad quality void so we kick DJIs arse back into the void

DJI is just a third rate drone manufacturer with market share. Like Nokia of old, who got kicked in the arse and where are they now... DJIs turn is coming, not if but when!

GOODBYE DJI

I believe that DJi is not unlike a lot of Tech innovators in that they tend to release products that may not have been thoroughly tested. The early adopters (i.e. customers) who purchase the latest gear unknowingly but effectively become Testing resources who provide free feedback to DJi support. As a software developer, I do the same thing but make it clear to my customers that there is always a degree of risk associated with installing the latest patch or incremental release. An in-house Testing regime can never replicate all of the various customer environments - it's simply not possible.

DJi is in a difficult position due to the potential for damage to or loss of drones as a result of bugs in the software. I think it's great that they offer insurance (DJi Refresh) and replace or repair damaged drones that can be recovered. In the case of irretrievable loss (e.g water landing) they probably should accept cached video as valid evidence along with the flight log etc. It's a dilemma to which there is no easy solution - you can't argue against DJi wanting to protect itself against unscrupulous individuals who would commit fraud.

A lot of people are justifiably unhappy with DJi releasing the Mavic 2 drones that did not include some of the fundamental capabilities of the superseded product such as precision landing and waypoints to name just a couple. Furthermore, DJI did not make this clear in any of the marketing material prior to or when releasing the Mavic 2 - it portrays a certain arrogance that says "Hey, we're the market leaders - like it or lump it".

I took delivery of a Mavic Pro a couple of weeks before the Mavic 2 was released because I knew that I would be buying a mature product with most if not all of the bugs ironed out and was also able to get a $200 discount. I intend to give DJi 12 months or so to perfect the Mavic 2 - at that stage I will most probably buy one.

 

[sar104]

If the commpas is calibrated 150 km from new takeoff site, that can couse the problem for the drone to back on HP! But, From experience I can say that not even 300 kn per line North/South is no problem but on the East/West route is a big problem.

Firstly - no - the FC calculates declination based on location - it doesn't measure it via the calibration process.

Secondly - no - I and others already analyzed the flight log, and there were no compass problems on that flight.

 

[Meta4]

If the commpas is calibrated 150 km from new takeoff site, that can couse the problem for the drone to back on HP! But, From experience I can say that not even 300 kn per line North/South is no problem but on the East/West route is a big problem.

The location and distance from your last compass calibration is completely irrelevant.
You could travel 5000 miles east/west or any direction and it will still make no difference.

 

[Meta4]

Thanks for your comments. The comments below are not strictly directed at you but to all those that are so sure they are correct without providing any real logic or reference material. The "you are all wrong the manufactures are wrong the world is wrong but I am right". Gang. Some of the comments are directed at your statements as I felt they need to be clarified. I don't like the reference you made that I am propagating misinformation when all I do is state fact based references. And that Meta4 is OK when all he does is post that anyone that disagrees with him is stupid. And provides no references only opinion.

I keep telling myself that I will ignore all the calibration misinformation but I think of all the members that are being confused. And really ....really. Will somebody post some reference based facts not allude to "Literature" or at least state some of their logic.

It's ironic how much misinformation you put out there after spraying anyone that has said different from the myths you cherish.

You really need to get an understanding of what compass calibration actually does and then you would understand when and why it might be required.

ps .. I have much, much more experience flying drones than you imagine.

 

[Dave A]

I'm no expert but I do understand the discussion on compass calibration and sar104 is 100% correct.

 

[BJR981S]

It's ironic how much misinformation you put out there after spraying anyone that has said different from the myths you cherish.

You really need to get an understanding of what compass calibration actually does and then you would understand when and why it might be required.

ps .. I have much, much more experience flying drones than you imagine.

Maybe you do. I have provided many times reference material here. Where is the reference material from you? I would like to read it so I can be as educated as you.

Please provide some level of confirmation of your views.

 

[Manicha]

Maybe you do. I have provided many times reference material here. Where is the reference material from you? I would like to read it so I can be as educated as you.

Please provide some level of confirmation of your views.

Can’t we just all get along and just stay confused as hell! I personally just calibrate compass when it asks me too and many times it doesn’t and I just fly when it says ‘READY TO FLY’ does it get anymore complicated?

 

[Manicha]

This isn’t just a signal loss, or the Mavic would RTH this is a sudden total power loss, as if the battery was there one minute, gone the next.

If this is in the UK, there is a lot going on over there in terms of research on equipment to intercept, shut down, overtake, retrieve, destroy etc....of Drones as a result of too many serious incidents/near misses with aircraft/Helicopters close to Airports. Maybe you’ve been a Guiné pig of a research test of some new anti Drone system.

 

[Manicha]

I’ve always felt, people miss handling, breaking rules, recklessly flying their drones when or where they shouldn’t be will band this hobby all together. Some Countries have already band the use of Drones all together. Some only by permission after a request made and even so, no pictures or video allowed while flying with the permission granted.

 

[DanMan32]

Battery on M2:
I tested the latches. First, only if both are disengaged does the app warn you about it.
Second: one of the latches on one of my three batteries doesn't stick out quite as far and spring not as strong. I was able to pull that side of the battery out of latch. I'm not sure but I think the battery was partially out, even on the other side after a hard flight (windy, high speed and turns).
Power is still engaged even when the battery is partially out, so that's good.
So for the M2, make very sure battery is completely in, and after flight, check that it is fully seated. If several of us see the batteries unlatching, DJI needs to know about it.

 

[Meta4]

I have provided many times reference material here. Where is the reference material from you? I would like to read it so I can be as educated as you.

Please provide some level of confirmation of your views.

SAR and BudWalker have eloquently explained already.
Reiterating would serve no purpose.
I'm unaware of a simple reference to quote.
The details of how it all works has been worked out co-operatively by a number of users over on PhantomPilots over a number of years.
I have been running a P4 pro in hard professional use for just a few weeks short of two years now without ever calibrating anything.
In that time I have traveled 3000 miles east/west and 1500 miles north/south.
That drone still runs as well as it did the day I took it out of the box and entirely without a glitch.

 

[BJR981S]

Thanks for going to the trouble of detailing a response. I appreciate that. I think we have agreed on most items now. I will take some of your comments onboard. But not all.

Deviation and declination are the same thing - no problem with you using deviation. But I'm confused - you asked for a link to literature showing that declination is computed by a global model, and then you provided the link yourself. Declination is not measured - it is calculated.

Can you please provide the link to the literature you mentioned? I would like to read the reference. Your statement was that it used a "global magnetic model" You now agree with me that it's is done by location but still refer to this model.

I would like to read that reference if it has any weight to our discussion.

It's certainly possible that the magnetic state of the aircraft might change over time, although the alternating field produced by the motor currents is unlikely to be the cause since alternating fields don't induce ferromagnetism and are actually the primary method for demagnetizing ferromagnetic materials.

Your description is of an AC motor that uses rings. These are brushless motors. There is no AC as such, AC means that it crosses the X axis from Positive to Negative. These use a Square Wave (Pseudo Sine Wave {+X Access offset for the P4P II, Mavic Platinum, Mavic II} ) pulse to magnetise the poles. So the magnetic field is expanding and contracting in one direction only. The positive x access. There is no -X axis field to erase the +ve field. The motor uses the pulse to magnetise in the same direction for each pole in the motor. At least that is my understanding. Please correct me if you think I am wrong.

Exactly. The compass is used to determine magnetic north. True north is offset from that by the location-dependent declination value. And the declination is calculated (not measured) based on your location from the global magnetic model in the firmware. You are now saying exactly what I said that you claim to disagree with.

It's the Global magnetic model that I disagree with. There is a global magnetic model of earths magnetic field but it is not used to calculate declination. Having this model in the drone would be pointless. There is a calculation done on location based on the location of the North Magnetic Pole and its reference to the True North. Plus provide a reference for the Global Magnetic Model that is in the FW.

Well no - the problem seems to be that you don't understand anything more recent than analog compasses. And, as I mentioned, there are no cases that I'm aware of where a change in declination due to a change in location has caused any problems at all, let alone uncontrolled flight or a fly away. Perhaps you could point to an example of such a case if you know of any.

Technology may change but Physics does not. I have over 100 drones from different manufacturers. I am not biased to DJI nor what I read in DJI flight records.

I will however do some additional research into the magnetometer used in the DJI drone. Do you have a reference to the actual HW unit used.

The Walkera drones that I own require calibration in all threee axis. Not just the 2 that are used in the DJI calibration routine. Same for my DJI Wookong and Naza H helicopters. DJI Helicopter FCs were available before the Phantom existed. I will see if I can find the DJI Wookong installation video. It showed how you needed to determine and install the Compass GPS module at the True North offset on the tailboom.

I will find the Drone flyaway for the traveller video. And post it here. As the drone was lost and it was a youtube video there is no flight record I can provide you.

Only one last question.

Why do you think DJI have introduced regular Compass calibrations in the M2s. I have given you are reasonable hypothesis. Do you have an alternate one?

And thanks again for your time and effort in this debate. I enjoy a good reasoned debate. I wish other posters in this discussion would do the same. At least state some reasons or reference material for their views. They add no value in just repeating I am right and you are wrong.

In a good reasoned debate both parties generally walk away with a better understanding of the subject.

I will not be able to continue this discussion for over a week. I am travelling interstate doing some technology strategy and transformation consulting to a Financial Institution in the morning. Have to feed my expensive hobby habit.


Cheers Brian

 

[BJR981S]

SAR and BudWalker have eloquently explained already.
Reiterating would serve no purpose.
I'm unaware of a simple reference to quote.
The details of how it all works has been worked out co-operatively by a number of users over on PhantomPilots over a number of years.
I have been running a P4 pro in hard professional use for just a few weeks short of two years now without ever calibrating anything.
In that time I have traveled 3000 miles east/west and 1500 miles north/south.
That drone still runs as well as it did the day I took it out of the box and entirely without a glitch.

Ok. this is what I referred to as "confirmation bias". You are assuming that this is the way things work based only on your experience and others that agee with you.

There is no reference that you can provide, there is no logic you can provide.

You assume that everyone, including the manufacturer and all the references are wrong because you haven't had a problem.

I consider this to an unreasonable attitude. This is what they call in religion "dogma".

I would expect more from someone here that professes to assist people in their understanding of this hobby.

 

[Meta4]

Ok. this is what I referred to as "confirmation bias". You are assuming that this is the way things work based only on your experience and others that agee with you.
There is no reference that you can provide, there is no logic you can provide.
You assume that everyone, including the manufacturer and all the references are wrong because you haven't had a problem.
I consider this to an unreasonable attitude. This is what they call in religion "dogma".

I would expect more from someone here that professes to assist people in their understanding of this hobby.

Sorry mate ... I don't think I can play by your rules.
We aren't all electronic engineers.
I'll never be able to give what you are asking for but don't see that makes my posting invalid.
My degree is in a different branch of science and I don't speak the same technical language.
I just have 30 years experience in navigation and using GPS for work and 5 years professional drone flying and 4 years involvement helping other drone flyers.
But what I see is you are participating in the dogma that you are so quick to accuse others of.
I don't assume the manufacturer is wrong.
I can see that DJI is hopeless at technical communication with users and their manuals are liberally sprinkled with inconsistencies, contradictions and errors.
Over five years, together with other intelligent users, what's true and what isn't has been worked out.
But just ignore me because I can't/won't supply the things you demand.

 

[BJR981S]

This also isn't correct. "Real Pilots" don't calibrate the compass during flight. They may set a new declination offset if required. But, this isn't anything like a compass calibration.

Please don't pick an argument based on a misunderstood lexicon.

It is called calibration. And indeed it is setting the offset.

I never said it was the same calibration as a drone. A bit hard to do a 2 axis rotation of a commercial aircraft in flight. I thought that would be obvious.

Just that it needed to be done as you travel a reasonable distances.

What was you point in your comment. What was it meant to achieve?

I give up. It is totally pointless to spend time with dogmatics.

Sar104 goes to the trouble of explaining his views I have respect for him.

Can anyone else here point me to any reference material other than personal views?

 

[BJR981S]

Sorry mate ... I don't think I can play by your rules.
We aren't all electronic engineers.
I'll never be able to give what you are asking for but don't see that makes my posting invalid.
My degree is in a different branch of science and I don't speak the same technical language.
I just have 30 years experience in navigation and using GPS for work and 5 years professional drone flying and 4 years involvement helping other drone flyers.
But what I see is you are participating in the dogma that you are so quick to accuse others of.
I don't assume the manufacturer is wrong.
I can see that DJI is hopeless at technical communication with users and their manuals are liberally sprinkled with inconsistencies, contradictions and errors.
Over five years, together with other intelligent users, what's true and what isn't has been worked out.
But just ignore me because I can't/won't supply the things you demand.

Thanks for detailing this to me. I do understand the position you take. I am just saying I disagree with it. You have a degree in another branch of science. Then you should understand that I am using the Scientific Methodology.

As I said to SAR104 healthy robust debate is good for all that are involved in the discussion.

Pitch an additional hypothesis to evaluate.

As you say there has been intelligent discussion in this Forum over quite a few years.

But that gets very stale if everyone that aggrees are the same vocal members. I get it, its a bit of an old boys club here.

In my line of work I find that a lot of organisations give statement like "This is the way we do things here" in a lot of cases this way of doing things was based on a false premise in the beginning.

You need to continually challenge your own way of thinking. Create a hypothesis based on sound logic and as much confirmed reference material (We all stand on the backs of others) as you can find. Then put that hypothesis to the test.

Some of the things you and others say may be correct. But I cannot determine the validity if it's not based on documented facts, or at least a solid hypothesis based on sound logic.

The reality of DJI now enforcing a regular compass calibration is a clear signal that you need to rethink your current position.

Wouldn't you agree?

 

[Meta4]

You need to continually challenge your own way of thinking. Create a hypothesis based on sound logic and as much confirmed reference material (We all stand on the backs of others) as you can find. Then put that hypothesis to the test.

That is how the understanding of the DJI compass has evolved over time.
It's been a gradual process that's been changed and refined to where it is now.

The reality of DJI now enforcing a regular compass calibration is a clear signal that you need to rethink your current position.

Wouldn't you agree?

Sorry but no.
Compare that to DJI's compass instructions for the P4 pro.
No mention of calibrating when new, or when changing locations or over time.

But DJI still couldn't get it completely right.
They leave users thinking that when they get a compass error (due to placing the drone almost on top of a steel or reinforced concrete surface) that they should recalibrate when the correct action is to move away from the issue the compass is warning of.

And they still can't tell users to avoid launching from steel or reinforced concrete surfaces - one of the most significant causes of lost/crashed drones.

DJI's documentation leaves a lot to be desired.

 

[BudWalker]

Please don't pick an argument based on a misunderstood lexicon.

It is called calibration. And indeed it is setting the offset.

I never said it was the same calibration as a drone. A bit hard to do a 2 axis rotation of a commercial aircraft in flight. I thought that would be obvious.

Just that it needed to be done as you travel a reasonable distances.

What was you point in your comment. What was it meant to achieve?

I give up. It is totally pointless to spend time with dogmatics.

Sar104 goes to the trouble of explaining his views I have respect for him.

Can anyone else here point me to any reference material other than personal views?

Sorry, this was a follow on to to the previous post that I had made. Just to be clear, declination isn't determined in the compass calibration process. It's determined before flight when enough GPS sats have been acquired to determine coords. This post - Mavic 2 Pro Flyaway - provides compelling evidence this is the case.

This has been pointed out to you twice now but you continue to ignore it. Maybe this is an example of confirmation bias that you have referred to.

 

[sar104]

Thanks for going to the trouble of detailing a response. I appreciate that. I think we have agreed on most items now. I will take some of your comments onboard. But not all.



Can you please provide the link to the literature you mentioned? I would like to read the reference. Your statement was that it used a "global magnetic model" You now agree with me that it's is done by location but still refer to this model.

I would like to read that reference if it has any weight to our discussion.

No - the argument was never whether declination varied by location - it was about your incorrect assertion that it was measured by the calibration process when in fact it is calculated by the FC from a global model.

Start here: NCEI Geomagnetic Calculators

Your description is of an AC motor that uses rings. These are brushless motors. There is no AC as such, AC means that it crosses the X axis from Positive to Negative. These use a Square Wave (Pseudo Sine Wave {+X Access offset for the P4P II, Mavic Platinum, Mavic II} ) pulse to magnetise the poles. So the magnetic field is expanding and contracting in one direction only. The positive x access. There is no -X axis field to erase the +ve field. The motor uses the pulse to magnetise in the same direction for each pole in the motor. At least that is my understanding. Please correct me if you think I am wrong.

There is a weak, permanent local magnetic field associated with the motors that does not change when they are running. It's simple to measure - you should try it. The bottom line is the motor currents do not appear to produce a significant steady field of any kind outside the motor, and that would explain why the magnetization state of the aircraft generally doesn't change with time.

It's the Global magnetic model that I disagree with. There is a global magnetic model of earths magnetic field but it is not used to calculate declination. Having this model in the drone would be pointless. There is a calculation done on location based on the location of the North Magnetic Pole and its reference to the True North. Plus provide a reference for the Global Magnetic Model that is in the FW.

This paragraph makes no sense at all. You are arguing that there is a model (if you know that why do you keep asking for references?) but that it is not used because it would be pointless. Why would it be pointless? - the declination at the flight location is exactly what the FC needs to navigate. You still haven't explained how you think the declination could be determined without a global magnetic model. Declination is the difference between true north and magnetic north. The FC knows where magnetic north is from the compass. How do you suppose that it knows where true north is relative to that, if not from independently knowing declination? I think you are simply being argumentative at this point - this issue is so fundamentally basic that I find it hard to accept that you still don't understand.

I'm not sure what's going to satisfy you in terms of "a reference for the Global Magnetic Model that is in the FW". That it is physically the only way to do this hasn't been enough. That the process is documented in the DAT file event stream log is not enough. I can tell you that back in 2014 when a number of us were beta testing firmware fixes for DJI in relation to the Phantom 2 j-hook problem, DJI confirmed that they were computing declination from a global model in the firmware, but I'm sure that's not going to convince you either.

Technology may change but Physics does not. I have over 100 drones from different manufacturers. I am not biased to DJI nor what I read in DJI flight records.

If you have ever read a flight log then I cannot understand how we can be having this discussion. And I know for certain, from this discussion, that physics is not your strong suit.

I will however do some additional research into the magnetometer used in the DJI drone. Do you have a reference to the actual HW unit used.

No - and it doesn't matter since they all work on the same principles. @Thunderdrones may be able to supply a part number.

The Walkera drones that I own require calibration in all threee axis. Not just the 2 that are used in the DJI calibration routine. Same for my DJI Wookong and Naza H helicopters. DJI Helicopter FCs were available before the Phantom existed. I will see if I can find the DJI Wookong installation video. It showed how you needed to determine and install the Compass GPS module at the True North offset on the tailboom.

Rotating the compass module to account for declination, similarly to setting declination on an analog compass, was common practice in older units that did not compute declination. DJI moved on from that 5 years ago. And, in any case, while an interesting observation, it doesn't remotely support your argument since the rotation was determined by looking up your local declination (calculated from the global model), not by any calibration process.

I will find the Drone flyaway for the traveller video. And post it here. As the drone was lost and it was a youtube video there is no flight record I can provide you.

I'm not even remotely interested in watching YouTube videos. YouTube is one of the biggest collections of junk science and misinformation that I have ever seen. Find me a flight log for a flyaway caused by a declination error.

Only one last question.

Why do you think DJI have introduced regular Compass calibrations in the M2s. I have given you are reasonable hypothesis. Do you have an alternate one?

Impossible to say, especially since they haven't changed the guidance for other models, which varies by model. They are certainly aware that calibration issues don't cause flyaways - when flyaways are due to compass problems they are always caused by local interference at takeoff. I've never even seen a case where DJI mentioned calibration as an issue in their response.

And thanks again for your time and effort in this debate. I enjoy a good reasoned debate. I wish other posters in this discussion would do the same. At least state some reasons or reference material for their views. They add no value in just repeating I am right and you are wrong.

In a good reasoned debate both parties generally walk away with a better understanding of the subject.

I will not be able to continue this discussion for over a week. I am travelling interstate doing some technology strategy and transformation consulting to a Financial Institution in the morning. Have to feed my expensive hobby habit.


Cheers Brian

The reason for the terse replies is that you can be quite infuriating to argue with - you just don't pay attention, and your accusations of confirmation bias when anyone references another post just look silly. It's tantamount to shouting "confirmation bias" when I cite any reference. You may be surprised to know that most of us have better things to do than explain, over and over again, the most fundamental concepts of solid-state magnetometer use.

So here's some more confirmation bias for you - I'll reference myself just in case this helps:

Mavic compass needs to be calibrated before each flight

 

[DanMan32]

I believe the motors are sent an AC waveform. Doing so provides control on 6 poles, not just 3. You can do that with brushless since each coil is controlled outside of the motor, but you can't with brush motors, as the brush contacts automatically reverse the current to the coils.