What the heck happened. P2V+ 45 degree crash into ocean.

[Aerial Entertainment Studios]

Hello all

I purchased my P2V+ in Hong Kong and have been flying around the world with it for 2 months. No issues whatsoever in Fiji, Hong Kong, Indonesia, Thailand. Sample footage here:

https://www.youtube.com/watch?v=4B1YPD5b5lk

However, I brought it with me to my home country of Canada (Quebec area) and the unit compass calibration got so out of whack that it decided to go for a dive in the ocean. It's now gone and I have lost $1500. I would need your help in understanding what happened as I've been flying this thing for months without any issues and pondering the option of buying another unit.

Here are a few more details.

- Initial takeoff was followed by a circling that got bigger and bigger (and faster) unless I intervened to stabilise the unit.
- After about 10 flights, the P2V software detects a compass calibration issue and requires an advanced calibration.
- I complete the advanced calibration (all confirmed in the software) and take the baby out for a test flight. Same issue occurs, but with less magnitude.
- I take it to the ocean front and take off with no issues. Stable for first 2-3 minutes (altitude about 100 feet) and then enters a high speed 45 degree dive straight into the water. I try to recover by sending the unit straight up but with zero response from the Phantom.

On the bright side, no one was injured (this has always been my worst fear), but at the same time, I don't believe I made a piloting error.

I am seeking for your opinion on what could have potentially happened. Was this an engine failure, compass issue, or something else?

 

[Dirty Bird]

I am sorry for the loss of your aircraft. I am not sure what caused your Phantom to fail but, from your description of the loss and events leading up to the crash, it appears a component was breaking down. Did you check the compass readings in the Assistant? Did you perform an IMU calibration? Please do not be offended, as I am not trying to be harsh or unsympathetic to your loss but, were I asked to classify this crash with the information I have on hand, I would classify it as Pilot Error. Why do I say that? If the aircraft was behaving perfectly normal up until this event, I would have said component failure. But you knew the aircraft was behaving abnormally, yet chose to fly it anyway. Once she began demonstrating a repeated pattern of anomalous behavior, you should have checked & calibrated everything, or sent it in for service & repair.

 

[Aerial Entertainment Studios]

Hi, no worries.

I checked the compass readings in the assistant and was seeing x&z axis with positive values of about 400 (give or take) while the y axis was at -289. Mod value was about 1400. This was clearly not normal which is why i did an advanced calibration. Thought this would fix the problem but apparently it didn't.

I clearly decided to fly even though the unit had some issues. That's the pilot error and I agree. Having said this, I have trouble understanding the 45 degree dive into the water.

I probably should have brought it in for checking, but given I am in Canada on holidays and wanted to film some stuff, and the fact that DJI customer service is absolutely brutal, I figured it would be easier to simply buy a new unit in Canada instead. Sending a P2V+ in for repairs has not proven to be a pain free process, even in Hong Kong.

 

[Double-D]

Sorry you lost your Phantom.
Shame it was over the water when happened.
I can't help with any reason why but look forward to some of the wiser forum users input.
The fact it was acting strange is a hard lesson learned.

 

[plasmo]

Sounds like you could have experienced the Toilet Bowl Effect. I had that happen to me. Did you by chance lift off from anything metal?

 

[cougar]

Hi, no worries.

I checked the compass readings in the assistant and was seeing x&z axis with positive values of about 400 (give or take) while the y axis was at -289. Mod value was about 1400. This was clearly not normal which is why i did an advanced calibration. Thought this would fix the problem but apparently it didn't.

I clearly decided to fly even though the unit had some issues. That's the pilot error and I agree. Having said this, I have trouble understanding the 45 degree dive into the water.

I probably should have brought it in for checking, but given I am in Canada on holidays and wanted to film some stuff, and the fact that DJI customer service is absolutely brutal, I figured it would be easier to simply buy a new unit in Canada instead. Sending a P2V+ in for repairs has not proven to be a pain free process, even in Hong Kong.

Hi, very sorry for the loss of your phantom, may I ask where you lost your phantom? Quebec?

 

[Aerial Entertainment Studios]

In the Gaspé area (more specifically Grande-Vallée town) as we were visiting the region.

Took off on a large pier. Flew fine for about 2-3 minutes and then went haywire.

I've been reading a lot since yesterday and it seems many people have complained of similar issues since the software update with the inclusion of waypoints. The toilet bowl effect was not there before I updated the software. Could this be related?

 

[Dirty Bird]

I checked the compass readings in the assistant and was seeing x&z axis with positive values of about 400 (give or take) while the y axis was at -289. Mod value was about 1400. This was clearly not normal which is why i did an advanced calibration. Thought this would fix the problem but apparently it didn't.

So you did check and your readings all appear to be perfectly normal. I feel better knowing that you checked. I have noticed some pilots mentioning 3.04 having a more pronounced toilet bowl effect which seems to improve after a few minutes. This does not, however, explain your Phantom's sudden kamikaze plunge into the ocean. I don't suppose you had a GetterBack installed, or any chance of recovery?

 

[cougar]

In the Gaspé area (more specifically Grande-Vallée town) as we were visiting the region.

Took off on a large pier. Flew fine for about 2-3 minutes and then went haywire.

I've been reading a lot since yesterday and it seems many people have complained of similar issues since the software update with the inclusion of waypoints. The toilet bowl effect was not there before I updated the software. Could this be related?

Reason I am asking is that I am not to far away from Quebec and have similar tendencies with booth my PV2+'s. Have done advanced calibration and all values are normal in Assistant software but have (since new) a problem with the phantom going in bigger and bigger circles and faster and faster (toilet bowl effect) I manually keep correcting and after 5-10 minutes it will correct itself and fly fine. The Phantom also do not fly straight but going to right of intended course.

After the GS app came out I tried it and are not able to "hit" any waypoint, it just goes around and around trying so I have to abort the mission. If I try the same thing after first letting the Phantom "find it self" it works fine. Problem is that by the time 5-10 min, only so much of the battery is left so only very short missions.

I believe that my problems and some of yours is caused by Magnetic declination and have contacted DJI about this issue. Testing are being done right now and hopefully we will have an update.
Hope that this will help in understanding what what might be some of the issues contributing to your crash, even if you might have had other issues.


Edit: Have tried booth 3.04 as well as 1.08 (that I prefer) but issues are the same.

 

[slothead]

First of all, great video pgaucher!

This issue reminds me that I would really like to see an FMEA (Failure Modes and Effects Analysis). This is a standard analysis done on all military and aerospace programs, and may be overkill for the Phantom Vision, but I would still like to see one. It is one of the things that I do as a consultant for NASA and DAPRA. It could be done a variety of levels of details, but at least needs to be done at the functional level. As an example, lets assume a motor failure (as you suggested). If the failure mode is an open winding in the motor, what is the effect (does it make a sideways dive as yours did)? and how can you compensate for this (if at all) and what indications could you have that this may be about to happen (if any)?

Obviously these Phantom Vision failure modes are very different all over the bird, but it would be very interesting to see how the analysis worked out and could also help determine future improvements and operating methods/recommendations (from the predictive indications and compensating provisions).

I'm going to start on this myself (using details posted in this forum and any inputs you guys are interested in supplying. Obviously since this is a voluntary effort on my part, don't expect anything immediately, but we'll see how the project progresses. As a starting point if any of you have any specific failure characteristics that you would like to see included, let me know.

 

[Dirty Bird]

Speaking of motor failure, I saw an interesting article and video a while back about a control algorithm that can allow a quad to safely stay in the air and land if one of the motors fails. Check this out:

Article:

http://diydrones.com/profiles/blogs/new-algorithm-can-save-a-quadcopter-after-one-motor-prop-failure

Video:

[youtube]https://www.youtube.com/watch?v=bsHryqnvyYA[/youtube]

 

[Double-D]

Speaking of motor failure, I saw an interesting article and video a while back about a control algorithm that can allow a quad to safely stay in the air and land if one of the motors fails. Check this out:

Article:

http://diydrones.com/profiles/blogs/new-algorithm-can-save-a-quadcopter-after-one-motor-prop-failure

Video:

[youtube]https://www.youtube.com/watch?v=bsHryqnvyYA[/youtube]

That is way cool. I'd feel safer with that feature implemented.

 

[Double-D]

As a newb pilot I have a problem with the calibration and the values we shoot for. It seems there is no definitive source for what they should be. The best I have found is the Colin Guinn video where I think he says we want between -300 to +300 and 1600 but if I hover over the question mark in the Naza software tools tab it says between -500 +500 and around 2400 other values that are not the same as what he says in the video. Now I am probably completely confused so anyone who wants to help correct the above it would be good. (the above paragraph is a mess.

It's not a mess. I understand perfectly and now have the same question.

I've always worked on the assumption that the ?/help information in the assistant was correct.
Since day one I've had one of the values in the +400 range though never 500 and my other value, can't remember what it's called, anywhere from +1300 to +1600 though usually mid +14/1500 range.

Is there a normal average?

 

[JWarren]

Had your Phantom been acting depressed? Sorry, I know it's no joke and sucks. I'd think it would be safe to say that it all goes back to the compass. Had you done a field calibration before the final flight?

 

[jimdenburg]

I'm gonna go out on a different limb and suggest maybe it was a power failure to one of the motors. I know on my P2V I had some bad solders that popped off mid flight and while I can't confirm the 45 degree angle (it was above the trees and my first indication of a problem was when I heard it crashing through the limbs), it would seem to me that if the craft lost one motor due to a bad solder or if it just crapped out, the lack of lift on that arm would most likely cause it to dip to that side, and with the other motors still spinning and perhaps trying to compensate, I could see them pushing the craft into ground (water). Would be nice if the tide washed it in or you had some other way to make a recovery and inspect it.

Looks like you put the thing to good use while you had it though. Sorry for your loss. I lost mine too after about 5 months of heavy use. I justified the monetary loss in my head by recalling that I've lost more than that on a weekend in Vegas, and I got a ton of enjoyment (and some great video/stills) out of the thing in the time that I had it. So I got 5 months in Vegas for my $.

 

[rrhansen]

As a newb pilot I have a problem with the calibration and the values we shoot for. It seems there is no definitive source for what they should be. The best I have found is the Colin Guinn video where I think he says we want between -300 to +300 and 1600 but if I hover over the question mark in the Naza software tools tab it says between -500 +500 and around 2400 other values that are not the same as what he says in the video. Now I am probably completely confused so anyone who wants to help correct the above it would be good. (the above paragraph is a mess.

It's not a mess. I understand perfectly and now have the same question.

I've always worked on the assumption that the ?/help information in the assistant was correct.
Since day one I've had one of the values in the +400 range though never 500 and my other value, can't remember what it's called, anywhere from +1300 to +1600 though usually mid +14/1500 range.

Is there a normal average?

I have the same questions about compass values and Collins video versus the Naza help function. I have similarly experienced initial flights with 20 to 30 foot radius and increasing circles prior to control inputs to intervene. Can any of you experienced pilots or compass gurus shed some light please? Thanks in advance.

 

[Gatordent]

For those of you having the toilet bowl flight pattern, try to do a field calibration in the same spot you take off from, flight a few feet in the air and land as soon as you see the phantom toilet bowl. Without powering down connect to assistant and read the compass values in the field if that's possible with a laptop or tablet. If they are within the threshold of colin's video +/- 300 and around 1400 then there might be something else to blame other than the compass.

 

[Double-D]

I should add that I've never had any odd behavior such as toilet bowl with the compass values I posted.
Just pure curiosity and making sure they are a safe value.

 

[Gatordent]

Speaking of motor failure, I saw an interesting article and video a while back about a control algorithm that can allow a quad to safely stay in the air and land if one of the motors fails. Check this out:

Article:

http://diydrones.com/profiles/blogs/new-algorithm-can-save-a-quadcopter-after-one-motor-prop-failure

Video:

[youtube]https://www.youtube.com/watch?v=bsHryqnvyYA[/youtube]

That is way cool. I'd feel safer with that feature implemented.

this looks like it would work with a lightweight symmetrical quad, DJI should get on this. not sure you could control it with three props but at least soften the landing a bit

 

[Wedeliver]

I just came across this in another thread and I thought the op might be interested. http://fpvfactory.com/products/mariner-accessory/c-12/ a waterproof quad...

Also, I thought I made a post here but I don't see it. Since sometimes I try and be funny maybe I said something that caused it to be deleted. If so, oops sorry. above is a serious reply to the original poster and his great over the water pics. I thought this might be an option for him.

maybe a better link.. http://fpvfactory.com/products/mariner- ... f-625.html