Caution: LONG post, but will hopefully help save your P2V+.

I have the new style P2V+.  Never crashed nor a hard landing, but have lost two ESCs: once from a simple tip-over when powering down motors, and I think when I clipped a small stand of desert grass.  The ESC actually failed on my next flight trying to take off.  Puff of white smoke.

I repaired one ESC by reverse engineering it for the schematic and replacing the MOSFETs.  Then with oscilloscopes connected to pertinent motor drive signals, characterized ESC performance under various conditions, including a 20-minute full power tethered flight to having some interuption to the props or sudden change in load - just to see what it takes to blow the MOSFETs on an ESC.  Turns out ... quite easy!  I repaired and intentionally blew up the ESC three times to verify the cause.

ANY INTERUPTION TO THE MOTOR ROTATION, SUCH AS A TIP-OVER OR GRAZING A LITTLE TREE BRANCH *WILL IMMEDIATELY BLOW THE ESC.*

Recommendations below ... but here is what seems to be the failure mode.

The ESC delivers the 3-phase drive pulses to the motor windings.  Two windings are energized, one is OFF at any given time.  The winding that is OFF is momentarily used as a generator, producing the back EMF (BEMF) voltage pulses that tells the MPU where the stator is compared to the windings to make motor speed adjustments and proper commutation sequencing.  All three BEMF windings are summed together to form the "common node" voltage.  This is the reference voltage for the MPU internal zero-crossing detector for determining rotor position.  If you have a tip-over or otherwise interupt the rotation of the motor, there is a sudden loss of BEMF pulses and the summed common node voltage goes to zero, and thus no zero-crossing rotor position.  The MPU suddenly doesn't know where the rotor is compared to the energized windings.  It doesn't know which windings to energize next to sequence the motor.  As a result, the energized winding, instead of being energized for only a few milliseconds, remains energized for a second or two (waiting for the next BEMF pulse).  The motor windings are a near short circuit, the reason they are "pulsed."  Now you have the MOSFETs trying to drive a near dead short, which would be nearly the full capacity (+11v 5A) of the P2V+ battery.  During this momentary short, the MOSFETs get very hot until they fail.  MOSFETs generally fail by the substrate melting, shorting the MOSFET (drain to source).  This extreme heat destroys the MOSFETs, as many of us have visibly seen.  The short circuit also causes the P2V+ battery to turn off to protect itself.

I do not know what can be improved on the ESC board to prevent this.  Once the battery detects a short and shuts down, it's pretty much over.

Thus, on a tip-over, as soon as the prop strikes even some soft sand and interrupts the motor rotation for a blink of an eye, the ESC will fail, a couple of MOSFETs go up in smoke, and you may find the battery has also suddenly turned itself off.  Now, imagine if that happens while in the air - skimming some leaves in a tree or something.  The blown ESC and battery shutdown will make your P2V+ fall from the sky like a rock that we have all read about  way too many times.

During my full power tethered flights, the wiring temperature was monitored.  It never exceeded 40C, below the melting point of wire insulation, though the insulation (actually, shrink tubing) was "mushy."  The MOSFETs were very hot by the end of the flight; the motors just kind of warm.  MOSFETs switching high current *do* get hot under normal conditions. Still, I didn't see any compelling signs of wire melting or any design flaws with the DJI ESCs.  Basically, they're quite clever and well designed.  I no longer believe the motor wires are too small and causing problems.   And, inspite of intentionally blowing up some ESCs for these tests, it never damaged a motor.

RECOMMENDATIONS
(The ones I will follow myself from here on out)

1. Don't EVER, EVER, EVER let your spinning props strike anything.  EVER.  It will blow the ESC.
    Trees, shrubs, grass ... anything that will come in contact with your spinning props.
2. From here on out, I will hand catch my P2V+ for landings.  (I practiced today.  After 2-3 times, it becomes duck soup).
3. I do not intend on getting prop guards (yet), but it does certainly add to the importance of having them.
    Especially if you don't do number 2.
4. To avoid any possible rotation interruption while in flight, I will no longer conduct abrupt course or altitude changes (even though
    I love watching my Phantom bank on a direction reversal!).  There's the possibility that an abrupt change in direction, for
    which the four motor speeds are suddenly changed and redistributed, combined with prop wash, *may* cause a temporary
    stall in rotation triggering an ESC failure.  We all know what happens then.  I don't know that for sure, but I plan on being more
    conservative and "smooth" in making direction, yaw and altitude changes.  Remember, the new 2312 motors have 25% more
    power and 25% more torque to make impressive direction changes.
5. This certainly implies to never fly your bird out of line of sight.  There might be a tree out there to snag your prop.
6. I also recommend to separate the 3 motor wires going to the ESC board so they do not touch each other, AND ensure they
    are above and away from direct contact with the six MOSFET chips.  On my P2V+, the motor wiring was slightly twisted in
    contact with each other and resting on top of a couple of MOSFET chips, causing excessive heating of the wires.  Some
    simple separation and avoid the heat from the MOSFETs will go a long ways keeping the wiring in good shape.

And lastly, I don't think I'll ever fly over water.  Not an ESC issue, but it sure seems it raises the unlucky factor by a ton. :-(

Finally, the signs of blown MOSFETs on your ESC is when your props try to move in one direction, then the other, seemingly hunting back and forth, seldom or never making a complete revolution.  If it does that, bad ESC, and likely not the motor.

I do not know how this issue might apply to all P2Vs falling out of the sky, but I am convinced it addresses the scores of reports of ESC failures following a simple little tip-over, including my own, or other inadvertent contact of something with a spinning prop.

I'd be interested in anybody who has also torn into the ESCs to figure them out, or your flight experience when you blew an ESC and if consistent with the above.  Hopefully being a little more cautious about tip-overs and a little conservative on ambitious flight control maneuvers will keep your P2V+ and many others in the sky for many enjoyable flights to come.

Paul
In the New Mexico desert

PS - I will prepare a document with the reverse engineered schematic diagram and the oscilloscope waveforms for troubleshooting failed ESCs, and will post shortly for those wishing to repair your ESC at the component level.

Thank you.  After reading this my prop guards are on for good.

Thank you for your post.
The most intriguing thing in your report for me, was that at the start you mention that you had a partial motor stall which did not take out the ESC until the next flight.

I wondered whether this could explain some of the failures we are seeing where ESCs are failing many minutes into a flight.  For example, ESC gets weakened by motor stall on one flight, but it still capable of spinning up and taking off on a subsequent flight.  Any undue stress on the ESC (such as sudden change of direction) which may not have otherwise taken out the ESC, may actually be enough to take out a damaged one.

Is there any credibility in this hypothesis.

If there is, then this surely means that corrective action would be to replace all affected ESCs after every crash or stalled prop incident.

Thoughts ?

Hi Paul,

Thanks for the update, I have a tip over a few weeks back but so far I do not have any issues with the phantom. Do you recommend to open up the shell to check and separate the cables ?

HI Paul,

Very interesting reading.

I am currently studying the forum as I intend to buy a P 3 when they appear in 2015; hopefully a lot of the issues herein will be resolved.

I have learnt a lot and now understand how to manage various bits, what addons i want from day one  and what to look for.

Your article is excellent.  If you have a photo or two of the wiring and where you recommend to seperate etc it would be appreciated. ( i dont have a Phantom so cannot open one up. However have read manuals  / you tube etc.

Thanks again.

Last edited by daveac@hotmail.co.uk In 2015-1-4 21:24 Editor


I have been reading about the esc failures, I have had my phantom for a week now with 4 x 20 minute flights, no problems as yet, should i be worried about the esc wires melting, I have seen a video on u tube of the wire size comparision. yes the wires look a lot smaller.

Should I contact my dealer and raise my concerns?.  I already had a flyaway with a blade qx350 v2 after just 4 minutes of flight!!!.
I bought the Phantom hoping it was a better product, now I am very nervous about the esc wiring, I can not afford to risk this machine.

If one motor spins slower then the rest does it mean that ESC is bad and should be replaced?

on my second flight I've tipped fantom and it burned ESC.

Replaced it and it worked good.

Now 20 flights later I tipped phantom again (on replaced ESC side) and that motor now spins slower then the rest on startup but flies just great. Should I replace the ESC?

Hi Paul I hope you don't mind I posted your hypothesis on Phantompilots.com  
http://www.phantompilots.com/vie ... 32368&start=200

Very well-reasoned treatise. Thank you for posting it.

So....likely the burned wiring that has been reported occurs after the MOSFET fries? By the time the wire gets that hot, the MOSFET is already dead and the Phantom is crashing?

Is it also reasonable to conclude that the reason the previous-version motors didn't seem to have this problem is the fact that they draw less current, giving more time in a motor stall before the MOSFET gets hot enough to die?

Some posters elsewhere have reported losing a prop during these events. Does the motor likely stop quickly enough that the prop can spin off by inertia? Feasible scenario?

This also fits with a video with sound that I saw where the Phantom, in the midst of an abrupt maneuver, just shut down. Not one motor failure, they all shut down at the same time in mid-air. I guess one fried MOSFET equal loss of all power as the battery turns itself off.

Not sure if I just got lucky, but had a tip over while landing yesterday. First incident since I got my P2V+, so I was a little freaked out. I was actually landed, then did CSC to stop motors when the tip occurred. Changed the 2 props and did a test flight, everything seems ok. Anything I should check specifically before next flight? Thank you for this informative post!

I agree, thanks for a great thread.

I have the Phantom 2 Vision---(no +). (small round white camera)
I bought it from DJI's website in Nov. for a black Friday price.
It has been tipped over on its props at least four times---a few times quite severely. (props trying to furiously spin)
I have also had to replace a few broken props from an encounter with a small tree.
Question:  Does MY phantom have the earlier ESCs or the new, problematic ones? It seems to be in perfect working order.  Other than one crazy incident over a power-line in week 1, my drone has done exactly what I've commanded it to do...

I have just repaired my Phantom Vision 2 + v3 after esc failure, noticed new one was now V2.1 (one I removed was V 2).

mikelycett@btin Posted at 2015-1-5 01:44
I have just repaired my Phantom Vision 2 + v3 after esc failure, noticed new one was now V2.1 (one I ...

I guess Paul's theory does add more importance to the firmware versions of the ESC's.

Can you post a picture showing the parts that burned up?

Here is a link to a pic  in another thread.

http://forum.dji.com/forum.php?m ... amp;page=1#pid25137

I'd like DJI to address this growing problem asap!

here's my old burned ESC (from my second flight, minor tip over on landing)

Where can I find replacement esc if and when I need one?

Pmcdn2014 Posted at 2015-1-5 04:22
I'd like DJI to address this growing problem asap!

I think they will once it becomes clear - but being as they were already tried, judged and hung for it being a problem with the motor wiring.....I think it's prudent that they have enough real data to work with.

I PMed one of their engineers about it and he is going to address it...my thinking is they want to see if they can duplicate or confirm the problem(s) first. After all, we can speculate here....but they have to actually address these things - and hopefully only once (a problem - if there is one - and a fix).

As a business if there are sufficient notes on an issue you will opt to evaluate scientificaly to prove it out but as MANAGEMENT you will  improve it on a temporary basis to minimise cost to the business. Whether the issue is real or not.  Profits are key.

You then await the outcome of the replication test.

You mean to tell me that there is no over-current protection on the ESC boards???  Looks like DJI needs to do more upgrades for the ESC boards!  Also, I would re-design the ESC PCBs, placing the output terminals for the motors as close to the motors as possible.  And also increase the wire diameters to the motors, like they were before the "upgrade".

Paul,

Very nice work and post.  I feel your synopsis and recommendations are spot on.

I will fly my V3.0 with renewed confidence.

Jamie

Awsome report, Paul. Thanks for sharing it.

How old is your Phantom? I wonder if there's any sort of differences between the same models/version of P2V+ (I mean different kind of MOSFETs or other electronics)

bloah Posted at 2015-1-5 04:45
here's my old burned ESC (from my second flight, minor tip over on landing)

I watched a video of a v3.0 lifting three other phantoms off the ground it may have been two anyway I would think that if anything would have put the esc under a heavy load and it didn't fail. How would you compare to a blade in grass situation?  I can only expect if the blades were held back from moving it would smoke the esc for sure.

jeepinocala Posted at 2015-1-5 08:22
I watched a video of a v3.0 lifting three other phantoms off the ground it may have been two anywa ...

na5n Do you think there would be any point in performing the same tests on a pre V2 ESC with the 2212 motors?  I am going to be upgrading my Phantom 2 v1 with the new V2.1 ESC's and 2312 motors (maybe not after reading this post ) and I would be willing to donate the hardware if you would like to do a comparison test.  Thoughts?

Interesting read Paul.

Paul, thank you very much. Do you, or anyone else, have a recommendation for a way to insulate the wires from the MOSFITs and each other, perhaps some heat insulation material? Thanks-

droneflyers.com Posted at 2015-1-5 05:10
I think they will once it becomes clear - but being as they were already tried, judged and hung fo ...

I have to wonder about the professional competence of a company (DJI in this case) if pretty much all of the scenarios that are likely to occur in the field by the average user (eg. tip-overs) were not fully tested prior to a products launch. This seems to me to be very basic research and development.

GREAT POST and I totally agree with it

yeah I'm a noob, but first flight was flawless, second flight landed with a bump and tipped, screwed the  rear left ESC, I replaced it myself, checked everything and had her flying perfectly @ about 10ft up, a gentle landing and I was happy, 5 seconds later, lot's of arcing and smoke coming from the opposite ESC, haven't opened her up yet, still waiting for my replacement to be delivered ( I have bought 5 ESC's )

I'd love to see a cage type cover that would remove any chance of the props coming to an abrupt stop and of course it would have to not affect prop performance ,that would definitely put my mind at ease LOL

Cheers
Stuart

skburnett@hotma Posted at 2015-1-5 21:41
yeah I'm a noob, but first flight was flawless, second flight landed with a bump and tipped, screwe ...

"I'd love to see a cage type cover that would remove any chance of the props coming to an abrupt stop and of course it would have to not affect prop performance ,that would definitely put my mind at ease LOL"

You mean in addition to the offering of prop guards, which are available?

yes, coz it was the prop guard flexing that stopped my prop when she tipped over

skburnett@hotma Posted at 2015-1-5 23:20
yes, coz it was the prop guard flexing that stopped my prop when she tipped over

Ah.  Hit the limit of what the prop guards are capable of defending against.

so you are saying when an ESC will fail for whatever reason the battery will not give any power to the system?  

No Deninho, this is not always the case, when I tipped over, the rear left ESC had one of it's MOSFET's damaged, Phantom still powered on, but the rear left motor was erratic, I can only assume that the rear right ESC had damage to it at the same time, but I never noticed (maybe a solder fracture?) I was pretty lucky to land when It did coz I can bet it would have been quite messy, the rear right died spectacularly and went out in a blaze of glory, when I now try and power up the phantom, the rear ESC LED's blink red, then the battery switches off, all other led's blink green, Ill replace the ESC and take it from there  

This is where the logic of the firmware needs to be changed to be, well, more logical.

First, if the position pulse doesn't arrive within a certain amount of time  (Minimal pulse time for the minimal RPM) then consider it a fault and try to restart the sequence  (Restart the motor).  If it doesn't restart after a few tries then disable it.  You're already going to crash so no since in burning everything up and temporary stops in rotation will not fry the ESC.

My milling machine uses a BLDC motor and if I stop it, it immediately flags Fault and stops the pulses.  Now, take it one step further and try to start the rotation again and check for that pulse.  If it doesn't appear then either retry or stop but don't keep sitting there waiting for a pulse when it's obvious the motor is not spinning.

Next, a current limiting circuit to the ESC.  Not a fuse, but something that would limit the maximum current one ESC could draw safely before going up in smoke.

And, while they are at it, rearrange the pads on the ESC for the motor wiring to the end of the ESC towards the motor to reduce the length of wire and to reduce shorting.  It would also keep the wire away from the top of the MOSFETs where all the heat is generated.

rwskinner@ccwip Posted at 2015-1-6 04:03
This is where the logic of the firmware needs to be changed to be, well, more logical.

First, if th ...

I wanted to add, if your commanding a certain speed from the BLDC, then you know in advance what the pulse time should be and that in itself is part of the error detection.  The exception is the initial start up, and sudden load changes, but that interval would never exceed very many milliseconds unless there was a problem.

Hi, Paul I cannot do anything but to confirm your tests. Small tip over and ESC says bye bye. But now the 2.1 ESC Firmware  is out there. You should test them too, to see if 2.1 is a solution to this problem.

In my opinion DJI should send all 4x revised ESC to EVERYBODY WHO OWNS P2 V2.0 and should say SORRY LOUD AND CLEAR.

But first we need to check if new revised ESC is really burn proof.

I'm a new pilot with a PV2+ Ver-2, and I've had at least 6 crashes from 10 feet up: poor compass calibration-panic crash, not enough satellites-panic crash, too much metal and rebar and interference-mind of its own crash, and at least 3 tip overs-landing crashes. Luckily no gimbal damage, and not a single ESC failure. These are the no-version# ESCs with the 2212 "old" motors. Hopefully I can learn how to fly before I need those V3 parts (which unfortunately I have already procured, six 2.0 ESCs and 2312 motors) Think DJI will let me return my 2.0 ESCs?

How do I tell if I have a version of the Vision + that might have this issue?

Paul, thanks for the post and it helps a lot.
About your issue and recommendation D-boy and me have already reported it to developers.
Let's wait for their feedback.
And thanks a lot for your work.