Anatomy of a DJI crash (and probably clues for all that inspire flyaways)

[pahl]

Hi guys,

i was reading DIY Drones site and saw a really interesting post about another post at Phantom Pilots:

http://diydrones.com/profiles/blogs/anatomy-of-a-dji-flyaway-1

the original post:

http://www.phantompilots.com/threads/anatomy-of-a-dji-flyaway.42997/

Here the PDF with all the details:

http://www.ianwood.com/docs/anatomy-of-a-dji-flyaway-v1.pdf

i found out REALLY interesting how the root of the problem seems to be something different to the elements we all try to analyze, such as IMU and compass calibration/MOD Values, GPS fix, etc...

A theory is that the problems could come from a defective DJI controller (in that case the Naza inside the Phantom), so....... in other words.... could the source of all these flyaways be a poor quality control at the manufacturing point? (the article talks about the idea of bad solder points creating issues after the first few flights). Maybe DJI is having trouble to keep up with the manufacturing volume?

i own an inspire 1 too and i´m concerned about two main issues: the image quality/compression (but thats another post) and, of course, the reliability as a professional flying platform (and obviously, this is VERY important).

It seems that the guy who created that post got all that info by attaching a flight logger (CAN bus) to the Phantom. I don't know which one he used but i own a flytrex core for a F450 i also own and the excel table looks very similar.

Do anyone (with more electronic technical knowledge) have any thought to share?

Again, could the source of the problems be hardware based instead of firmware?

I love my inspire and never had a problem (besides one MC recoding error once) but i need to be sure that its a reliable platform for aerial filming. I already have an insurance (the laws in my country says you must have one if you are a RPAS operator) BUT i don't want to "use" it after a uncontrolled crash...

 

[The Editor]

Hi guys,

i was reading DIY Drones site and saw a really interesting post about another post at Phantom Pilots:

http://diydrones.com/profiles/blogs/anatomy-of-a-dji-flyaway-1

the original post:

http://www.phantompilots.com/threads/anatomy-of-a-dji-flyaway.42997/

Here the PDF with all the details:

http://www.ianwood.com/docs/anatomy-of-a-dji-flyaway-v1.pdf

i found out REALLY interesting how the root of the problem seems to be something different to the elements we all try to analyze, such as IMU and compass calibration/MOD Values, GPS fix, etc...

A theory is that the problems could come from a defective DJI controller (in that case the Naza inside the Phantom), so....... in other words.... could the source of all these flyaways be a poor quality control at the manufacturing point? (the article talks about the idea of bad solder points creating issues after the first few flights). Maybe DJI is having trouble to keep up with the manufacturing volume?

i own an inspire 1 too and i´m concerned about two main issues: the image quality/compression (but thats another post) and, of course, the reliability as a professional flying platform (and obviously, this is VERY important).

It seems that the guy who created that post got all that info by attaching a flight logger (CAN bus) to the Phantom. I don't know which one he used but i own a flytrex core for a F450 i also own and the excel table looks very similar.

Do anyone (with more electronic technical knowledge) have any thought to share?

Again, could the source of the problems be hardware based instead of firmware?

I love my inspire and never had a problem (besides one MC recoding error once) but i need to be sure that its a reliable platform for aerial filming. I already have an insurance (the laws in my country says you must have one if you are a RPAS operator) BUT i don't want to "use" it after a uncontrolled crash...

It's a good theory but I do not think it is the case with reported Inspire problems.

Many of the early Phantoms suffered what appeared to be stuck accelerometers. In fact in a few cases they had to be reset by a short sharp knock on a table to release the sensor. This could be done with both the Naza in the Phantom and the stand alone Naza FC.
The symptoms I have seen on every single Inspire 'flyaway' that has been video recorded do not exhibit the same symptoms as a stuck accelerometer.

In every case I have seen, the aircraft first starts to toilet bowl and then gets out of control with an ever increasing magnitude of error. This is classic compass related problem.

I would love to be proved wrong but to date I have not seen any evidence that suggest the 'uncontrollable nature' of the flyaways have been anything but poor aircraft preparation/knowledge coupled with too many calibrations in unknown environments with a bit of pilot inexperience, panic and a failure to switch into atti mode AS SOON as a problem manifests itself. I think the combination of all these things goes a long way towards explaining the large proportion of 'problems'

 

[blade strike]

I agree with the editor.. I will also say that the FC logic is completely new from the inspire up. The inspire and p3 have logic to stop these kind of build up of errors.

 

[pahl]

well... so i guess it might be a software/firmware problem.

Hoping a new firmware update increases reliability (and also improves the camera compression/quatily) *fingers crossed

Thanks for your responses!!

Great forum, by the way!!

 

[Daniel Mascheck]

How do you access all this data from your
inspire?

 

[pahl]

Hi Daniel,

that data is obtained from a can bus logger (such as FlyTrex) on a phantom and it´s available for naza based platforms.

I believe that they are creating a inspire 1 version but i´m not sure that´s available right now.

 

[ianwood]

I don't know how much is similar between A2 (FC on which Inspire / P3 are derived) and the NAZA M V2. I know there's enough similarity that they can share different CAN components. The messaging has a different structure but the underlying mechanisms are actually the same.

As for differences in the way the FCs work, it's hard to tell but we've seen enough A2 flyaways by people who know what they're doing to know that it too is not perfect. If I recall correctly, Inspire had some early GPS issues and then the ongoing compass and app disconnects on both Inspire and P3.

On the DIYdrones link, someone mentioned a buffer overrun. The data in my logs does look consistent with that type of a bug.

P.S. I've put a video together of the event. It only has a few seconds of the flyaway due to loss of power on the GoPro. However, it gives you enough to see it got bad quickly.

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies

 

[Kauaigene]

"Loss of power on the GoPro"...?
What's a GoPro have to do with this?

 

[The Editor]

"Loss of power on the GoPro"...?
What's a GoPro have to do with this?

Because Ian wasn't flying an Inspire at the time of this incident but an Naza equipped platform. The Inspire FC shares some of its design with/from the Naza/A2.

 

[mdomeny]

"Loss of power on the GoPro"...?
What's a GoPro have to do with this?

seems like it was a phantom with gopro (do not know which one that is) i am guessing but the emphasis here is on the fly away and why it can happen or rather trying to get to the root cause... nice vid and the music... 28 days... hope i will have more lol

ok i did not read the whole thread before posting
i have no clue what the gopro has to do with it...