IRIS+ flew off while in loiter - logs linked - IRIS

Hi All,

New IRIS+ owner here that's *really* been enjoying flying for the past several days. I've been really impressed with the Iris+, and am having a lot of fun learning.

I haven't had any real problems (other than feeling a little underpowered with the gimbal / gopro) until today. While I was flying it around the yard today in loiter mode (very slow flight, low altitude), it just shot off to the right about 40+ feet and hit a tree in my neighbors yard.

I completely lost all control, and the Iris was not responding to any input from the remote. I've verified that it was in loiter mode, batteries were good, I wan't more than 20 feet from the unit, so the RF signal should have been fine.

From what I can tell so far, I ended up with a broken arm (right rear), and maybe a broken gimbal (it was twitching a lot - it seems better now, but I won't know for sure until i fly again).

I believe I followed the right steps to get the last flight logs from the unit ( if not, please forgive the newb here - I promise I searched and spent time reading before posting :-) )

If anyone with more experience that I have so far would take a look and offer any input on what went wrong, I'd love to know. I've linked to a zip file below.

Logs: https://www.dropbox.com/s/pwaeiqiq5dkm6g8/2014-11-08%2016-31-22.log.zip?dl=0

Any questions, or anything else I can do the help with the diagnosis, please let me know.

Thanks!

-Ed

Ed,

May as well give support a quick call and see how they want to handle the glitch, or not.

In the arducopter code there's a different subsystem and fail safe error code for losing GPS lock and regaining it than what you experienced:

http://copter.ardupilot.com/wiki/common-diagnosing-problems-using-l...

What happened to you is logged by the code literally as "glitch" - which is pretty vague. Do you have a faulty GPS module? If you replace an arm (I've replaced many Iris arms, it's slow going but easy enough) and put her in the air, how likely is it that your GPS module or whatever 'glitched' today will glitch again?

We've experienced low quality GPS signals - not seeing enough sats to have an HDOP that provides for accurate, safe loiter / alt hold / auto flying. What you experienced does not appear to be that, at all. As SkyRover points out, after the glitch you have the same sat number and the same good HDOP you had previously.

I think you could make the case to 3DR support that if they cannot give you a reason for the "glitch" after reviewing the logs, then your confidence in the hardware they sold you is low enough to warrant a refund or replacement.

Most GPS units we've seen on UAVs aren't crammed into such a tight space as the Iris right next to other electronics, and even with the little shield they added to the Iris+ there's a certain level of doubt that it's a good idea to have it all that close. We do fly an original Iris (and the one before that in beta) and *most* of the time, she's solid. We've had a crazy fly away. We had abysmal or erratic HDOPs with the very first unit when compared to an APM quad sitting yards away in the same open field.

Look at the masts they use on the Y6 (which we also fly) and the X8 - they *could* slap the GPS module where the mast is attached, but they don't...

Replies

- aforsyth > Ed H November 9, 2014 at 7:52am
  
  It is possible that you would have been able to switch to stabilize and bring it under control. Tricky to re-orientate your perspective to "front" during the confusion of a fly away, while also making that split second decision of "do I leave the throttle in the middle and bring it under control, or just try to go UP and avoid collisions, and which way do I throw the stick to stop its mad dash?"
  
  I think that it's a natural thought process to fly low to avoid coming down from a great height and therefore doing more damage to property and UAV - not to mention wanting to get back on the ground when there's a problem - but my business partner pointed out to me when we were flight testing fixed wing UAVs that altitude often saves you from a stall or a bad turn because you've got time to correct. Either way the "open space in all directions" idea helps :) but we won't always be flying in spaces like that. When it doesn't seem like something is under control and you're fighting to regain, going UP will avoid hitting trees and houses and people, but it's counter intuitive to a degree.
SkyRover November 8, 2014 at 3:51pm

yup, definitely a gps glitch. even managed to pick back up several sats before it came down
aforsyth November 8, 2014 at 3:22pm
- Ed H > aforsyth November 8, 2014 at 3:30pm
  
  Thanks, aforsyth, for the quick response.
  
  Is this something I should call 3DR support about, or is this just par for the course?
- aforsyth > Ed H November 8, 2014 at 4:16pm
  
  Ed,
  
  May as well give support a quick call and see how they want to handle the glitch, or not.
  
  In the arducopter code there's a different subsystem and fail safe error code for losing GPS lock and regaining it than what you experienced:
  
  http://copter.ardupilot.com/wiki/common-diagnosing-problems-using-l...
  
  What happened to you is logged by the code literally as "glitch" - which is pretty vague. Do you have a faulty GPS module? If you replace an arm (I've replaced many Iris arms, it's slow going but easy enough) and put her in the air, how likely is it that your GPS module or whatever 'glitched' today will glitch again?
  
  We've experienced low quality GPS signals - not seeing enough sats to have an HDOP that provides for accurate, safe loiter / alt hold / auto flying. What you experienced does not appear to be that, at all. As SkyRover points out, after the glitch you have the same sat number and the same good HDOP you had previously.
  
  I think you could make the case to 3DR support that if they cannot give you a reason for the "glitch" after reviewing the logs, then your confidence in the hardware they sold you is low enough to warrant a refund or replacement.
  
  Most GPS units we've seen on UAVs aren't crammed into such a tight space as the Iris right next to other electronics, and even with the little shield they added to the Iris+ there's a certain level of doubt that it's a good idea to have it all that close. We do fly an original Iris (and the one before that in beta) and *most* of the time, she's solid. We've had a crazy fly away. We had abysmal or erratic HDOPs with the very first unit when compared to an APM quad sitting yards away in the same open field.
  
  Look at the masts they use on the Y6 (which we also fly) and the X8 - they *could* slap the GPS module where the mast is attached, but they don't...
- Ben Norris > aforsyth November 13, 2014 at 5:59am
  
  HDOP is not something you can just faithfully rely on. Arducopter should be filtering out unrealistic position changes too.
- Ed H > aforsyth November 9, 2014 at 7:22am
  
  Thanks for all the detail - I appreciate it! :-)
  
  I'm going to give 3DR support a call on Monday and see what they say - this has actually happened twice now (see update to main thread).
  
  You're right - I am concerned about the GPS glitching again. Based on what I saw yesterday, I'd be scared to fly it anywhere but a wideo open field.
  
  thanks.
aforsyth November 8, 2014 at 3:21pm

Something glitched with your GPS, the HDOP spiked and it tried very very hard to go where it thought it suddenly needed to be.

Adding a screen shot. Have you used Mission Planner much? It has canned graphs to look at GPS, mechanical and other log info to troubleshoot failures.

of 2

This reply was deleted.