Welcome Inspire Pilots!
Join our free DJI Inspire community today!
Sign up

Firmware Bug Safety Issue - (Ver 1.04 and earlier)

Joined
Jan 3, 2015
Messages
40
Reaction score
15
Age
61
When using the dual controls I have found that the camera operator using the slave controller is able to select "take off" from the screen. It doesn't always do it. It seems if you go into the controller settings and reselect slave mode then go back to the main screen, the slave controller can then engage the take off and land controls on the screen.

This is a safety concern. As the rotors can be engaged while the pilot is not in control. You could be changing a battery and have someone start the rotors inadvertently.

I have tested this with another Inspire using different tablets with the same result. This was tested with both controllers using iPads as well as with an iPad on the master and the slave using Android with the same results.

This has been tested with the current version as well as 1.03 with the same result. This has been reported to DJI but they just want the unit sent in for repair. Testing shows that this is a firmware/app issue and not related to particular hardware.

Any beta testers out there that can replicate and report it before we have an unplanned take off?

B Dog
 
  • Like
Reactions: JoffaDan
I understand the point you are making but cannot see how someone would start the motors during a battery change! While you are swapping the battery, you cannot possibly start the motors as the battery needs to be switched on first.
Secondly, the Inspire will not start unless it has gone through its initialization sequence and the motors will make the usual rising tones sound alerting anyone that the Inspire is 'live'.
Thirdly, assuming you have a checklist, why would your payload operator even attempt to call up the auto take off dialogue and then further confirm the action with a swipe?
Is it not a serious safety issue and bug that while you are driving on the freeway at 70mph your passenger could reach the gear selector and push it to neutral if they wanted to or maybe reach over and kill the engine?... No because you tell them not to do it!
 
  • Like
Reactions: Nic
It seems if you go into the controller settings and reselect slave mode then go back to the main screen, the slave controller can then engage the take off and land controls on the screen.
This is a serious issue. When you say you go to the controller setting is this on the slave controller? Is this related to the firmware or the latest GO app (2.2.0 iOS)?
 
I understand the point you are making but cannot see how someone would start the motors during a battery change! While you are swapping the battery, you cannot possibly start the motors as the battery needs to be switched on first.
Secondly, the Inspire will not start unless it has gone through its initialization sequence and the motors will make the usual rising tones sound alerting anyone that the Inspire is 'live'.
Thirdly, assuming you have a checklist, why would your payload operator even attempt to call up the auto take off dialogue and then further confirm the action with a swipe?
Is it not a serious safety issue and bug that while you are driving on the freeway at 70mph your passenger could reach the gear selector and push it to neutral if they wanted to or maybe reach over and kill the engine?... No because you tell them not to do it!

Ed. Everything's a safety issue when in the wrong persons hands. What I find a problem with here is the response from DJI. This is something that should not be allowed to happen no matter how stupid the second operator may be. "Send it in for service"....wrong answer!
 
Ed. Everything's a safety issue when in the wrong persons hands. What I find a problem with here is the response from DJI. This is something that should not be allowed to happen no matter how stupid the second operator may be. "Send it in for service"....wrong answer!
I know - having a bad day here. I come from 30 years of R/C flying and have seen fingers ripped to shreds by people flick starting glo plug engines and hands severed by 700 size heli blades but nobody shouted " Its a serious bug, its the manufactures fault we're all going to die - we're doomed"
No, we just wrapped up the guys hand, drove him to emergency and said he will be more careful next time.
This can be a dangerous or even fatal hobby! Unfortunately you can never legislate or foresee stupid in all circumstances.
I will get off my soapbox now and disappear back into my man cave (and probably wait for the abuse). :rolleyes:
 
Ed. Everything's a safety issue when in the wrong persons hands. What I find a problem with here is the response from DJI. This is something that should not be allowed to happen no matter how stupid the second operator may be. "Send it in for service"....wrong answer!
I have spent some time trying to replicate the fault you described as it would be a safety issue. Everything I tried failed to allow me to allow me to arm the motors or take off using the Slave unit with the Master connected. I tried turning the Slave RC to Off and back to Slave. The closest I got was setting the Slave RC as Master which turned the Auto Takeoff button on. Pressing this button displayed the Auto Takeoff dialog with the execute slider. However, moving slider had no effect and the function timed out. The aircraft did not arm or take off. Is this what you observed?
 
When using the dual controls I have found that the camera operator using the slave controller is able to select "take off" from the screen. It doesn't always do it. It seems if you go into the controller settings and reselect slave mode then go back to the main screen, the slave controller can then engage the take off and land controls on the screen.

This is a safety concern. As the rotors can be engaged while the pilot is not in control. You could be changing a battery and have someone start the rotors inadvertently.

I have tested this with another Inspire using different tablets with the same result. This was tested with both controllers using iPads as well as with an iPad on the master and the slave using Android with the same results.

This has been tested with the current version as well as 1.03 with the same result. This has been reported to DJI but they just want the unit sent in for repair. Testing shows that this is a firmware/app issue and not related to particular hardware.

Any beta testers out there that can replicate and report it before we have an unplanned take off?

B Dog
I have tried a number of times now to duplicate this issue you state you have, with no success. There is no way I can start the motors or effect a take off or landing of the craft with the "slave RC".
I can bring up the auto take off slider but nothing happens when I use the slide control, it just times out.
 
  • Like
Reactions: Scotflieger
I know - having a bad day here. I come from 30 years of R/C flying and have seen fingers ripped to shreds by people flick starting glo plug engines and hands severed by 700 size heli blades but nobody shouted " Its a serious bug, its the manufactures fault we're all going to die - we're doomed"
No, we just wrapped up the guys hand, drove him to emergency and said he will be more careful next time.
This can be a dangerous or even fatal hobby! Unfortunately you can never legislate or foresee stupid in all circumstances.
I will get off my soapbox now and disappear back into my man cave (and probably wait for the abuse). :rolleyes:
I agree,
 
I understand the point you are making but cannot see how someone would start the motors during a battery change! While you are swapping the battery, you cannot possibly start the motors as the battery needs to be switched on first.
Secondly, the Inspire will not start unless it has gone through its initialization sequence and the motors will make the usual rising tones sound alerting anyone that the Inspire is 'live'.
Thirdly, assuming you have a checklist, why would your payload operator even attempt to call up the auto take off dialogue and then further confirm the action with a swipe?
Is it not a serious safety issue and bug that while you are driving on the freeway at 70mph your passenger could reach the gear selector and push it to neutral if they wanted to or maybe reach over and kill the engine?... No because you tell them not to do it!
I agree with this, if your second controller camera operator is that stupid you need a new second controller operator immediately. There is no possible scenario where he should even be attempting to do such a thing anyway.
 
I agree with this, if your second controller camera operator is that stupid you need a new second controller operator immediately. There is no possible scenario where he should even be attempting to do such a thing anyway.
I'm afraid the world is full of stupid. If it can happen someone will do it. Neither @Derfmen nor I were able to reproduce @B Dog's reported problem other than being able to call up the non-functional auto launch dialog on the Slave unit.
 
I'm afraid the world is full of stupid. If it can happen someone will do it. Neither @Derfmen nor I were able to reproduce @B Dog's reported problem other than being able to call up the non-functional auto launch dialog on the Slave unit.
true enough. it seems most bugs reported here even if occurring to many people arent occurring to everyone. there are many factors, which app version is used on either android or apple, firmware version, inspire model version and so on, any number of those different possibilities would have to match perfectly to make the bugs 100% reproducible. At any rate even if it is a legit bug it sounds mostly like a non issue as no second controller operator would/should ever even attempt to activate that feature in the first place.
 
  • Like
Reactions: Scotflieger
have seen fingers ripped to shreds by people flick starting glo plug engines
That brings back some good old memories. I still miss the connection with a couple of nerves in some fingertips. Not good for guitar or piano playing. (among a lot of other things).
 
  • Like
Reactions: The Editor
I agree with this, if your second controller camera operator is that stupid you need a new second controller operator immediately. There is no possible scenario where he should even be attempting to do such a thing anyway.
I'm fortunate to have a buddy who's also RPAS certified and flies an inspire. We team up when needed and both are pilot, observer or gimbal controller, depending on who's I1 is used.
It's great to have 2 pairs of pilot eyes and skills at the more complex jobs.
 
I understand the point you are making but cannot see how someone would start the motors during a battery change! While you are swapping the battery, you cannot possibly start the motors as the battery needs to be switched on first.
Secondly, the Inspire will not start unless it has gone through its initialization sequence and the motors will make the usual rising tones sound alerting anyone that the Inspire is 'live'.
Thirdly, assuming you have a checklist, why would your payload operator even attempt to call up the auto take off dialogue and then further confirm the action with a swipe?
Is it not a serious safety issue and bug that while you are driving on the freeway at 70mph your passenger could reach the gear selector and push it to neutral if they wanted to or maybe reach over and kill the engine?... No because you tell them not to do it!

I think you missed the point of my post. The fact is that the firmware allows this to happen. When I say while changing the battery I mean the whole process. So you've just flown and are just getting ready to turn the battery off (so it's still on and doesn't need to initialize) when someone plays with the slave remote and fires it up thinking that it can only control the camera. Of course if you always have the same payload operator this should never happen but in reality sometimes at a scene I have new payload operators that I can't necessarily guarantee that they won't play with the controls.

The analogy to a car is totally not the same. In that analogy I would have to have the passenger having their own pedals and steering wheel that aren't supposed to operate the car but they might... In your analogy my payload operator would have to grab the pilot controller out of my hands. Not the same thing.

But I digress, I have dealt with firmware bugs for many years and the fact it that you should not be able to put the device into a mode that has no reward but causes a risk. Regardless of the chance of it happening - it should never happen. There are plenty of examples of lawsuits for devices causing injury when the device did something dangerous after someone did something that no person should have ever done - but did. Did the manufacturer do everything they reasonably could to prevent the accident? In this case it's just a firmware fix, so I'd say they haven't.

I have tested the bug again and it does actually take off. Here is what you need to do in order to test it. This bug was discovered after trying to confirm that the slave was in fact in slave mode and then seeing the take off option was enabled.

Here is how to test this:

Connect Ipad to main.
Turn on main controller.
Connect S5 or Ipad to slave.
Turn on Slave Controller.
Turn on Flier Try to take off with slave application - will not work.
Go to controller settings on slave S5 or Ipad and select Slave mode (even though it's already in Slave mode).
Press X to close.
Try to take off, You can now slide to take off.
Flier lifts off to 5ft. press button again to land. Press ok to land Flier lands is place.
 
regardless, the chances of a random camera operator doing that are slim to none and you shouldnt be so worried about it. If youre hiring camera operators who dont understand that they have zero expectation or reason to ever try and control the aircrafts flight and they randomly play with the settings such as you describe and then still try to control flight functions id take a hard look at your hiring process and maybe change a few things.
 
It seems if you go into the controller settings and reselect slave mode then go back to the main screen, the slave controller can then engage the take off and land controls on the screen.

Just be happy thats fantastic! If it really works you could send someone with the 2nd remote somewhere away and land it there safe then.

However you have to test if it really works outside the 5.8 Mhz range.
 
regardless, the chances of a random camera operator doing that are slim to none and you shouldnt be so worried about it. If youre hiring camera operators who dont understand that they have zero expectation or reason to ever try and control the aircrafts flight and they randomly play with the settings such as you describe and then still try to control flight functions id take a hard look at your hiring process and maybe change a few things.

I'm not in fact worried that it is actually likely to happen to me or one my team. My main point is that any programming bug like this that can be recreated is easy to fix and shouldn't be there in any professional product. Could it happen to someone out flying with their kid who plays with the slave remote? Possibly, you can't control every possibility but you can reduce the chance but fixing bad code like this. If the feature had a use then you would mitigate the risk but if it has absolutely no use then it is a bug and should be fixed. As a maufacturer I would want to know get this feedback and fix the issue in the next version.
 
I'm not in fact worried that it is actually likely to happen to me or one my team. My main point is that any programming bug like this that can be recreated is easy to fix and shouldn't be there in any professional product. Could it happen to someone out flying with their kid who plays with the slave remote? Possibly, you can't control every possibility but you can reduce the chance but fixing bad code like this. If the feature had a use then you would mitigate the risk but if it has absolutely no use then it is a bug and should be fixed. As a maufacturer I would want to know get this feedback and fix the issue in the next version.
fair enough, however no one else was able to recreate it they said. Not to mention this isnt the dji official forum and dji is rarely on their own forum, highly doubt any of them are on this one. Besides theres a billion things dji should worry about fixing or adding before they worry about this bug in my opinion, but none of them are likely to ever get fixed.
 
fair enough, however no one else was able to recreate it they said. Not to mention this isnt the dji official forum and dji is rarely on their own forum, highly doubt any of them are on this one. Besides theres a billion things dji should worry about fixing or adding before they worry about this bug in my opinion, but none of them are likely to ever get fixed.

I'm afraid you're probably right. I was hoping that one of the firmware beta testers on here that have a contact in DJI could include it in a report. As you say DJI support doesn't work to address all sorts of issues so why should this one be any different. Thanks for your comments.
 

Members online

No members online now.

Forum statistics

Threads
22,277
Messages
210,655
Members
34,324
Latest member
Charlesssouth