EdisonW1979 Posted at 2018-6-25 12:20
Sorry, but I've only collected all the complaints and issues and compiled them into one thread, a thread that you and your friend hallmark007 have littered with attacks against those who dare speak against your two opinions.

In this case yes, insufficient GPS locks may be the culprit, but as I stated, in no way should that prompt a drone to spontaneously want to kill itself! That is not misinformation, that is common sense observation.

Again post correct information, advice I gave was to ship drone back and some have as you will see in this thread , and now are flying happily. While others sit on this forum complaining but still having to wait. Why not hit dji where it hurts, their profits.
I still haven’t seen anyone being helped with your informational thread.

If you know exact spot (within a few feet) you placed your drone for TakeOff ---- that area can be checked for a source of Magnetic Field that could cause Compass Interference.
If you are interested in how, let me know.

HedgeTrimmer Posted at 2018-6-25 16:17
If you know exact spot (within a few feet) you placed your drone for TakeOff ---- that area can be checked for a source of Magnetic Field that could cause Compass Interference.
If you are interested in how, let me know.

Checking the compass error readings in the sensor section of DJI Go 4 will indicate if there is anything influencing the compass.

Aardvark Posted at 2018-6-25 16:48
Checking the compass error readings in the sensor section of DJI Go 4 will indicate if there is anything influencing the compass.

Better the Devil you know..

hallmark007 Posted at 2018-6-25 16:49
Better the Devil you know..

Déjà vu   

HedgeTrimmer Posted at 2018-6-25 16:17
If you know exact spot (within a few feet) you placed your drone for TakeOff ---- that area can be checked for a source of Magnetic Field that could cause Compass Interference.
If you are interested in how, let me know.

unless calibrated at the take off location which exhibited mag interference it shouldn't really matter that much i dont think.. unless its stupidly strong interference.  but once few feet in the air, things should and usually do return to normal or close enough..  

i take off from hood of my car while its running all the time, and im pegged out on both compasses for interference, but never negatively impacts my flight once its in the air...  


i do not recommend a complete RTH even in a situation where you take off from a known hotspot of interference, best to bring it back yourself if you can.

my point is a properly functional drone shouldn't get lost just because the take off spot if a hotspot of magnetic interference.   once it clears that it should be happy, a normal functional drone shouldn't then go crash itself into a river or fly forever away just because took off in a hotspot

HereForTheBeer Posted at 2018-6-25 16:53
unless calibrated at the take off location which exhibited mag interference it shouldn't really matter that much i dont think.. unless its stupidly strong interference.  but once few feet in the air, things should and usually do return to normal or close enough..  

i take off from hood of my car while its running all the time, and im pegged out on both compasses for interference, but never negatively impacts my flight once its in the air...  

I took off from the sidewalk in front of my house. I've done this multiple times from this spot and never had an issue. I actually use the empty lot across the street for practice. No high power lines are anywhere near the area.

I actually flew earlier that day in ATTI mode in the exact same spot and took a pano of the neighborhood.

HedgeTrimmer Posted at 2018-6-25 16:04
Thanks for information and reply.
Several more questions:  
What kind of surface/area did your drone takeoff from?

Took off from a level sidewalk with no metal or power lines nearby.

Aardvark Posted at 2018-6-25 16:48
Checking the compass error readings in the sensor section of DJI Go 4 will indicate if there is anything influencing the compass.

Like I showed here: Follow up small drones magnetic intereference sub-subsurface

Take note of first picture.  The R.C. shows ''Ready to Go'' despite detected magnetic interference.
Going back to Main Menu of GO-4 app also showed it was okay to fly, Green with no errors.

HereForTheBeer Posted at 2018-6-25 16:53
unless calibrated at the take off location which exhibited mag interference it shouldn't really matter that much i dont think.. unless its stupidly strong interference.  but once few feet in the air, things should and usually do return to normal or close enough..  

i take off from hood of my car while its running all the time, and im pegged out on both compasses for interference, but never negatively impacts my flight once its in the air...  

From testing I have done thus far, it takes a lot of magnetic interference to cause problems at takeoff.
For example: Building wall causing 15-degrees of Magnetic Interference at 36-inches or 0.91-meter to Handheld GPS's compass.

Strong sudden gusts of winds have been a bigger issue for me.  But give credit where credit is do, the Mavic Pro has thus far not crash.  Some hair raising moments during takeoff and landing.

HedgeTrimmer Posted at 2018-6-25 17:19
Like I showed here: Follow up small drones magnetic intereference sub-subsurface

Take note of first picture.  The R.C. shows ''Ready to Go'' despite detected magnetic interference.

In picture 1, compass one is yellow (good), and you will note that compass one is the active device (blue circle indicating it is in use) of the two. So compass 2 reading has no influence on the devices decision.
So what the device is telling you is correct.

Aardvark Posted at 2018-6-25 17:30
In picture 1, compass one is yellow (good), and you will note that compass one is the active device (blue circle indicating it is in use) of the two. So compass 2 reading has no influence on the devices decision.
So what the device is telling you is correct.

True, but far from being Excellent.  In that case, GO-4 App on main screen should have issued a warning.

Example of Mavic Pro picking wrong compass to use.  It goes with Yellow (Good) vs. Green (Excellent).

HedgeTrimmer Posted at 2018-6-25 17:30
From testing I have done thus far, it takes a lot of magnetic interference to cause problems at takeoff.
For example: Building wall causing 15-degrees of Magnetic Interference at 36-inches or 0.91-meter to Handheld GPS's compass.

all i know is what i seen and what i tried.. which is def more haphazardly than your testing.. mine is the hold my beer and watch this method of putting my drone at direct risk.  

but still my results have shown, once if gets out of the interference zone, its probably okay assuming have a properly functional drone with a calibration that isn't jacked up.

and no if im taking off from my car where basically a magnetic nightmare, mine doesnt tell me ready to go, it says "Mag interference" on remote. and red banner telling me my compass is messed up basically, but once got a few feet of distance its happy, even if yellow compass it does not freak out and throw itself into a wall or some trees..  seems effectively..normal

HedgeTrimmer Posted at 2018-6-25 17:39
True, but far from being Excellent.  In that case, GO-4 App on main screen should have issued a warning.

Example of Mavic Pro picking wrong compass to use.  It goes with Yellow (Good) vs. Green (Excellent).

Why is that the wrong compass ?

It must be border line between excellent & good, but if I read the values correctly then compass 1 has the lower error number measurement 213 as opposed to 232. In practice little difference, so compass one would be the logical one to use. The lower the error value the better.

j2thafree Posted at 2018-6-25 17:19
Took off from a level sidewalk with no metal or power lines nearby.

The sidewalk may or may not have steel rebar or steel mesh in it.
Really need to check that sidewalk area thoroughly before attempting another flight from same spot (after you get your drone repaired).

I know from testing I have done, you can position Mavic Pro over an area that has sufficient steel rebar to cause magnetic interference and either get a compass Error or message about compass needing calibration.  Rotate the Mavic Pro 90-degrees over same spot, and magnetic interference drops to where GO-4 shows Green good to go.

Aardvark Posted at 2018-6-25 17:49
Why is that the wrong compass ?

It must be border line between excellent & good, but if I read the values correctly then compass 1 has the lower error number measurement 213 as opposed to 232. In practice little difference, so compass one would be the logical one to use. The lower the error value the better.

Can't speak for numbers.  Only that Mavic Pro has identified Compass #2 to be Excellent, and Compass #1 to be Good, yet chose Compass #1 to use.   

Making me wonder if some of Compass related crashes are result of problems in choosing which compass to use, failing to warn Pilot about detected magnetic interference, and allowing takeoff when two compasses are in disagreement about level of interference.

HereForTheBeer Posted at 2018-6-25 17:47
all i know is what i seen and what i tried.. which is def more haphazardly than your testing.. mine is the hold my beer and watch this method of putting my drone at direct risk.  

but still my results have shown, once if gets out of the interference zone, its probably okay assuming have a properly functional drone with a calibration that isn't jacked up.

Because of testing I have done, what you are talking about has been running through my mind.
Problem is how to setup such tests without putting drone at risk.  I give you a for being bolder than me.

After doing quite a bit of research and reading up on magnetic interference and Toliet Bowl Effect, I am coming away with something that fits with your experience.

Theory:

It isn't merely magnetic interference that causes Toilet Bowl Effect.  It isn't simply clearing an area of magnetic intereference that results in compass error flagged crashes.
Instead, Toilet Bowl Effect and compass error crashes are result of compasses that are out of calibration or were poorly calibrated - being used to close to sources of magnetic interference.

HedgeTrimmer Posted at 2018-6-25 18:15
Because of testing I have done, what you are talking about has been running through my mind.
Problem is how to setup such tests without putting drone at risk.  I give you a  for being bolder than me.

i believe your theory is correct!  and it doesn't help that the app says to calibrate the compass....people will be like.. "oh... okay ill do it right here then"


can someoen verify my theory... that these drones for the most part atleast, are factory calibrated..?  because i never did a compass calibration until more recently, i nearly owned my mavic pro for a year and traveled around alot with it.

HedgeTrimmer Posted at 2018-6-25 17:59
Can't speak for numbers.  Only that Mavic Pro has identified Compass #2 to be Excellent, and Compass #1 to be Good, yet chose Compass #1 to use.   

Making me wonder if some of Compass related crashes are result of problems in choosing which compass to use, failing to warn Pilot about detected magnetic interference, and allowing takeoff when two compasses are in disagreement about level of interference.

Only one is active, it wouldn't be comparing the two. If the active compass shows an unacceptable error level, the redundant compass will become active. So they never disagree. Lets suppose that one compass were to fail completely, then the other takes over. That is the whole point of redundancy, Likewise the dual IMU, if one fails in operation the other will take over.

Aardvark Posted at 2018-6-25 18:21
Only one is active, it wouldn't be comparing the two. If the active compass shows an unacceptable error level, the redundant compass will become active. So they never disagree. Lets suppose that one compass were to fail completely, then the other takes over. That is the whole point of redundancy, Likewise the dual IMU, if one fails in operation the other will take over.

I know only one is active at a time.  You are missing the point.  

When one compass (magnetic field strength sensors for XYZ axis) is showing a real difference in magnetic fields, then one of compasses is either out of calibration or drone is being affected by a source of magnetic interference very near the drone.  

Either way, take off is a risk.   In case of source that close to drone, the drone needs to be moved.  In case of calibration, If something goes wrong with drone selected compass, the drone is left switching to out of calibration compass.

HereForTheBeer Posted at 2018-6-25 15:08
dji is worse than apple about admitting issues.. this makes it hard to figure out the real source of the issue on software or hardware difficult..   i do believe the mavic air has some fundamental flaw in it, either in firmware/software or in its hardware causing common issues.

yes you have a golden example mavic air, as does several others here, and that's awesome, i'm happy for you and others with golden examples.. similar story with my mavic pro, its effectively a golden example, barely any issues.. but just look around.. both platforms have their issues.  mavic pro has a NASTY history with some flaky models still in existence on shelves being purchased. mavic air has some flaw in it causing problems as well.

The main point for me is that it does not help anyone to generalise issues. Each case needs to be looked at individually. I dare to say that my Mavic would have behaved exactly like the one in this particular case and the reasons for that are quite obvious. There is no indication for a faulty device or firmware here.

HedgeTrimmer Posted at 2018-6-25 19:41
I know only one is active at a time.  You are missing the point.  

When one compass (magnetic field strength sensors for XYZ axis) is showing a real difference in magnetic fields, then one of compasses is either out of calibration or drone is being affected by a source of magnetic interference very near the drone.  

As you say, in your experiment, as a precaution it would be sensible to move aircraft to a better place. Which is exactly what the instruction given by the system would be if significant interference were detected.

As to the OPs incident, it doesn't look like the compass had any bearing on the outcome.

Aardvark Posted at 2018-6-26 01:12
As you say, in your experiment, as a precaution it would be sensible to move aircraft to a better place. Which is exactly what the instruction given by the system would be if significant interference were detected.

As to the OPs incident, it doesn't look like the compass had any bearing on the outcome.

Had seen this posted, along with others discussing Compass.

Aardvark Posted at 2018-6-26 01:12
As you say, in your experiment, as a precaution it would be sensible to move aircraft to a better place. Which is exactly what the instruction given by the system would be if significant interference were detected.

As to the OPs incident, it doesn't look like the compass had any bearing on the outcome.

Which is exactly what the instruction given by the system would be if significant interference were detected.

Except it isn't warning with significant interference.

DJI and others here have repeatedly pointed to magnetic interference as cause of Crashes and in some cases Pilot Error, thus denying any warranty claim.  

Wachtberger Posted at 2018-6-25 21:44
The main point for me is that it does not help anyone to generalise issues. Each case needs to be looked at individually. I dare to say that my Mavic would have behaved exactly like the one in this particular case and the reasons for that are quite obvious. There is no indication for a faulty device or firmware here.

i'm not sure about that..  im willing to bet that there is some hidden issue that logs arent showing/obtaining properly either.  

good example of logs not quiet being in tune with reality:

  sometimes my mavic pro shows a tiny bit of drift, IMU/GPS speed is 0.00 even if its correcting its drift, no speed indicated on IMUs or GPS.. however the mavic pro sees the drift and correcting for it.. not just VPS doing that either.  i disable visual system and it correcting for its drift here and there but always 0 speed.   not sure if that's normal or not, but i would think it some speed getting logged if it has to correct for it, specially with VPS turned off.  

another example is my first DJi spark, not sure why my stick inputs got logged, when the drone completely ignored it during some of its random moments of not listening to its owner.  basically if logs were anything to go by, i was in control of it but in reality it was drifting away wildly on its on even if my only command was forward gently,  it shot sideways...  even if thats atti mode why would it shot sideways with forward command?  that's like it has its brain on sideways or something..

HereForTheBeer Posted at 2018-6-26 20:08
i'm not sure about that..  im willing to bet that there is some hidden issue that logs arent showing/obtaining properly either.  

good example of logs not quiet being in tune with reality:

VPS won’t ever correct drift that’s not its purpose, IMU will try to correct but only on instruction from controller.
Best way you can show this drifting is by putting up your logs, or sending dji your .dat file, wind will also be a factor with drifting , VPS can also cause drifting if light is bad and your flying low.

Vision Positioning System (VPS)

The system that gives the Phantom 4 its ability to see consists of two front-facing and two downward-looking optical sensors, coupled with two dedicated image processors. The system extracts 3D information from the digital images captured by its sensors through stereo vision, or stereopsis as it is called when the human brain does it, comparing the information about a scene from two vantage points and using the difference in position for each object in the two 2D images to triangulate their real-world locations.

By working closely together with its satellite positioning system, the aircraft is able to locate itself within this map, with an accuracy of +/- 0.1 m vertically and +/- 0.3 m horizontally. This way, the aircraft is able to maintain and re-adjust its hover position even if it is pushed by an external force or carried away by strong winds. The VPS also increases hover accuracy and makes it possible to fly indoors where no GPS signal is available.

https://www.dji.com/newsroom/news/inside-a-drone-computer-vision

HedgeTrimmer Posted at 2018-6-26 23:47
Vision Positioning System (VPS)

The system that gives the Phantom 4 its ability to see consists of two front-facing and two downward-looking optical sensors, coupled with two dedicated image processors. The system extracts 3D information from the digital images captured by its sensors through stereo vision, or stereopsis as it is called when the human brain does it, comparing the information about a scene from two vantage points and using the difference in position for each object in the two 2D images to triangulate their real-world locations.

You should read what the poster wrote, VPS cannot correct drifting, it can stop a craft drifting but needs conditions to be right, the poster aircraft would not be corrected but drift can be avoided by a good working VPs , what you quote is 3D , poster flies a Mavic Pro not a phantom Pro .

hallmark007 Posted at 2018-6-25 15:57
I have left video below, it explains why you don’t always get a warning on the ground.

https://youtu.be/ZL6NbpzJVJI

hallmark007 This is a great video. I saw all 3 and they are very informative.

kchevy Posted at 2018-6-27 04:44
hallmark007 This is a great video. I saw all 3 and they are very informative.

Great they offer a very detailed explanation of how these aircraft work and explained in very simple terms.

GPS did not have a srtong enough signal. almost happened to me in Hawaii. Thought 7 was enough, now wont go above vision sensors capability without 10 bars

hallmark007 Posted at 2018-6-27 00:48
You should read what the poster wrote, VPS cannot correct drifting, it can stop a craft drifting but needs conditions to be right, the poster aircraft would not be corrected but drift can be avoided by a good working VPs , what you quote is 3D , poster flies a Mavic Pro not a phantom Pro .

Description of how VPS works was best described by one accompanying Phantom.  That is why I posted the description.


VPS by DJI --> This way, the aircraft is able to maintain and re-adjust its  hover position even if it is pushed by an external force or carried away  by strong winds.

HedgeTrimmer Posted at 2018-6-25 17:39
True, but far from being Excellent.  In that case, GO-4 App on main screen should have issued a warning.

Example of Mavic Pro picking wrong compass to use.  It goes with Yellow (Good) vs. Green (Excellent).

Ahh, my eyes!

HedgeTrimmer Posted at 2018-6-27 07:56
Description of how VPS works was best described by one accompanying Phantom.  That is why I posted the description.

That’s for 3D which MavAir phantom 4 Pro and inspire use, read your manual.

HedgeTrimmer Posted at 2018-6-27 07:56
Description of how VPS works was best described by one accompanying Phantom.  That is why I posted the description.

That’s for 3D system, MavAir, P4Pro, inspire, not Mavic Pro, read your manual it tells you there regarding VPS for Mavic Pro.

hallmark007 Posted at 2018-6-27 08:16
That’s for 3D system, MavAir, P4Pro, inspire, not Mavic Pro, read your manual it tells you there regarding VPS for Mavic Pro.

You mean like this:
The down- ward Vision System uses ultrasound and image data to help the aircraft maintain its current position. With the help of the Downward Vision System, your Mavic Pro can hover in place more precisely and  y indoors or in other environments where a GPS signal is not available.

From Page 24 of Mavic Pro User Manual.

It's not all bad news @j2thafree , you appear to be the holder of a Guinness World record! If my fingers and toes are correct your Mavic travelled over 34 million feet in 2 minutes which equates to over 250 times the speed of sound or Mach 252. Awesome.

HedgeTrimmer Posted at 2018-6-27 07:56
Description of how VPS works was best described by one accompanying Phantom.  That is why I posted the description.

They are quite different systems read what 3D does, as I explained to HFTB VPs won’t correct your drone if it drifts while flying, yes it does help with hovering both indoors and below 30 ft or somewhere around that , but it doesn’t correct when flying if you Aircraft has already drifted.

HedgeTrimmer Posted at 2018-6-27 08:27
You mean like this:
The down- ward Vision System uses ultrasound and image data to help the aircraft maintain its current position. With the help of the Downward Vision System, your Mavic Pro can hover in place more precisely and  y indoors or in other environments where a GPS signal is not available.

Look you put up what is a VPS system for a P4Pro , which it different from Mavic Pro, I was answering HFTB regarding system for Mavic Pro , and explained correctly what VPS does on Mavic Pro and it’s not the same as P4Pro .

hallmark007 Posted at 2018-6-27 09:49
Look you put up what is a VPS system for a P4Pro , which it different from Mavic Pro, I was answering HFTB regarding system for Mavic Pro , and explained correctly what VPS does on Mavic Pro and it’s not the same as P4Pro .

Please post evidence from DJI that shows VPS for Mavic Pro is overall different from Mavic Air, Phantom 4 Pro, Inspire, etc.  

I don't mean difference's between circuit board or chips set.  I am talking about VPS as DJI uses term to refer to it's Visual Positioning System.