P3 Unexpected Loss of Control upon landing
- Thread starter Serpent
- Start date
The OP stated "- Note: Visual Positioning System turned off (some flights ago)."
Does the VPS turn itself back on?
Meta4
Premium Pilot
- Joined
- Jan 8, 2014
- Messages
- 15,300
- Reaction score
- 9,242
DJI make it very clear that you shouldn't calibrate the compass near large metal objects.
Meta4
Premium Pilot
- Joined
- Jan 8, 2014
- Messages
- 15,300
- Reaction score
- 9,242
That's something that can be very easily discounted.
1. The moisture in trees (or any water) has no magnetic effect.
2. The problem observed is not compass related.
- Joined
- May 20, 2014
- Messages
- 94
- Reaction score
- 15
DrewFiles: The descent was on RTH and then i took control back as it neared the ground. I've never had a problem landing in dust/dirt before. Agree for not pointing the camera down to avoid possible lens damage.
EnvisionAbove: I've flown near to trees many times before without difficulties. The tree was mainly to provide shelter from the sun. Yes, it's possible the corrugated roofing may have thrown out one of the sensors. That said I've landed on structures where i'd previously had compass errors when starting off from without problem.
IflyinWY: No underground utilities. We're up in a remote place where there is no cell phone signal, hardly any electric and they are in the midst of creating a well for water.
DrChris: No. It's at 40/41 on all of the sensitivities. Gains are as per default.
snerd: The video is speeded up X16 during the RTH descent to reduce the video size (it shows that in text over the video). The problem wasn't the descent but came after touch down.
Meta4: I agree with you (except VPS was turned off). I don't think this is a compass issue. It was an sudden and erratic burst in power gaining height which then became barely controllable, with it losing height nearly as fast as it gained height after. I was countering the motion as best i could. At the end the drone wouldn't come to a stop with the usual downward left stick position.
EnvisionAbove: I've flown near to trees many times before without difficulties. The tree was mainly to provide shelter from the sun. Yes, it's possible the corrugated roofing may have thrown out one of the sensors. That said I've landed on structures where i'd previously had compass errors when starting off from without problem.
IflyinWY: No underground utilities. We're up in a remote place where there is no cell phone signal, hardly any electric and they are in the midst of creating a well for water.
DrChris: No. It's at 40/41 on all of the sensitivities. Gains are as per default.
snerd: The video is speeded up X16 during the RTH descent to reduce the video size (it shows that in text over the video). The problem wasn't the descent but came after touch down.
Meta4: I agree with you (except VPS was turned off). I don't think this is a compass issue. It was an sudden and erratic burst in power gaining height which then became barely controllable, with it losing height nearly as fast as it gained height after. I was countering the motion as best i could. At the end the drone wouldn't come to a stop with the usual downward left stick position.
At least SOME of that can be attributed to ground effect. During our test for our pilot licence, they produced that exact effect with a P2 bouncing up and down to demonstrate the effect that ground effect has on rotorcraft. With the right amount of power you can keep it going indefinitely without any input from the controller.
Ground effect creates an area where much less power is required to produce lift, so the aircraft lifts, but gets out of ground effect and no longer has enough power, and descends again, only to now rise again when it gets back into the ground effect again. You can keep the bounce going indefinitely with exactly the right amount of power.
Come down slowly and gently, preferably not purely vertically, then when about a foot above ground drop power cleanly to pass through the area of GE, moving forward keeps the disturbed air from developing underneath you, and then flare slightly on landing to remove the forward momentum. Watch how real helicopters land, they're very aware of GE and that's how they land to reduce the risk.
Ground effect creates an area where much less power is required to produce lift, so the aircraft lifts, but gets out of ground effect and no longer has enough power, and descends again, only to now rise again when it gets back into the ground effect again. You can keep the bounce going indefinitely with exactly the right amount of power.
Come down slowly and gently, preferably not purely vertically, then when about a foot above ground drop power cleanly to pass through the area of GE, moving forward keeps the disturbed air from developing underneath you, and then flare slightly on landing to remove the forward momentum. Watch how real helicopters land, they're very aware of GE and that's how they land to reduce the risk.
Just a guess ... but it looks to me like ground effects followed by either a return to RTH mode, full up on the throttle, or initiation of an auto-takeoff.
Is there a log file available?
-- Roger
Is there a log file available?
-- Roger
- Joined
- Apr 21, 2015
- Messages
- 69
- Reaction score
- 10
- Age
- 40
Looks like ground effect. Particularly with debris on the ground nearby. I've noticed that my P3 is a lot smoother to land when there's plenty of breathing room around. It's also always smoother on open pavement than open grass.
The other thing I'd look at is control inputs. They seem a bit jerky prior to landing so I'm wondering if throttle was also jerky. The app should log stick inputs. Give those a look to see if you were smooth.
It's most likely a combination of several things.
The other thing I'd look at is control inputs. They seem a bit jerky prior to landing so I'm wondering if throttle was also jerky. The app should log stick inputs. Give those a look to see if you were smooth.
It's most likely a combination of several things.
ok guys. we are a group of 5 friends that we all have ph3. two of us have the pro version and the other 3 the advanced version . This is an RTH issue most probably . what you should take in consideration before take off is the following : always make sure to tun on the remote and after the bird. switch off the bird and then the remote. never ever calibrate near metal objects. The other day that one of us tried to calibrate on top of the bird box that had the car keys in it and we faced issues with calibration. Also what we found very very useful...never land our drones totally to the ground...leave it to hoover about your high...take it with your hand from the base and move the left joystick down to switch off...much more safer this way...if you are always trying landing it to the ground you have chances of flipping it over ...or damage the camera lens if there is rough ground below it.
Yes, I read the directions. and I tested it. it's trueDJI make it very clear that you shouldn't calibrate the compass near large metal objects.
Maybe a bad or intermittent barometric reading. It did look like it was bouncing between two points in a control feedback loop. My phantom has done that when flying indoors in a confined space.
GPS mode + VPS = Feedback from visual sensors + GPS + Accelerometers + Barometer.
ATT Mode Alone + VPS = Feedback from Accelerometers + Barometer + Visual sensors
Maybe when the quad got into Ground Effect (Close to the ground) the barometer sensed a sudden change in barometric pressure, thus causing the phantom to think it was lower than it actually was. This caused the quad to climb to compensate.. Then it saw a sudden decrease in Barometric pressure, causing the quad to sense a higher altitude than it expected to be and it sent a control pulse to descend. This feedback loop happened a few times until you landed it.
GPS mode + VPS = Feedback from visual sensors + GPS + Accelerometers + Barometer.
ATT Mode Alone + VPS = Feedback from Accelerometers + Barometer + Visual sensors
Maybe when the quad got into Ground Effect (Close to the ground) the barometer sensed a sudden change in barometric pressure, thus causing the phantom to think it was lower than it actually was. This caused the quad to climb to compensate.. Then it saw a sudden decrease in Barometric pressure, causing the quad to sense a higher altitude than it expected to be and it sent a control pulse to descend. This feedback loop happened a few times until you landed it.
Now that is a good thought, perhaps the dust confused the VPS as to the actual distance to the ground.
If VPS is turned off it isn't surprising that the results are unpredictable. The software has to take several inputs and respond to each separately. It needs to stay level, control the downward speed, check the distance from the ground by two separate methods. It also has to measure the response to individual actions. At some point the software needs to have avoiding action which in this case seems to be just to go straight up.
Indoors the effect of turning off VPS is different.
Indoors the effect of turning off VPS is different.
- Joined
- May 20, 2014
- Messages
- 94
- Reaction score
- 15
StevenQX posted this in another thread (jump to 39 seconds into the video):
This is similar to what I experienced with the bouncing up and down effect.
The other effect is the P3 no longer maintains stable hover and is losing altitude without left stick upwards input.
In the video he had his VPS on. I'd previously set VPS 'off', unless it goes back to 'on' across resets.
The other effect is the P3 no longer maintains stable hover and is losing altitude without left stick upwards input.
In the video he had his VPS on. I'd previously set VPS 'off', unless it goes back to 'on' across resets.
Last edited:
N017RW
Premium Pilot
Lint?
The surface area of the contacts would pretty much rule this out.
The FC and other components cannot tolerate truly intermittent power as this would cause resets of all the processors.
If that occurs it's coming down.
The surface area of the contacts would pretty much rule this out.
The FC and other components cannot tolerate truly intermittent power as this would cause resets of all the processors.
If that occurs it's coming down.
I blame user error: those kaki shorts and crazy shirt you had on were enough to confuse Einstein.
- Joined
- Jan 29, 2015
- Messages
- 341
- Reaction score
- 77
The new firmware update addresses P3 altitude stability. That might be the solution.
- Joined
- May 20, 2014
- Messages
- 94
- Reaction score
- 15
Thanks Johnp. I will wait a week or so for other's reports prior to taking the leap.
Similar threads
