Dangerous behaviour of my new p4

[Rutzki29]

Bought yesterday a new p4 and finally got to fly it today. Calibrated the compass and IMU (firmwares up to date according the app) and the whole first flight in beginner mode went well except a little twitching of the bird while steering with the right stick. This did not bother me since i hadd full control and the bird did smooth and slow movements exactly how it was ment to do. Second flight the horrible problems started. Calibrated the compass again and the sliders were green and the values showed 1. Took the beginner mode off and lifted the bird up and flew perfectly fine the first 5 minutes. First problems i noticed when i was howering maybe 20m over myself and saw that the bird was not hovering still, it was doing a circle with the radius getting bigger and bigger. Tried flying forward and at this point it did it, started to look for a safe place to land and noticed that i was having close to zero possibility to control the bird anymore. Throttle up seemed to work fine and managed to dodge a tree with throttling up. Throttle down did not bring it down like it normally does. I have experience from cheaper birds with no aid tech, and without trying can say with confidence that i could fly the p4 in atti mode and control it properly. But this was no atti mode, i had zero control of direction of the bird most of the time. Tried even holding the rth button at some point but all it did was stop for a few seconds and then the gliding back and forward all around with pretty high speeds. Got it to land on uneven ground and powered it off. I was really confused of what happened, but decided to move to the field and try again on low altitude and see what happends. Everything worked nicely again at the beginning except minor twitching a few times. After a few minutes i noticed again that the bird was not hovering still and i started to look for a place to land. With close to zero control again i got it to land on the sandroad, only thing i was touching at this point was the throttle down, and i had it fully pressed down when the bird touched the ground. Suddently the bird flips the wrong way around and the propellers start to spin against the sandroad with really high speed. At this point i still had the throttle stick fully pressed down. Tried the emergency shutdown with failed succes a few times, and it finally worked and the bird shut down.
So my question is, what the hell can cause this kind of behaviour? File attatched shows the aftermath.

 

[msinger]

what the hell can cause this kind of behaviour?

It sounds like you have a bad compass calibration. Try calibrating the compass again outdoors in an area that is free of magnetic metal objects. That includes metal that might be on your person and/or in the ground. After you have a good calibration, follow this guide.

 

[Mark The Droner]

Yep. When the craft circles like you describe, we call that TBE, short for Toilet Bowl Effect and indicates a bad compass calibration.

 

[Meta4]

Second flight the horrible problems started. Calibrated the compass again ... and saw that the bird was not hovering still, it was doing a circle with the radius getting bigger and bigger.
But this was no atti mode, i had zero control of direction of the bird most of the time. Tried even holding the rth button at some point but all it did was stop for a few seconds and then the gliding back and forward all around with pretty high speeds.
So my question is, what the hell can cause this kind of behaviour?

By now you know.
Recalibrating the compass is unnecessary and as you've seen it can cause problems.
Get a good compass calibration and stick with it.
Here's an example.
This guy was one of those calibrate-every-time flyers.
He recalibrated right on top of a bunch of steel. (Reinforced concrete is half steel).

 

[Rutzki29]

It sounds like you have a bad compass calibration. Try calibrating the compass again outdoors in an area that is free of magnetic metal objects. That includes metal that might be on your person and/or in the ground. After you have a good calibration, follow this guide.

Thanks for the replies! Too bad this did not do the trick.
Gained some confidence and walked out to the field and did the recalibration of the compass, all data was ok after this and 17 satellites with strong gps signal. Did not have anything on me and the controller was about 10m away from the bird. Dedided to play it safe and lifted it up via the app. The bird only rose to maybe 1meter and started to do the same circular movement with rising radius. At maybe 2m circle i took it down slowly and when it hit the ground i pushed the left stick fully down, the bird tried moving forward on the ground again but this time it shut down before it flipped the wrong way up like last time. Had a real struggle of getting a working bird from the store, two times i got one with a faulty battery and now i finally got a working one yesterday, so im not giving up easily.

 

[Mark The Droner]

Is it a sports field? Is it possible metal was layed down underneath the dirt when they built the field?

 

[Rutzki29]

Is it a sports field? Is it possible metal was layed down underneath the dirt when they built the field?

Its a old traditional potato field with a sandroad going through it, family owned for over a century so i am sure of that there is no metal underground.

 

[GMack]

How did your VPS calibration go using Assistant 2.0? At <30 feet, I believe the VPS bottom camera and sonar have a lot to say about how the landings go.

 

[Rutzki29]

How did your VPS calibration go using Assistant 2.0? At <30 feet, I believe the VPS bottom camera and sonar have a lot to say about how the landings go.

That is the only thing i left un-calibrated. That is the next thing i will be trying. Problem is that when the first time the problems occured (circling while hovering) i was at +100 feet atleast. In beginner mode first time flying the bird it landed steadily and smoothly several times.

 

[GMack]

On my second P4, the bottom VPS was not working at all being new. First P4 was fine there, but had other issues with a crooked sensor. I returned both for a third (Love Best Buy!) and it went fine, although it does have its own personality too at times as it drifts when hovering by maybe a meter or two which I think is within their hover specs. Altitude is never good though on the GO screen and it waivers by 50 feet at times (Landing is never same altitude as take off was, maybe -45 feet, etc.). DJI's Q.C. really seems all over the place with these things - not to mention each bird's unique personality traits or the RC unit.

 

[m0j0]

That's the toilet bowl effect. Usually means bad calibration.


Sent from my iPad using PhantomPilots

 

[Rutzki29]

Now i got the forward and down sensors calibrated. Had to do it twice since the first time the software detected screen size wrong by 20 inches. Firmware update tab says that there is a month newer firmware available than the one i have installed. Weird thing since dji go tells me im up to date with the current. Hope they dont release any beta firmware.
Edit: it was a battery update

 

[MacCool]

That's the toilet bowl effect. Usually means bad calibration.

Or bad compasses/circuitry/part 43 mainboard. The OP sounds like he's calibrating correctly. People are just lobbing in "compass calibration" as the usual mantra that's expressed when anything unusual happens on a given Phantom flight.

I don't fully understand why DJI requires calibration. Why can't they program declination in firmware so that the compass is automatically calibrated for any given GPS location?

 

[GMack]

....

I don't fully understand why DJI requires calibration. Why can't they program declination in firmware so that the compass is automatically calibrated for any given GPS location?

Interesting idea.

I know the compass in my phone can act strange after calibration with some compass apps. Some want you to wave it a certain way, or spin it thusly, etc. Then it screws up and North is South, etc. Might explain the video above where he is flying backwards although camera is pointing forwards.

However, there are odd areas where the geology has some highly magnetic rocks and moving plates (West coast USA.) that might skew the compass if not calibrated for that specific area. Some coastal areas have those weird iron rocks too.

Could be it might make for more work too if DJI had to calibrate each and every drone made if the chip requires calibration when installed. So we get to do it for them.

 

[MacCool]

However, there are odd areas where the geology has some highly magnetic rocks and moving plates (West coast USA.) that might skew the compass if not calibrated for that specific area. Some coastal areas have those weird iron rocks too.

Could be it might make for more work too if DJI had to calibrate each and every drone made if the chip requires calibration when installed. So we get to do it for them.

I'm aware. Over the last 35 years, I've lived smack in the middle of both of the two largest areas of iron ore reserves in the USA. As a long-time full-scale pilot and hiker that often relies on magnetic compasses, I understand the regional and even local magnetic compass variations that can occur. IMHO, those things can be accounted for in software.

I do agree, however, that it's cheaper for DJI to let the consumer do it manually and put the responsibility on them. The problem is that there are so many opportunities to mis-calibrate when doing it manually.

 

[Rutzki29]

A little update on the situation. Had a friend over who has a p2 himself and he had a different technique in doing the calibration. He did not spin the bird in the same spot as i have done but he stood in the same spot and spun it around himself with his arm almost straight. Then he repeated the same process with nose tilted down. Question is, is that the right method or not? The app also asked me to calibrate the rc, wich i had not done before either. Worth to mention that we were in a different location too, all the calibrations were done there offcourse. The bird flew perfectly, no problems with anything at all. Interested of what happends in the future and how the bird reacts from revisiting the place where problems occured. Who knows, the soil might be rich of iron as mentioned in other replies. Thanks all for the help!

 

[msinger]

Had a friend over who has a p2 himself and he had a different technique in doing the calibration. He did not spin the bird in the same spot as i have done but he stood in the same spot and spun it around himself with his arm almost straight. Then he repeated the same process with nose tilted down. Question is, is that the right method or not?

That's how I do it. Both methods are okay.

 

[Erised]

Sounds like a compass but did you properly calibrate your RC?

 

[moswissa]

I was a calibrate everytime
Flyer until I learned my lesson! Only calibrate when you are at a new location!!


Sent from my iPhone using Tapatalk

 

[Monte55]

Or bad compasses/circuitry/part 43 mainboard. The OP sounds like he's calibrating correctly. People are just lobbing in "compass calibration" as the usual mantra that's expressed when anything unusual happens on a given Phantom flight.

I don't fully understand why DJI requires calibration. Why can't they program declination in firmware so that the compass is automatically calibrated for any given GPS location?

How would that make up for any accessories added to the bird or removed?