Hi all.
I could really use some help trying to figure out if Arducopter 2.6 is ever going to work for me.
I have had a total of 4 units 2x genuine 3DR APMs and 2 generic arducopter.
2 buddies also have 1 unit each.
We all have major issues with all units.
to cut a long story short. We can have 3x perfect return to launches, then the the 4th one will result in a toiletbowl effect. Just randomness like that.
I've given up on using ANY failsafe as APM is just too unreliable.
That was a week ago, now things have gotten much worse.
Even stabalize mode does not work. If the quad makes it off the ground it will sometimes dart off in one direction, other times it just falls out of the sky The quad also "pulses"
Some times the quad will just stop responding to the controls and has to be knocked out of the sky to stop it from flying.
Othertimes the unit will just disarm on takeoff.
Have tried multiple recievers, boards etc, NOTHING fixes these issues. Both my buddies have given up on their board and now fly perfectly with another (closed source) autopilot. I would like to get this going correctly but the APM just seems so unstable.
I have attached some logs in case that helps.
Please help!
Replies
I just had a quick look at you log, and one thing I see is that your refresh rate of your ESC is set at 490hz,
I know for a fact that Not all ESC do like this if they are designed for 400hz.
the ESC may work, but are prone to Losing Sync in the air which makes them irresponsible.
Usually that happens on one ESC at the time, which is causing a crash if the drone is not high enough to catch it during the fall (while the ESC is finding sync again if you are lucky)
I have seen this happen in one of my first flight, completely unexpected and already flying around for 5 minutes..
In other words their is simply no quarantee, that even if it looks fine, you will stay out of problems..
However if you reduce the frequency to 400 hz of even lower these sync problems do disappear.
I am flying for at least the last 30 flights with 200hz setting in this parameter.
From a software perspective I do support the feedback from everyone in this thread that it is extremely well thought out.
I flew a mission of 3.5 km around a lake (FPV) on a v3.20 RC7 based on pure trust in the beta version which was released at that time..
If you set your fail safes correctly a jump to 300m could be prevented, by maximizing the altitude to 50m and the circle radius to 50m...
This way you know for sure it will return to the takeoff position, while trying the different settings..
From a support perspective I think that these Developers are extremely helpful with suggestions...
Fact is that you will need to follow up to get to the core of your problems,.,
So if someone is specifying that your VCc may drop below 5v, first step is to correct that...
Best,
Erik
Shane, where are your logs? I'll have a look.
You made a comment about there not being any sanity checks on the baro. I can't recall if there are in 3.1.5... maybe not, but there is in 3.2 for sure.
Thank you Rob.
2014-10-22 08-50-47 57.kmz
2014-10-22 08-50-47 57.log.gpx
2014-10-22 08-50-47 57.log
Auto analysis says
Log File C:/Program Files (x86)/Mission Planner/logs/Shane/2014102208504757.log
Size (kb) 15.4755859375
No of lines 435
Duration 0:00:00
Vehicletype ArduCopter
Firmware Version V3.1.5
Firmware Hash 3c57e771
Hardware Type APM 2
Free Mem 1002
Skipped Lines 0
Test: Autotune = UNKNOWN - No ATUN log data
Test: Balance/Twist = GOOD -
Test: Brownout = FAIL - Truncated Log? Ends while armed at altitude 13.25m
Test: Compass = GOOD - No MAG data, unable to test mag_field interference
Test: Dupe Log Data = GOOD -
Test: Empty = GOOD -
Test: Event/Failsafe = GOOD -
Test: GPS = FAIL - Min satellites: 0, Max HDop: 0.0
Test: IMU Mismatch = UNKNOWN - No IMU log data
Test: Parameters = GOOD -
Test: PM = UNKNOWN - No PM log data
Test: Pitch/Roll = GOOD -
Test: Thrust = GOOD -
Test: VCC = FAIL - VCC below minimum of 4.6v (4.487v)
So voltage drop and no autotune has been performed.
COMPASS_OFS_X 0
COMPASS_OFS_Y 0
COMPASS_OFS_Z 0
So compass has never been calibrated.
Pitch has not been manually tuned.
Roll has not been manually tuned
The IMU data is not being logged but from the baro vs the IMU altitude the vibration isolation is probably not adequate
The Vcc definitely drops below minimum
You'll find the documentation on analyzing the logfiles here. http://copter.ardupilot.com/wiki/common-diagnosing-problems-using-l...
It looks to me like you have not completed even the most basic setup on this vehicle and you are not providing an adequate power supply to the board.
Hi Thanks for the reply.
I 100% did do the compass calibration, so I dont know what happened there.
I also did the magnetic declination manually using the mapping site.
I will go back and check, I have loaded the firmware so many times that maybe I forgot to do the settings on this flight. I was using stabalization mode only, so as to isolate things a bit.
also, I'm using a 3DR power module, so I don't know why there should be a power issue. Should I be using BECS?
Anyway, thank you VERY much for that. I have enough parts here to rebuild the unit and will make sure IMU logging is on and check the calibration.
>>>I also did the magnetic declination manually using the mapping site.
It doesn't look like it from your log file.
COMPASS_AUTODEC 1
COMPASS_DEC 0
The 3DR power module will provide adequate power when you have an adequate power source for it. The logs show that as your throttle goes up that your Vcc goes down which generally means the input from the battery to the power module is not adequate.
>>>I have loaded the firmware so many times that maybe I forgot to do the settings on this flight.
Why are you reloading firmware? It worked correctly the first time you installed it and every time after that.
I hope you can appreciate the developers being sensitive about your comments about the autopilot not being reliable. We're still here helping you but it seems to me that you could be helping yourself by making sure that your vehicle is assembled, wired, and configured correctly.
Thanks for the replies all.
I actually bought the NAZA first, but didn't like the fact that its a closed source product non modifiable product. There is a lot I would like to do with arducopter, if I could just find a fix for my issues.
I am an avid Linux user and have a Ham radio license, so experimenting and fault finding is not new to me. Its hard to see the fault lying in several brand new kits after they worked with NAZA 100% Also how I can have 10-15 perfect flights, then to have the unit flip upside down and crash, etc. I admit I do not understand the logs very well but at the same time, are not getting a lot of support here on the forums. I know everyone donates their time and are busy, so I do not expect everyone to rush out and help me.
It does not help that 3DR shipped me 2 faulty products in a row, then got into some trouble and wouldn't ship me a replacement.
Hardware maybe? get a pixhawk? For example, in alt hold mode while flying FPV, the unit, after flying for about 2 minutes just shot up at full power. I had to manually override after it got to about 300m I'm thinking the code should of picked up the fact that such a dramatic change in baro reading should have resulted in maybe not reacting so violently.
Maybe that's part of the problem. If all parameters are 100% things go fine, it may be that if something goes wrong with a sensor, the code just reacts without much cross-checking?
I fully expect to crash and make mistakes, but I don't seem to be able to fix my issues. I have also seen several people go through the exact same issues. I have the log from my last crash, if anyone is able to help me (and call me an idiot if its warranted) then go right ahead!
Maybe the cows in the paddock we fly in are plotting against me, they look evil.
Did anyone pick up he is using V2.6 ?
Shane, I'm using V3.1.5 on a 3DR frame, ESC's, and motors.
I had a 3DR APM 2.5 in, and had no problems/failures, other than a compassmot of 37.
I wanted a 3DR 2.6, the bottom fell out of our currency, and I had to get a HobbyKing clone.
Loaded with V3.1.5, it still performs flawlessly.
(Also, I needed the APM for a fixed wing project, where compassmot isn't so crucial)
Admittedly, I haven't set up any failsafes, other than battery, but I have flown so many missions, loiters, and RTL's, that it is becoming boring.
That would be in winds ranging from nothing, to 40 Km/H this past weekend.
So, I have to make the statement, that for my specific airframe, the code, and the hardware, is perfect, after "autotune".
A friend of mine still flies with APM 2.5 and V3.1.5, sometimes he has a bit of "toilet bowling", due to bad compassmot.
But he flies missions with me.........
Maybe it is your version of the code, maybe construction, tuning, or hardware problems.
Too many people are hugely successful with APM, to write it off as a "bad" product.
My fixed wing UAV Talon, flying with the old 3DR APM 2.5, also is performing flawlessly.
Just my observations......
Shane, nobody will apparently be able to help you out of your arducopter frustration. If you are happy with a closed commercial system such as Naza, then stick with it. This simply shows some people can't handle and/or do not like integration projects ; and this is why there is a market for closed commercial systems.
That is a rude and inappropriate comment. There is a difference between "integration projects" and beating your head against a wall. Why is it anytime someone identifies a possible problem with the system they "can't handle" it? You think there are no bugs or issues to ever deal with APM?
-
5
-
6
-
7
-
8
-
9
of 17 Next