Did he fly it into the ground?

Joined
Nov 6, 2014
Messages
8
Reaction score
2
One of my pilots flying a mapping mission (GS Pro) and P4Pv2 went into "Smart Power Go Home" mode. P4P was about 0.25 mile from PIC when in it went into RTL. As it was returning, he applied down throttle to bring it down as it returned. (This is somewhat standard procedure.) I think he didn't have it in site and flew it into the ground.

I pulled the DAT file from AC. I extracted the data using DatCon.
Can someone verify that the Controller:ctrl_thr: D represents his stick input?
RTL started at 875.104sec
Ground impact at ~906 sec - battery was ejected.
There are also 3 "fly_away" messages in log (see screenshot below) that occurred during the decent.
What else should I be looking for?
Note 1: I didn't include GPS:Lat as it was flying almost due west and value didn't change much.
Note 2: Vision safety is disabled as we have an external GPS unit that mounts beneath the downward vision.
201909_P4Pv2_287.png


1602369076193.png


1602369468590.png

Trees have been cleared since this image was taken.
 
I pulled the DAT file from AC. I extracted the data using DatCon.
Can someone verify that the Controller:ctrl_thr: D represents his stick input?
Does GS Pro create a .txt file log like DJI Go 4 does?
If so post that as it will clearly show joystick inputs.
 
That is another interesting thing, all of the text files from GS Pro from the last 2 outings have not been readable. I use dronelogbook to track flights and it couldn't read the files and I just tried the one at phantomhelp and it couldn't read them either. Is there another tool to try?
 
FWIW, these are after the upgrade to iOS 14 on iPad... just in case there is any connection.
 
One of my pilots flying a mapping mission (GS Pro) and P4Pv2 went into "Smart Power Go Home" mode. P4P was about 0.25 mile from PIC when in it went into RTL. As it was returning, he applied down throttle to bring it down as it returned. (This is somewhat standard procedure.) I think he didn't have it in site and flew it into the ground.

I pulled the DAT file from AC. I extracted the data using DatCon.
Can someone verify that the Controller:ctrl_thr: D represents his stick input?
RTL started at 875.104sec
Ground impact at ~906 sec - battery was ejected.
There are also 3 "fly_away" messages in log (see screenshot below) that occurred during the decent.
What else should I be looking for?
Note 1: I didn't include GPS:Lat as it was flying almost due west and value didn't change much.
Note 2: Vision safety is disabled as we have an external GPS unit that mounts beneath the downward vision.
View attachment 120332

View attachment 120335

View attachment 120336
Trees have been cleared since this image was taken.
I don't think I've ever held the throttle stick down when my drone was out of sight so I can't add much here. But I think if I ever did that my wife would tell me there was "a loose nut on the stick" and forbid me to ever touch a drone again. But that's just me...
 
As it was returning, he applied down throttle to bring it down as it returned. (This is somewhat standard procedure.) I think he didn't have it in site and flew it into the ground.

Might want to rethink your op procedure. You should always have the aircraft in sight (the pilot). I seem to recall seeing that somewhere. If the RTH function is programmed properly (which should be on your checklist for each flight and checked prior to taking off) then it will function as advertised; at least in the past four years mine over two DJI aircraft and my big bird have never failed me (to date). Applying down stick to an aircraft you can't see is a really ....bonehead move and in this case with predictable results.
 
  • Like
Reactions: Phantomwolf
To clarify my previous statement, it is common to bring it down as it is returning home, but with the implication that it is in view. Otherwise, you let it return at RTH height. Why it wasn't in view at only 0.25 mile, I don't know yet. There are times when you take your eye off of it, or it blends in with the background, and you don't actually see it for a few seconds and you pick it back up because you know the general location and know where to look. If this happens, the operator should stop descending until they have it in view again.
 
To clarify my previous statement, it is common to bring it down as it is returning home, but with the implication that it is in view. Otherwise, you let it return at RTH height. Why it wasn't in view at only 0.25 mile, I don't know yet. There are times when you take your eye off of it, or it blends in with the background, and you don't actually see it for a few seconds and you pick it back up because you know the general location and know where to look. If this happens, the operator should stop descending until they have it in view again.
Correct me if I'm wrong but doesn't the operation of any of the controls while in RTH cancel the RTH and return to manual control? At least that's what my P4 owners manual states.
 
To clarify my previous statement, it is common to bring it down as it is returning home, but with the implication that it is in view. Otherwise, you let it return at RTH height. Why it wasn't in view at only 0.25 mile, I don't know yet. There are times when you take your eye off of it, or it blends in with the background, and you don't actually see it for a few seconds and you pick it back up because you know the general location and know where to look. If this happens, the operator should stop descending until they have it in view again.

I'm just pointing out there are flaws in your process. RTH is autonomous and does not require pilot interaction. It does work. I use it in one form or another every pre-programmed flight we fly. I have had to manually intervene twice in the four years we've been mapping. Both were due to a loss of the GPS signal at the start of the flight. The aircraft on both occasions went into a stable hover until I manually took control and landed it.
Take this as an opportunity to seriously review your processes and rewrite your operations manual then retrain your pilots. No VLOS is a reg violation. Descending without VLOS while in an automated routine would make me seriously question said pilots employment. That was extremely poor judgement on your pilot's part.
I mean this in a constructive way. I am the UAS program manager for facilities. We went through an extensive documentation process before we first took flight for 107 operations. All of my pilots (and myself) are bound by our operations manual and the respective aircraft flight manuals for each of the aircraft we fly. It's never complete and always a work in progress based upon problems we encounter in day to day operations. If it needs to be stated in writing and taught that you don't interfere with an autonomous operation unless there is an emergency, that needs to be placed in print and taught to everyone who operates one of your aircraft.
 
One of my pilots flying a mapping mission (GS Pro) and P4Pv2 went into "Smart Power Go Home" mode. P4P was about 0.25 mile from PIC when in it went into RTL. As it was returning, he applied down throttle to bring it down as it returned. (This is somewhat standard procedure.) I think he didn't have it in site and flew it into the ground.

I pulled the DAT file from AC. I extracted the data using DatCon.
Can someone verify that the Controller:ctrl_thr: D represents his stick input?
RTL started at 875.104sec
Ground impact at ~906 sec - battery was ejected.
There are also 3 "fly_away" messages in log (see screenshot below) that occurred during the decent.
What else should I be looking for?
Note 1: I didn't include GPS:Lat as it was flying almost due west and value didn't change much.
Note 2: Vision safety is disabled as we have an external GPS unit that mounts beneath the downward vision.
View attachment 120332

View attachment 120335

View attachment 120336
Trees have been cleared since this image was taken.
Controller:ctrl_thr: D is the thrust setting that is being used by the FC. Almost always it's the just the control stick input. However if the P4Pro is being flown by a WP mission the control inputs used to implement the mission can be found in the serial_api_inputs record. So in this case Controller:ctrl_thr: D will be a combination of both serial_api_inputs:sdk_thr_z:D and the RC thrust input (which doesn't appear in the P4Pro .DAT).

Looking at your chart my guess is that Controller:ctrl_thr: D reflects just the pilot input. If it were under mission control I would expect it to be constant - at least step-wise constant.

Can you provide the .DAT? I'm kinda interested in the fly away messages in the eventLog stream.
 
Controller:ctrl_thr: D is the thrust setting that is being used by the FC. Almost always it's the just the control stick input. However if the P4Pro is being flown by a WP mission the control inputs used to implement the mission can be found in the serial_api_inputs record. So in this case Controller:ctrl_thr: D will be a combination of both serial_api_inputs:sdk_thr_z:D and the RC thrust input (which doesn't appear in the P4Pro .DAT).

Looking at your chart my guess is that Controller:ctrl_thr: D reflects just the pilot input. If it were under mission control I would expect it to be constant - at least step-wise constant.

Can you provide the .DAT? I'm kinda interested in the fly away messages in the eventLog stream.
Thanks. Sent link via DM.
 
  • Like
Reactions: dewens
Correct me if I'm wrong but doesn't the operation of any of the controls while in RTH cancel the RTH and return to manual control? At least that's what my P4 owners manual states.
No .. moving the joysticks during RTH does not cancel RTH (and there's nothing in your manual that suggests this).
Try it for yourself sometime to find out.

Pressing the RTH button or moving the Flight Mode Switch will cancel RTH.
 
One of my pilots flying a mapping mission (GS Pro) and P4Pv2 went into "Smart Power Go Home" mode. P4P was about 0.25 mile from PIC when in it went into RTL. As it was returning, he applied down throttle to bring it down as it returned. (This is somewhat standard procedure.) I think he didn't have it in site and flew it into the ground.

I pulled the DAT file from AC. I extracted the data using DatCon.
Can someone verify that the Controller:ctrl_thr: D represents his stick input?
RTL started at 875.104sec
Ground impact at ~906 sec - battery was ejected.
There are also 3 "fly_away" messages in log (see screenshot below) that occurred during the decent.
What else should I be looking for?
Note 1: I didn't include GPS:Lat as it was flying almost due west and value didn't change much.
Note 2: Vision safety is disabled as we have an external GPS unit that mounts beneath the downward vision.
View attachment 120332

View attachment 120335

View attachment 120336
Trees have been cleared since this image was taken.
May I ask why you feel it’s standard procedure to lower the altitude of the quad as it’s returning? I wouldn’t even include that in the operating procedure. Of course the drone can be lowered when you have it close to you in sight (duh). I also don’t advocate using auto return all the time as while It works most of the time, it’s that 1% and you’ve lost it. I rely on the map to guide me home keeping the altitude high until it’s over me.
 
  • Like
Reactions: R Martin
Actually if the AC wasn’t in sight, how can you know if he didn’t hit a power line etc on RTH??? Holding the throttle down on RTH just doesn’t make any sense!?!?
 
  • Like
Reactions: R Martin
Actually if the AC wasn’t in sight, how can you know if he didn’t hit a power line etc on RTH??? Holding the throttle down on RTH just doesn’t make any sense!?!?
I think it's a new landing technique where you make a "final approach" and come in on a "glide slope" to land. A technique for stubborn RC airplane pilots who can't stop trying to mix old habits with new technology.
 
To clarify my previous statement, it is common to bring it down as it is returning home, but with the implication that it is in view.

I do this all the time. It's more efficient to create a pseudo "glide slope" than to allow the bird to fly to the RTH point at altitude. That said, I ALWAYS have EITHER line of sight OR I use the camera. I've gotten some really kewl footage that way at the end of a mapping day. But I digress...



Otherwise, you let it return at RTH height. Why it wasn't in view at only 0.25 mile, I don't know yet.

1300 feet out, very possible. My world record is about 2500 under perfect conditions. Under average conditions, 2000' is about where I lose the Phantom. 1300 feet is pretty easy to see the Phantom.




There are times when you take your eye off of it, or it blends in with the background, and you don't actually see it for a few seconds and you pick it back up because you know the general location and know where to look.

Yes. But I'll either look at the camera view or stop descending. ONE TIME I didn't follow my own advice and actually nicked a telephone pole. No damage. No crash. Lucky day.




If this happens, the operator should stop descending until they have it in view again.

Yep. OR...check the camera view.

D
 
Me too prefer descending path on return to me. But NEVER if I don't see it.
The last sequence of the video if you descending on return is much more nice to watch.
But I must admit that many times I lost the sight of the drone when I'm checking the flight parameters. It happen somewhere at 600 - 900 m (depends on light and background). This is for white Phantom, Mavics fade away at shorter distances.
 

Members online

No members online now.

Forum statistics

Threads
143,066
Messages
1,467,355
Members
104,934
Latest member
jody.paugh@fullerandsons.