P4 Crash - Human Error or Malfunction?

Joined
Jul 10, 2016
Messages
21
Reaction score
4
Age
40
I've attached a link to a composite video as well the DJI Flight Record and an additional .info file (not sure of relevance). The composite video, 2:00 in duration, first displays the video that was retrieved from the sd card upon finding the P4 - you can witness the crash in it's entirety. The second half of the video contains 3 screens (1) the top left is the video capture from the sd card, (2) the top right is the video capture from an nvidia shield (the tablet connected to the RC, which natively captures the video being seen on the screen), and (3) the bottom center, shows a video of the DJI Flight Record - all synced to easier understand when the signal from the RC was lost and the subsequent events that followed. Lastly, a small clip has been added showing the current performance issues.

To provide an idea of context, my home point (and where I resided throughout the duration of the flight) is on a roadway, located just to the right of the tree line that is being shown in the middle of the screen. The road has an elevation difference of approximately 8' over the elevation of the field. I have an unobstructed, clear line of site over the extents of the field. This is my last shot of the evening, a slow forward pan over a ditch and across the far edge of a wheat field as an adjacent combine is harvesting. Upon capturing the shot, in GPS mode, I have the P4 pointed directly towards its home position. I change the flight mode from GPS to sport, shortly after the change of mode, the signal to the RC is lost. When you watch the video, you can see that shortly after the signal is lost, the P4 takes a downward trajectory, clearing some distance before hitting the wheat field. Having lost signal on numerous previous occasions, I maintained my position on the roadway when the signal first dropped, assuming the P4 would enter the RTH sequence and the signal would be re-acquired shortly. 5 minutes passes, nothing. Approaching 10 mins and now I am assuming something has gone wrong. I proceed walking into the field in the direction of the P4's last signal, thinking perhaps it was hovering and a closer proximity to the RC may re-acquire. Nothing. The field is 100 acres in size, the wheat growth is approximately 36-42" high, the time of day is +/- 8:50pm with sunset fast approaching.

The P4 was recovered. The search lasted just shy of 4 hours, well into the night. Rain started to fall, light to moderate, approximately 2 hours into the search (southwestern Ontario is supposedly on the verge of severe drought, of course it rains for the first time in weeks the night my P4 is lying in a field). To my surprise, the P4 did not have any immediate signs of physical damage (the shell, arms, propellers, lower armature, gimbal, etc.. all appeared to have been in good order). I cannot speak to the effect of the impact force on any of the internal components. The P4 was cleaned and let to dry for 36hrs, adjacent to a dehumidifier and light fan with desiccant.

My initial attempt to power up, the P4 would turn on in what seemed to be the usual fashion - audible and visual ques, gimbal extents check, prop twitch etc... The RC could not establish a connection to the P4. Multiple attempts without luck.

My subsequent attempts (this morning) have been hit and miss. The connection to the P4 from the RC had been established on a couple of occasions, of which, 2 of 3, the DJI Go App would identify my device as being a P3P. Turning off and on again several times I was able to get the RC to connect to the P4 and the DJI Go App to identify the device as a P4. I powered up and took off, maintaining a 4-5' hover as I begin to test several items. The P4 seems responsive to all movements (with light inputs in GPS mode), however, the gimbal and camera system have suffered some extent of damage that is prohibiting its function. See end of video for a capture of the gimbals response to yaw movements. In addition, the record ability in both picture and video will not engage.

In reviewing the flight record and videos, I cannot help but think that a component (hardware or software) malfunction was the result of the initial RC signal loss and subsequent downward pitch into the field. In all prior experiences with signal loss, the P4 would immediately cease its current operation and commence the RTH sequence.

 

Attachments

  • DJIFlightRecord_2016-07-07_[20-36-07].txt
    1 MB · Views: 256
I've attached a link to a composite video as well the DJI Flight Record and an additional .info file (not sure of relevance). The composite video, 2:00 in duration, first displays the video that was retrieved from the sd card upon finding the P4 - you can witness the crash in it's entirety. The second half of the video contains 3 screens (1) the top left is the video capture from the sd card, (2) the top right is the video capture from an nvidia shield (the tablet connected to the RC, which natively captures the video being seen on the screen), and (3) the bottom center, shows a video of the DJI Flight Record - all synced to easier understand when the signal from the RC was lost and the subsequent events that followed. Lastly, a small clip has been added showing the current performance issues.

To provide an idea of context, my home point (and where I resided throughout the duration of the flight) is on a roadway, located just to the right of the tree line that is being shown in the middle of the screen. The road has an elevation difference of approximately 8' over the elevation of the field. I have an unobstructed, clear line of site over the extents of the field. This is my last shot of the evening, a slow forward pan over a ditch and across the far edge of a wheat field as an adjacent combine is harvesting. Upon capturing the shot, in GPS mode, I have the P4 pointed directly towards its home position. I change the flight mode from GPS to sport, shortly after the change of mode, the signal to the RC is lost. When you watch the video, you can see that shortly after the signal is lost, the P4 takes a downward trajectory, clearing some distance before hitting the wheat field. Having lost signal on numerous previous occasions, I maintained my position on the roadway when the signal first dropped, assuming the P4 would enter the RTH sequence and the signal would be re-acquired shortly. 5 minutes passes, nothing. Approaching 10 mins and now I am assuming something has gone wrong. I proceed walking into the field in the direction of the P4's last signal, thinking perhaps it was hovering and a closer proximity to the RC may re-acquire. Nothing. The field is 100 acres in size, the wheat growth is approximately 36-42" high, the time of day is +/- 8:50pm with sunset fast approaching.

The P4 was recovered. The search lasted just shy of 4 hours, well into the night. Rain started to fall, light to moderate, approximately 2 hours into the search (southwestern Ontario is supposedly on the verge of severe drought, of course it rains for the first time in weeks the night my P4 is lying in a field). To my surprise, the P4 did not have any immediate signs of physical damage (the shell, arms, propellers, lower armature, gimbal, etc.. all appeared to have been in good order). I cannot speak to the effect of the impact force on any of the internal components. The P4 was cleaned and let to dry for 36hrs, adjacent to a dehumidifier and light fan with desiccant.

My initial attempt to power up, the P4 would turn on in what seemed to be the usual fashion - audible and visual ques, gimbal extents check, prop twitch etc... The RC could not establish a connection to the P4. Multiple attempts without luck.

My subsequent attempts (this morning) have been hit and miss. The connection to the P4 from the RC had been established on a couple of occasions, of which, 2 of 3, the DJI Go App would identify my device as being a P3P. Turning off and on again several times I was able to get the RC to connect to the P4 and the DJI Go App to identify the device as a P4. I powered up and took off, maintaining a 4-5' hover as I begin to test several items. The P4 seems responsive to all movements (with light inputs in GPS mode), however, the gimbal and camera system have suffered some extent of damage that is prohibiting its function. See end of video for a capture of the gimbals response to yaw movements. In addition, the record ability in both picture and video will not engage.

In reviewing the flight record and videos, I cannot help but think that a component (hardware or software) malfunction was the result of the initial RC signal loss and subsequent downward pitch into the field. In all prior experiences with signal loss, the P4 would immediately cease its current operation and commence the RTH sequence.

Not an expert on the subject here , but what it looks like is that you lost VIDEO control , but not RC Control and you drove that bad boy into the ground. Did you release the sticks after you lost video? Nice low flying effect , but is it worth losing your drone over? Any change in elevation? Also in Sport mode if you are flying fast , once you let go of the sticks , the P4 will drop about 8 feet. This has been discussed all over these threads.
 
Not an expert on the subject here , but what it looks like is that you lost VIDEO control , but not RC Control and you drove that bad boy into the ground. Did you release the sticks after you lost video? Nice low flying effect , but is it worth losing your drone over? Any change in elevation? Also in Sport mode if you are flying fast , once you let go of the sticks , the P4 will drop about 8 feet. This has been discussed all over these threads.

Familiar with the elevation drops associated with a sudden stick release... I would assume that if it was a loss of video control only, the flight record would have maintained the diagnostics to indicate as such. Once the signal dropped, I let go of the sticks. When looking at the flight record and synced videos, you'll see a change in downward trajectory (when the signal cuts) while velocity remains the same, not consistent with the slowing of velocity and elevation drop when sticks are released. No expert either, just trying to better understand what happened here.
 
I would assume that if it was a loss of video control only, the flight record would have maintained the diagnostics to indicate as such.
The TXT flight log stops recording as soon as the downlink (video) is lost. The internal DAT log contains a complete log though. Unfortunately, only DJI is able to read that log.
 
What is your loss of signal setting on the settings page? By default, the drones are set to stop/slow down reposition facing towards home and return direct to home. Doesn't seem to be the case here as the drone continues to move forward full speed but I guess the question here is if the drone takes a bit to slow down if the setting is set.
 
What is your loss of signal setting on the settings page? By default, the drones are set to stop/slow down reposition facing towards home and return direct to home. Doesn't seem to be the case here as the drone continues to move forward full speed but I guess the question here is if the drone takes a bit to slow down if the setting is set.
Loss of signal settings have not been modified from their defaults.
 
The TXT flight log stops recording as soon as the downlink (video) is lost. The internal DAT log contains a complete log though. Unfortunately, only DJI is able to read that log.
Unfortunate indeed. Thanks for informing.
 
Any other experienced pilots out there care to weigh in?

Look at the video at 1:10, you'll see that forward motion is the only stick applied. The signal cuts on the RC / flight record and your left with the video capture of the sd card. At the loss of signal, the P4 slightly changes to a downward trajectory and carries on over some distance before crashing.

Even if video only loss occurred, it's such an irrational (imo) action, to only slightly adjust my elevation to take it into the ground... Signal dropped (whether it was video only or video and RC) and with it the sticks.

In commencing a claim with DJI, I've inquired if they'd provide a copy of the DAT data file, which I'd love to see.

If it was human error, I have no problem eating it, it just seems very odd to me...
 
Anytime your bird pitches forward, it loses a little lift (translating into a loss of altitude). The faster and/or the quicker you increase speed, the more likely it is to lose some lift. The barometer altimeter isn't accurate enough to detect minor changes in altitude and therefore won't compensate by increasing altitude to mitigate the loss caused by pitching forward. As low as you were flying, as it increased pitch and speed, it lost just enough altitude to crash into the field. Signal to RC was not lost long enough to initiate RTH (3 sec). Signal was probably lost due to low altitude and the oblique angle the antennas on the AC would have been in (relative to the antenna on the RC) while it was pitched forward. If contact was lost immediately before the crash, it would continue on it's increased pitch/decreasing altitude path for 3 secs before RTH kicked in. It was down in the field before the 3 secs had passed.
 
Last edited:
With the latest update addressing the problem of altitude stability, you have a strong case that you are not responsible.


Sent from my iPad using PhantomPilots mobile app
 
What it appears to me to be is:
1. Drone was on a slight downward trajectory
2. Switched to Sport Mode as it was approaching the vegetation, pitched forward and naturally lost some elevation.
3 At this point the radio transmission was probably largely absorbed by the grain as it was just above the field, and getting very close to the grain.
4. Signal lost, and RTH never had a chance to initiate, as it takes a minimum of 3 seconds from signal loss. I'd say it hit the grain just after about 2 to 3 seconds. RTH never had a chance to operate.

In summary:
- Cause of crash, flying too low, across a signal absorbing land feature.
 
As an aside, I have an nVidia Shield K1 too, but I haven't tried out the screen video capture yet.
I'd assumed it would capture the screen you see, with the telemetry, map, and other details.
Does it only capture the video transmission?
 
As an aside, I have an nVidia Shield K1 too, but I haven't tried out the screen video capture yet.
I'd assumed it would capture the screen you see, with the telemetry, map, and other details.
Does it only capture the video transmission?
Geoff,
The K1 only captures the video transmission, it does so be default. The flight record was a separate recording.
 
Update and some additional thoughts...

I've been in contact with DJI, a description of events and a link to the video has gotten me a prepaid shipping label to Carson. Here's to hoping additional diagnostics and a DAT file can shed some light.

DJI Tech Support has informed me that they will make DAT files available upon request when you have an item that's been sent in for diagnostics / repair.
  • DJI Flight Record is actively receiving data up until 1:11, all numbers stop, however, the signal does not cut until 1:16.
  • If the P4 maintained its speed at 52mph, it is covering approximately 76'/s.
  • The P4 is in the air for 4s after the signal cuts, covering approximately 300'
  • The flight mode is switched to sport mode at 1:14
  • At 1:17 / 1:18, do you think that is a change in trajectory or camera movement? I'm certain there is no input to the camera or to the left stick coming from my end.
 
Geoff,
The K1 only captures the video transmission, it does so be default. The flight record was a separate recording.

Not sure about yours but when I use the screen record (under share) mine records exactly what I see on the screen. The go app or litchi, or whatever I am using and exactly as I would see it on the screen.


Sent from my iPad using PhantomPilots mobile app
 
Not sure about yours but when I use the screen record (under share) mine records exactly what I see on the screen. The go app or litchi, or whatever I am using and exactly as I would see it on the screen.


Sent from my iPad using PhantomPilots mobile app
Thanks. I've yet to use the screen record function, certainly of greater use than video only. Am still pleasantly surprised that it's capturing content regardless of my actions... helpful when something goes wrong.
 
Update and some additional thoughts...

I've been in contact with DJI, a description of events and a link to the video has gotten me a prepaid shipping label to Carson. Here's to hoping additional diagnostics and a DAT file can shed some light.

DJI Tech Support has informed me that they will make DAT files available upon request when you have an item that's been sent in for diagnostics / repair.
  • DJI Flight Record is actively receiving data up until 1:11, all numbers stop, however, the signal does not cut until 1:16.
  • If the P4 maintained its speed at 52mph, it is covering approximately 76'/s.
  • The P4 is in the air for 4s after the signal cuts, covering approximately 300'
  • The flight mode is switched to sport mode at 1:14
  • At 1:17 / 1:18, do you think that is a change in trajectory or camera movement? I'm certain there is no input to the camera or to the left stick coming from my end.

There does seem to be a discrepancy in recorded speed. At the end of the recorded data, around 14:42, the log file indicates a speed of about 32 mph, while the screen recording indicates 52 mph. 32 mph is actually consistent with the rate of change of location from the log file, and appears to be correct. The displayed value of 52 would be correct if the units were km/hr, but it is not traveling at 52 mph. This kind of units error was present in previous versions of the Go app, but I thought that it had been fixed. I doubt this has anything to do with the crash, though.
 
What it appears to me to be is:
1. Drone was on a slight downward trajectory
2. Switched to Sport Mode as it was approaching the vegetation, pitched forward and naturally lost some elevation.
3 At this point the radio transmission was probably largely absorbed by the grain as it was just above the field, and getting very close to the grain.
4. Signal lost, and RTH never had a chance to initiate, as it takes a minimum of 3 seconds from signal loss. I'd say it hit the grain just after about 2 to 3 seconds. RTH never had a chance to operate.

In summary:
- Cause of crash, flying too low, across a signal absorbing land feature.

It does appear that video feed was lost but the control link was still up, with full forward pitch in sport mode. It was close enough to the ground that it only needed to lose a few feet of elevation to crash.
 
It does appear that video feed was lost but the control link was still up, with full forward pitch in sport mode. It was close enough to the ground that it only needed to lose a few feet of elevation to crash.
At 10' above take-off point with a field 4-6' below the roadway take-off... Even assuming 10-12', that's more than just a few feet of elevation difference. I'm a bit surprised that between GPS and sonar, the discrepancy in elevational stability is that large.
 
At 10' above take-off point with a field 4-6' below the roadway take-off... Even assuming 10-12', that's more than just a few feet of elevation difference. I'm a bit surprised that between GPS and sonar, the discrepancy in elevational stability is that large.

No sonar in Sport mode, is there? Plus the corn field is not likely to produce meaningful sonar returns. Without that I think you are probably expecting too much in terms of altitude stability during relatively high-speed maneuvering, even if the actual speeds are closer to 30 mph.
 

Recent Posts

Members online

No members online now.

Forum statistics

Threads
143,094
Messages
1,467,600
Members
104,980
Latest member
ozmtl