Disarming when flying in stab mode!

[GoogleCodeExporter]

What steps will reproduce the problem?
1.When flying at 6mtr it automatic disarmed the motors and the X650 felt down 
like a rock.

Used beta version 2.9R7
Crius AIO v1

Original issue reported on code.google.com by voyager4...@gmail.com on 13 Mar 2013 at 10:55

[GoogleCodeExporter]

Had one drop out of the sky.
I was using auto pilot mode on a tablet over Bluetooth.
The quad flew to the waypoint and then the motors cut out.
The HUD on the tablet froze for a few seconds then I tried pushing up the 
throttle on the radio and it took off again so it hadn't disarmed, just froze 
for a few seconds.

Original comment by pk...@sitedesign.net on 13 Apr 2013 at 8:38

[GoogleCodeExporter]

[deleted comment]

[GoogleCodeExporter]

Mine fell out of the sky also a few days ago.
Disarmed in AUTO mode, when only 10 meters away and 10 meters high.
Stabelize and loiter were fine.
Running MPNG 2.9 R6 on a Crius AIOP V1.1 board, ESC's Turnigy 30A Plush
@ Alex, if you want I will send you my log file.
Is it safe to use the R300 version now?

Original comment by Polleke1...@gmail.com on 13 Apr 2013 at 1:55

[GoogleCodeExporter]

Sir Alex is there anything I can do logging wise to help you diagnose the issue?

Original comment by pk...@sitedesign.net on 15 Apr 2013 at 8:08

[GoogleCodeExporter]

I think I may have found the problem.
I have just noticed that some of the quads that I had built for people had got 
mild steel plates in the base instead of aluminium (vibration dampening). This 
caused major compass noise.
I wonder if the extreme compass errors conflicting with the GPS data caused 
overload to the routines?

Original comment by pk...@sitedesign.net on 16 Apr 2013 at 12:15

[GoogleCodeExporter]

" This is take off in stabilize then switch to auto pilot. It heads to the 
first waypoint but is going sideways, it cuts out at or just before the 
waypoint."

I had the same issue the other evening. I was in Stab mode, went to auto to 
preform the same waypoints I had done on a previous flight. The quad went into 
a high speed run to waypoint one and then went in. After a night of searching I 
found it once I had daylight. Luckily (amazingly) only had a broken arm. I 
brought it back and checked it over only to find that all the waypoints that 
were stored in the crius AIO were correct. My quad has no metal plates on it.

Original comment by Escavin...@gmail.com on 19 Apr 2013 at 5:14

[GoogleCodeExporter]

I think a lot of these issues being seen now are compass related.
I have done a bit more testing and found if you do the compass live calibration 
in Mission Planner the values should all be between -150 and plus 150. Around 
90, -10 and +10 would be about right.
If you do the same with the motors running they should not be far away from 
that. If you do see big changes with the motors running then that would cause 
too much compass interference so affect the direction calculations.
I am sure these compass errors are causing the failures.
Also make sure you disable compass learn in the adv param list after 
calibrating the compass as it tries to calibrate during flight which give mad 
figures sometimes.

I will be doing some more instructional videos on YouTube later this week to 
cover this.
http://youtube.com/user/westmidscom/

Original comment by pk...@sitedesign.net on 22 Apr 2013 at 6:53

[GoogleCodeExporter]

I have different problem with CRIUS AIOP V2,i am using on Tricopter ,flys well 
on stabalize mode but when switch to guided mode it does not follow waypoint 
,instead it goes to loiter moder and keeps hovering on home point! Can any one 
tell whats the problem? Firmware 2.8 R3

Original comment by praveen....@gmail.com on 24 Apr 2013 at 10:55

[GoogleCodeExporter]

Well i think i got the same problem. Crius all in one v1.1: MegaPirateNG_2.8_R3 

I got it on video! 

http://youtu.be/kN3FjU58ppA

I was flying FPV in 'Loiter'  mode. Suddenly high in the aire all 4 motor
s stops... 
It look likes it was disarming itself fore some no fucking damn reason :P
This is not and tx/rx issue i have it programmed for failsave. 

A lot of dammage :(

Hope you guys can help me out, i was planning to use this drome on a event. But 
i dont trust flying in anymore. 

Original comment by DroneSh...@gmail.com on 25 Apr 2013 at 12:11

[GoogleCodeExporter]

While you were in Loiter did you touch the rudder stick? Mine disarmed in 
Loiter when I yawed to the left.

Original comment by ajrma...@gmail.com on 25 Apr 2013 at 3:23

[GoogleCodeExporter]

[deleted comment]

[GoogleCodeExporter]

Now with and working link!

http://youtu.be/kxvGvuRCu60

RE: 

Well i think i got the same problem. Crius all in one v1.1: MegaPirateNG_2.8_R3 

I got it on video! 

http://youtu.be/kxvGvuRCu60

I was flying FPV in 'Loiter'  mode. Suddenly in the aire all 4 motors stopped 
working... 
It look likes it was disarming itself fore some no fucking damn reason :P
This is not and tx/rx issue i have it programmed for failsave. 

A lot of dammage :(

Hope you guys can help me out, i was planning to use this drome on a event. But 
i dont trust flying this quad anymore. 

Original comment by mast3rw...@gmail.com on 25 Apr 2013 at 4:03

[GoogleCodeExporter]

Is there any work being done on this disarming issue with 29r300? No new posts 
from Alex for a month or more now.

Original comment by ajrma...@gmail.com on 30 Apr 2013 at 7:52

[GoogleCodeExporter]

It is the biggest issue!
Why is it only MEDIUM priority?

Original comment by moonsurv...@gmail.com on 3 May 2013 at 9:36

[GoogleCodeExporter]

Guys,

I do not think anyone is working on this issue now ...

It seems to be solved by Alex first, but as I can see from your posts - the 
problems is coming back from time to time.

What would be my advice for those who are still getting the disarming issue is 
to try this port of mpNG >>>> http://code.google.com/p/megapirateng-add-ons/ 
(use SVN or file download)

It's logicaly the same as rev.300 but has some important code changes (or in my 
private opinion improvements - resulting with lower CPU load) .. and what's 
important - it never disarmed.

I have tested this build on AIO v.1 with PPM_SUM receiver on A8 and with sonar 
disabled - as described.

>>> That's the only configuration you will get serious CPU load benefits over 
rev300 <<<<

For AIO v2 you will get replaced PPM decoder only - I am not sure if it helps 
on disarming as I can not test this config.

I have no reports on disarming from people downloading this build for weeks now 
but have no idea what config they use :)

Original comment by pa...@kulma.pl on 5 May 2013 at 10:09

[GoogleCodeExporter]

any news about the fixing?

Original comment by 78mar...@gmail.com on 16 May 2013 at 7:20

[GoogleCodeExporter]

Used megapirate-add-ons on Crius v2, was in loiter moved to stab, all motors 
cut... crash... while falling (7m or so) tried changing modes, increasing 
throttle, nothing.

One important thing is that while on the ground, I had throttle down by 
instinct but I accidentally touched it and the motors started moving without 
arming them, so its not actually disarming, maybe a freeze/hang, not reboot as 
it would have come back disarmed. I used PPM on PL1 (for v2) when I should have 
used PPM on A8 to disable sonar and free CPU cycles (had it configured and not 
yet pushed to FC), I didn't touch any log related stuff, maybe I have a log, I 
will look today.

Original comment by shays....@liveperson.com on 5 Jun 2013 at 8:46

[GoogleCodeExporter]

My quadcopter also drops out of the sky. When in stabmode and hovering all 
motors stops. Dont know what is wrong. Motors weren't disarmed!
Use MPng 2.9.1 (88) with add-ons.

Original comment by voyager4...@gmail.com on 11 Jun 2013 at 8:58

[GoogleCodeExporter]

Have sent you an email.
Paku

Original comment by pa...@kulma.pl on 11 Jun 2013 at 9:54

[GoogleCodeExporter]

Any issues with 2.9 R5? I have it running on two hexes and no problems but 
missions are important and just want to make sure it is a solid version.

Cheers

Original comment by bigsky...@gmail.com on 4 Jul 2013 at 4:21

[GoogleCodeExporter]

it happend to me too. First 4 flights no problem. only switching from stabilize 
to pos hold and pid tunning. the 5th flight was with badluck. i took off and 
went away and high becouse i intended to put my goggles but the motors stopped 
and the quad dropped from 30 m like a rock. happy that only one arm broken. 
after reashearching found that i am not the only one. and today in enterd the 
mpng add ons and found that they have a update from 5 days and they say the 
disarm problem identified and corrected . should i try it or flash the mpng 
2.8? or just go tu multiwii becouse of it's more stable ? i have a aiop 1.1 . 
this is the question :) thanks ans sorry for my bad english. :P 

Original comment by djliv...@gmail.com on 7 Jul 2013 at 4:55

[GoogleCodeExporter]

My tricopter felt down while being at Alt Hold mode. I dont know if it was 
disarmed, but suddenly the motors stopped working. My setup was Crius AIO v1.0 
with MPNG 2.9R7 (rev 300) with standard receiver. 

Original comment by josemanu...@gmail.com on 21 Jul 2013 at 4:56

[GoogleCodeExporter]

Hy there!

So, in the end, whats going on?! Should we all forget the MPNG-Project and try 
finding another as an alternative (maybe MWC or the like) as not everybody is 
able to debug or even rewright the code?
This is not meant as an attack to Alex: hats off to so much work invested! But 
it should be a good manner to inform the users which are sitting on hot coals 
what is next and if the development is going further in the next time or if it 
simply stoped.

And now back to topic: 
I got my criusV2 just 4 weeks ago and finaly I've got it working. But from time 
to time a got a slightly different problem: if I switched in Mid-Air from 
stable to AH, the motors where spinning down but did not turn off. As I was 
around 6 meters above ground, I've managed to switch back from AH to stable and 
just some centimeters bevor the copter hits the ground the motors where 
spinning up and it was stable again - phew...
As it was flying I could not see if the LEDs where showing "disarmed" while 
falling.
One more thing was the following: as flying in Loiter mode I can turn the 
copter cw and ccw but when I turn it cw (to the right) it starts sinking and if 
I turn it left it climbs to the previous altitude - no turn off or disarming yet
Most things are working pretty fine, RTL, Loiter, AH, PH... Simple mode does 
some crazy thing indeed, but there should be more investigation of me.

My setup:
Hexa (x-style), 20A SimonK (flashed myself), MPNG r300, Graupner 35MHz Rx with 
diy-ppm (as my frsky d4r-ii got some problems after reflashing it to 27ms)
And Yes! I would take the risk for some beta-testing with the hope in getting a 
good photo copter in the end.

Just my 2 cents to the whole situation (esp. for the crius boards)
I find it unfortunate that the pcb is still under lock and seal. As some guys 
reports disarming while the birds are on desk, it should be possible to go 
deeper when there where more informations about it. Maybe you could use another 
pin for debugging light or for measuring or what ever.
And let me say it as I'm not a lover of the far eastern business practices - if 
they want to rebuild it, they will do it! So, what is the intention of holding 
it closed?

Original comment by d.schulz...@googlemail.com on 6 Aug 2013 at 10:31

[GoogleCodeExporter]

I'm currently working on porting 3.0

Actually, this problem not solved in 2.9. looks like it performance issue, but 
nobody knows whats exactly causes motors to stop.

Paku made some cuts in code(removed some parts, not used by all) and seems it 
help, but... since problem not 1005 sorted out, I cant say it's 100% was fixed.

So latest stable version still 2.8
Currently I'm working on MPNG 3.0 (most parts of code rewritten, so I hope we 
will not see this bug).

Project news and updates available on home page of this site or at RCGroups:
http://www.rcgroups.com/forums/showthread.php?t=1382503

Original comment by Alexey.K...@gmail.com on 6 Aug 2013 at 10:42

[GoogleCodeExporter]

Awesome work Alexey as usual. I'm running the latest revision and have still 
had the motors disarm - but only once while switching out of auto to stab. 
Earlier revisions I had it many times. I'm looking forward to your port of 
arducopter 3 very much.

Original comment by netwimbe...@gmail.com on 6 Aug 2013 at 11:51

[GoogleCodeExporter]

Ah, hello Alexey, nice to meet you, that's a good statement to see you are 
still alive ;)

As I stumbled through the changelist I've noticed that there are many small 
changes in code - are this bugfixes or are this sometimes other improvements 
apart from ardupilot?

Original comment by d.schulz...@googlemail.com on 6 Aug 2013 at 3:52

[GoogleCodeExporter]

Alex:

It's sorted out.

Could be I did not expressed it directly but the fault was in the scheduler 
function calculating main loop lock-out time. Code expected two time values to 
be always set so first time is before second (T2-T1 always >0) while it was not 
always a case as code was interrupt driven and some sheduler jobs takes LONG 
times and from time to time T2-T1 < 0. Next, as you may know, T1 and T2 was 
defined and calculated as 32 bit integers so the negative numbers were 
represented as integer, mostly big one, always bigger then 2 second scheduler 
lock-out limit. In such case scheduler just stops the motors - as expected 

Fault is inside original ArduCopter code, had some talks with AC guys on it. It 
could be the fault showed up in mpNG code as there are much more interrupts and 
longer running drivers around scheduler work 

Anyway inside my build faulty code is "secured" (checks for reasonable times 
values) and interrupts number is limited a little - had no reports on disarms 
for last build.

All the case is solved here in line 53

http://code.google.com/p/megapirateng-add-ons/source/diff?spec=svn160&r=158&form
at=side&path=/trunk/ArduCopter/failsafe.pde&old_path=/trunk/ArduCopter/failsafe.
pde&old=2

Have fun working on 3.0 :)

Paku

Original comment by pa...@kulma.pl on 6 Aug 2013 at 5:40

[GoogleCodeExporter]

@paku

Thanks for your work, but: on 6.Aug you wrote "Its sorted out" - but if I take 
a look to the sources the last changes made are from 1.Jul. So what took this 
time?

Original comment by d.schulz...@googlemail.com on 14 Aug 2013 at 12:32

[GoogleCodeExporter]

His code located here: http://code.google.com/p/megapirateng-add-ons/source/list

Original comment by Alexey.K...@gmail.com on 14 Aug 2013 at 12:35

[GoogleCodeExporter]

It's sorted out by the build committed on July 1.

Sounds better :)

And yes pls do not "spam" Alex forum, if you have any questions ask me using 
link above.

Original comment by pa...@kulma.pl on 14 Aug 2013 at 2:04

[GoogleCodeExporter]

I had it happen to me yesterday, on my hexacopter, with my new brushless 
gimbal, over the frikken sea!!!

All electronics totally fried.  Circa $800 damage..  Gutted.

Original comment by pcburn...@gmail.com on 19 Aug 2013 at 12:13

[GoogleCodeExporter]

Sorry, some more details, it happened when I switched from stabilise to loiter, 
all motors stopped.  Bill has now hit $1100..  

Original comment by pcburn...@gmail.com on 19 Aug 2013 at 4:33

[GoogleCodeExporter]

@pcburn*:

More details would be interesting: what flight controller, which firmware 
version, which kind of ESC is used. What RC do you use: sumsignal or single 
channels?

Original comment by d.schulz...@googlemail.com on 19 Aug 2013 at 7:18

[GoogleCodeExporter]

More details don't really matter in this case. The big is with the current 
release and previous release of MPNG 2.9. You can either use Pakus build or you 
run the risk of getting motor switch off. Most of the time it happens while 
switching flight modes (exactly as described).

I know what its like to lose $1100 worth of flying machine, in this case you 
should have probably researched a bit and known about this bug (or it should be 
mentioned in red on the main page) I would not fly over water knowing this bug 
exists even if I had never had it happen to me (I have, 3 times).

Arducoter 3 is busy being ported now so I don't think they will bother to fix 
this in the MPNG original source. 

Original comment by glenwimb...@gmail.com on 19 Aug 2013 at 7:33

[GoogleCodeExporter]

This is okay, but I think it should be a matter of good manners that the base 
details of the "flying system" are mandatory if someone is searching for help!
Especially on a system with many possible hardware setups and many software 
versions!

Original comment by d.schulz...@googlemail.com on 19 Aug 2013 at 7:41

[GoogleCodeExporter]

Witespy Multiwii Pro Ezy 3.0
MPNG 2.9
Skywalker 40A ESCs
Spektrum DX8.

But as Glenwimb stated, this is all fairly irrelevant when the issue has 
already been identified.

@Glenwmb, where would I be able to get a copy of Paku's build?

Original comment by pcburn...@gmail.com on 13 Sep 2013 at 2:48

[GoogleCodeExporter]

Same problem here!
Happened to me last night flying in Stab mode using 2.8_r3 (stable?). Luckily 
there was no harm done.

I'm using a HK Multiwii Pro with GPS, and a Bluetooth module to get the 
telemetry. I was thinking: It's known that when connecting the FC through USB a 
reset occurs. Does the same happen when connecting through Bluetooth or 3Dr? 
What if the Bluetooth dongle loses connection with my phone, and the phone 
tries to reconnect and resets the FC while flying?

Just an idea.
I'm not really confident flying now...

Original comment by mattogo...@gmail.com on 7 Oct 2013 at 4:20

[GoogleCodeExporter]

Same problem here.
I am running 2.8_R3 and never had this issue before, but due to earlier crash I 
had to replace my GPS from original MTK (nmea) to NEO-6M (ublox).
After this change the quad suddenly cuts motors in mid air, luckily only from 1 
meter above ground.
Could it have something to do with the ublox GPS setting?
Running on a HK Multiwii Pro and Bluetooth module.

Original comment by andersso...@gmail.com on 10 Oct 2013 at 12:48

[GoogleCodeExporter]

Still thinking it's a Bluetooth connection issue. Everybody who experienced 
this had a Bluetooth module or a 3D radio?

Original comment by mattogo...@gmail.com on 10 Oct 2013 at 1:45

[GoogleCodeExporter]

This issue already fixed by Paku.

Everybody can download fixed version from: 
http://code.google.com/p/megapirateng-add-ons/source/list

I'm closed support of 2.9 and this site (sources moved to GitHub), since I'm 
working on 3.x

Original comment by Alexey.K...@gmail.com on 10 Oct 2013 at 1:48

[GoogleCodeExporter]

The fixed version is 2.9? I crashed using 2.9, so I'd like to stick to 2.8. Any 
solution?
Happened to me again last night (twice) :-(

Thanks!

Original comment by mattogo...@gmail.com on 15 Oct 2013 at 3:47

[GoogleCodeExporter]

@mattogo... : No, a partly fixed version is from paku but his code is on 
another repo: http://code.google.com/p/megapirateng-add-ons

Alex is working on porting the the 3.*.* to MPNG - and there were many 
improvements, especially to the timing issues was payed attention to.

Original comment by d.schulz...@googlemail.com on 17 Oct 2013 at 8:31

[GoogleCodeExporter]

Little warning on my build.

I had some reports on motors spikes using Std Rx configuration. 
http://code.google.com/p/megapirateng-add-ons/issues/detail?id=2

I am not going to fix it as most use CPPM and Alex is working on 3.0.

No problems with CPPM Rx config as for now. 
Std Rx /Spikes problem is not so frequent as well (few reports for hundreds 
downloads)

Original comment by pa...@kulma.pl on 17 Oct 2013 at 9:44