Please Help Quad Destroyed

[Sandman1962]

In regards to Frank's photos....
Is it an illusion, or does it look, in the second photo, that the black wire is almost ""melted" into the MOSFET beside it?
( aside:It's like the Zapruder film isn't it?)

 

[Hughie]

Ultimately the ESCs overheat trying to push that power demand through.

The load pulls the power, the supply does not push the power. When an ESC blows it is usually because the motor is pulling more current than the final driver stage can deliver.

On the subject of why an ESC blows when a propellor is stalled, this is because the back EMF produced by the motor drops when the motor is stalled meaning that the only resistance remaining is that of the winding. Effective resistance drops, current goes up, something has to give.

 

[Hughie]

Hughie, if you are referring to the very center of your image, then no I have nothing like that on any of my ESC's. I don't know if I should have, but I don't.

Well perhaps we should consider what the issue might be if the faulty component is not an ESC. Perhaps it would be good to get them tested somehow ?

 

[flyNfrank]

Hughie, if you are referring to the very center of your image, then no I have nothing like that on any of my ESC's. I don't know if I should have, but I don't.

Well perhaps we should consider what the issue might be if the faulty component is not an ESC. Perhaps it would be good to get them tested somehow ?

Well but now these ones I have here are in bad shape. One has the black canister looking thing broke off, and a different one has the connecter broke off that the 4 wire plug goes into. I have two that might be useable for testing.

 

[mede8er]

I've seen the new motors up close and personal, the windings on the 2312 look superior to the original motors. Of course looks aren't everything....

Whether the MOSFET's failed first or the motor stalled, causing them to overload is the big question...

 

[flyNfrank]

Also not sure this is worth telling or not. But I watched a video where a V3 tipped over and when the props hit a solid surface, they just stopped altogether as if they had a clutch in them opposed to trying to rotate. Right away I figured the motors having sealed bearings were the reason, or, they had some kind of limit switch now that will help from getting as much damage.

 

[mede8er]

Yep...seems the majority of failures are not from modded birds, but from factory delivered birds....BTW FWIW, I don't think your build was faulty from what I can see. Not the prettiest job, but it looks like the solder points maintained integrity until impact.....

 

[Hughie]

Here we have 2 very experienced pilots, both of whom have recently done the v3 mods (motors and esc's) and both experience the same horrific results... I believe DJI has a serious quality control issue and a recall is in order....

Well, maybe that's the root of the problem. Maybe the main board in a v2 doesn't like something about a v3 ecs and motor. I assume DJI is selling those parts for a v3 and has never said that they could be used to upgrade a v2 -- not that I'm defending DJI. Just looking at what the cause could be. Considering that quality control is WAY LOW on the list for Chinese manufacturing, it could also just boil down to a non-obvious faulty part or connection anywhere on the bird -- which could be just waiting to fail on all Phantoms. But the first thing I would ask DJI is whether v3 ecs/motor set can be used in a v2 without problems.

I appreciate this is a Vision+ forum, but the same parts are fitted in P2 non vision and we should probably assume that the general problem affects both classes of P2. The version of the P2 NV with the new ESCs/Motors is V2 not V3. This makes the above a bit confusing.

There is upgrade information provided by DJI with the ESCs. It says that
New Esc compatible with new motor
New Esc compatible with old motor
Old Esc compatible with old motor
Old Esc NOT compatible with new motor

Nothing is mentioned about the main board.

 

[MacCool]

Ultimately the ESCs overheat trying to push that power demand through.

The load pulls the power, the supply does not push the power. When an ESC blows it is usually because the motor is pulling more current than the final driver stage can deliver.

Yes. How would smaller gauge wire lead to burning up the ESC? More likely, it would burn up the motor, or melt the wiring.

 

[flyNfrank]

I just realized I have never added my opinion on what I think the actual problem is. Due to the fact I actually got to see a problem take place right in front of me as I was looking right at it, is rare. I have had to think out other people scenarios so many times now that I'm shocked I was even looking at this when it happened.

Because the quad maintained altitude all the way to the point in which it rolled over to it's right side. And even held the altitude when it was slowing down quickly when I still have the lever pressed in the forward position.... I have to blame this problem on whatever controls the ESC's. If that would be the software, then I'm pointing at the software as the main contributor.

I do have the flight data from two separate flight loggers. Both are missing the end data. But I think the flytrex data shows more then the other one did. And maybe something will show up there? Wow..... I just looked at the data from the Flytex and there is some strange stuff coming into play. I think I will work on taking some pic's to share with those of you that wish to know as much as possible with these crashes. Not to keep anyone on a cliff hanger until I get back with images, What I'm seeing has to do with voltage & satellites received.

 

[Hughie]

cant wait :?

 

[flyNfrank]

Sorry I took so long to get this up!
I took images of the last 50-60 lines of data I have from the flight. I also added what each column is for a reference above & below the 3 .csv file images.

One most apparent piece to this puzzle is the Number of Satellites. The fact that the number of satellites is changing while the altitude is hardly changing. Next thing to pay attention to is the Voltage, and look close to see in most cases the Voltage change dictates the Number of Satellites. The 3rd thing to use as a tool is the Altitude & Speed. If you are familiar with looking at flytrex data on your own quad, the strange areas seen here will stand out like red flags.

I wish I was able to show this data with the video of the flight. It may help some understand better what may or may not be taken place. I do know the sound of the quad was normal when looking at the data you see. The column F shows distance and at the last line shows 233ft and if you look at a monitor pic it shows distance at 215ft. So this data stop very close to the point in which the quad was dropping in the sky. I would have to check with Amit at Flytrex to see if log in the device would have the complete file. The flytrex should have recorded to the point in which the battery was ejected.

Ok now here comes the most valuable question of all. And if you have any ability, or a resource that can help, please check into what I'm about to bring up.... Our quads have a internal data logger that logs what the quad does as soon as it is powered on. It saves error logs ect to a folder, in a compressed file I don't know. We need that info. We need to be able to access such info. My thought is that with the SDK setup we might be able to learn something of this there? If you have the ability to look into this, would you please. Or, if you know of someone who might know something about it, could you please ask them about it? Thanks everyone that gets involved and gives support, even if it is in the most simple way.

The 3 images are set so that the last line of data is in the 3rd image at the bottom. Don't forget you should be able to click on each image to see a larger and more clear image.

A-Latitude
B-Longitude
C-altitude(feet)
D-ascent(feet)
E-speed(mph)
F-distance(feet)
G-max_altitude(feet)
H-max_ascent(feet)
I-max_speed(feet)
J-max_distance(feet)
K-time(milliseconds)
L-datetime(utc)
M-datetime(local)
N-satellites
O-pressure(pa)
P-tempature(f)
Q-voltage(v)



A-Latitude
B-Longitude
C-altitude(feet)
D-ascent(feet)
E-speed(mph)
F-distance(feet)
G-max_altitude(feet)
H-max_ascent(feet)
I-max_speed(feet)
J-max_distance(feet)
K-time(milliseconds)
L-datetime(utc)
M-datetime(local)
N-satellites
O-pressure(pa)
P-tempature(f)
Q-voltage(v)

 

[Mopar Bob]

One most apparent piece to this puzzle is the Number of Satellites. The fact that the number of satellites is changing while the altitude is hardly changing.

Frank, aren't the number of satellites always changing. If you just hover, the sats above are revolving around the earth and passing command to the appropriate one as they go out of range. If you check the satellite predictor tool, maybe you will see a connection with the reduced sats.

 

[flyNfrank]

Bob, I have 250-300 recorded flights from flight loggers and I have never seen anything like that, ever.

I do want to make sure that because I pointed out about the satellites, that I'm trying to say they are part of the blame, cause I'm not. I was showing how one thing had an effect on another. Overall what I am working towards is the fact that there is some confusion going on, by something.

 

[Great Pumpkin]

flyNfrank -
Eventually when you've been able to take a deep breath and can find the time, it would be extremely instructive to the rest of us if you would write and post a summary of the entire affair, from the mods you made, to the Phantom's flight history from the mods to the crash. Add your observations while it was in flight and while it was falling. Then include a detailed description of the damage, your opinions about what caused the crash, and THEN if possible, the real cause of the failure and crash. Much of this information is already in this thread, but it is scattered and written while you were still in panic mode and puzzled.

 

[justin00]

Frank, nah they all have V2 on the part code in gold, even really old ESC. It is the software version that is written on the QR code.. e.g 2.0 or nothing in our case.
The QR code identifies them.. its a pretty stupid system.. but that is what DJI told me a few weeks ago. You've got the same as Bob and I. Ours also have v2 in gold, but apparantly v2.0 due to no beeps and red leds after starting motors.. Kinda irrelevant now and probs the last thing on your mind, but did you happen to get the 100ft time ?

But yeah, probs would of happend even if you had the one with QR 2.0.. Heaps of crashes with that combo. I think DBS has 3 2.0 and 1 2.1.. Dunno where the hell you get 2.1 from...

It is a shame you are one of the only few willing to take the top off during a crash. If others did they same we could look for similarities and do DJIs job for them.

Hughie, I only flew with default gains.

If I'm to blame for what happen I have absolutely no problem facing what it may have been, and wearing it. I did things and took precautions that I felt were over & above what was necessary. Example, after completing the 12 motor to ESC soldering points had enough time to cool, I lifted on each wire to see if it would either break off, pull out of the solder, or if the solder would break free of the board. I lifted to the point that the quad would just begin to slightly raise from the table.

Justin00, the ESC's show V2.0 on the stamp in gold lettering. I was under the understanding that with the combo of V2.0 Esc's and Firmware 4.6 you were good to go. When searching for where to find firmware 4.6 would be listed at, I learned that if you had the v3.08 firmware it contained the 4.6 firmware for the ESC's.

So while looking over the mess after removing the top cover, I looked for anything that appeared burnt, black, or melted, and nothing stood out anywhere. Someone mentioned a mosfet looking suspect. The discoloring that appears on the mosfet is I use flex resin and that is what it looks like. It comes right off like grease. There was several mosfets(board connectors) pulled off the ESC & main board. The main board is broken and the modfets closest to the break were the ones that had pulled from the board. The gps puck is shattered. Several bosses were broken and I had to use a screw driver to line up the holes in order to remove the screws. 3 of the motors turn hard with one unable to do a full rotation. Dirty Bird is right, the impact is what messed them up to the condition they're in.

The condition of the quad after the crash is one thing. But as some have wisely pointed out, there are so many others experiencing the exact same thing. They all have the same characteristics from start up to crashing. It because of that fact, you just about have to eliminate the possibility that the pilot is at fault.

 

[Mopar Bob]

Justin, your point is a good one about doing DJI's job for them. If DJI would take a step towards resolving some of the issues pointed out on this forum, it would save everyone a lot of time and money. These machines are not cheap. Loosing one sucks. Frankly, if it wasn't for this forum and other forums like it, many of us would have our fingers up our asses trying to keep our Phantoms flying. Major contributors like Frank, DirtyBird and a few others have unknowingly helped me understand some of the gobbly-gook that comes out of DJI. The Phantom is a nice machine. But if there was some decent competition I would Ebay the Phantom and buy from a company that gave a ****.

 

[justin00]

Yeah true.. I have a few intermittent issues.. such as black screen instead of video when tele works.. and you open app a few times and it suddenly works.
On the DJI forum I have seen heaaapps of v3 owners say after they updated they only get a black screen.. So I'm lucky I guess...
So many problems.... Its amazing.. Seems like the vision + was starting to work pretty well, and you'd hope the v3 would be thoroughly tested.. but seems like the v3 has so many problems.. I would not want to get one.

Lets hope DJI say something soon... as you said, these are not cheap at all. I paid $1800 AUD... crazy money...

*edit* I had a look at your 3 data sheets Frank, which column where you concerned about ? The satts ? It just looks like you lost a few... seems normalish? I mean I can loose a few sats by just flying 100ft horizontally in some areas where I fly.

Also as you probably figured out, the crash file just tells you which line in which file (and the line number in the file) the app had an issue with.
If you are not familiar with decompiling java apps, a really really easy way is to download an android app from the play store called "show java". Load her up then select the DJI App.. it will take 30-60 minutes to decompile. Then to make it easier to search / goto line numbers, connect your phone to your pc and copy the files across that it creates. The files will be on your phones storage in a directory called java or show java or something to do with java, its quite obvious. Otherwise you can look at the files in the show java app but you can't search or anything, so much easier to use pc.
Decompiling java apps requires a few programs in windows usually (well for an amatuer like me it does) so the show java app makes it so easy to do.

I'm pretty sure the crash file is relating to the DJI App, as opposed to your Phantom, but I suppose if the app has an issue god knows what could happen. Might be worth looking at each line number in the relevant file and having a look at what the function is.

If people are bored I would recommend everyone does the above to have a look at all the functions. Its quite educational, even if you don't know java.. Basically shows you how the DJI App works... Like for e.g the IP Addresses are hardcoded in, as well as the ports, so if you were to change the ip alot of stuff would stop working.

 

[eyecon82]

Mine just almost fell out of the sky yesterday http://youtu.be/ogExIuRQHC0

 

[eyecon82]

And I have a v3