A2 Problems Catalogue

MombasaFlash

Heli's & Tele's bloke
Hi has any one able to figure out the possible specific problem with the A2?


Amidst the deafening silence from DJI, and in an effort to breathe some new life into the subject as the "A2 FC Firmware 2.3 - Here We ALL go Again" thread is not shedding any clear light on it all, perhaps we could try to assemble all the failure and crash reports into one place in order to try and narrow down the circumstances under which folks have been having unexplained issues with the infamous A2 Flight Control.

Maybe all those who have suffered could list the conditions and gear they were using at the time of their incident.


Please feel free to add pertinent categories

1. Which firmware version?
2. What radio transmitter and was it a duplex with 2.4G telemetry?
3. Was the A2 built-in receiver used or bypassed. If it was bypassed, which receiver was used?
4. What Flight Mode was engaged when things got funky?
5. Were any other onboard transmitters being used at the time?
6. Was it normal transmitter flight or programmed waypoint flight?
7. ......?
 

ACP

Member
Where do I start.
OK
EVO s800
1. V2.3
2. Jeti DC-16 duplex 2.4g telemetry
3. A2 built-in receiver bypassed Duel RSat2 PPM
4. GPS
5. Dji 5.8G Video TX
6. Norman GPS
7. FLIP OF DEATH at 10m altitude.
Smashed 3 booms, and 4 Props.
Smashed BeSteady one
Sony NEX7 don't work no more....


Cinistar 8

1. V2.3
2. Jeti DC-16 duplex 2.4g telemetry
3. A2 built-in receiver bypassed Duel RSat2 PPM
4. GPS
5. Dji 5.8G Video TX, Teradek Bolt 2000 5Ghz
6. Norman GPS
7. IMU LOCK FLIP OF DEATH at 35m altitude. Then all motors powered down.
Smashed 4 booms and 5 Props.
Distorted Movi10 Gimbal
Ripped off
Canon CN7x17 KAS S/E1 - 4K Cine-Servo Lens (ON LOAN FROM CANON FOR REVEW) form my:.........
RED EPIC-M DRAGON (CARBON FIBER) destroying the sensor and demolishing the brain covering (and god knows what else until RED report back)
All in all £27005.346 ($44772.30) worth of video kit.

OK not A2 but Ace Wavepoint £7,500 ($12434)
SkyCam (2m signal rotor heli)

1. V4.02
2. Jeti DC-16 duplex 2.4g telemetry
3. Jeti R18 2x RSat (18 channel PWM)
4. GPS
5. Teradek Bolt 2000 5Ghz
6. Cruse GPS
7. Altitude 9m
8. Custom Gimbal
9. Canon 1Dc camera + 4K recorder +
Canon CN-E15.5-47mm T2.8 L S EF mount 4K / Super 35mm Prime Lens (£15,405)

Well EM.. I have flown this setup many times without any trouble until one day a certain government body decided they were going to some GPS testing.
To cut a long story short and not to disclose too munch info (they may be listening hehehe).
Anyway they swamped GPS for a "TEST"!!!
This would have been ok since I am good at single rotor manual flying. But then this happened the clever little Dji Ace wavepoint thought "Ah something funny is going on with the GPS so I will drop out of GPS and go to Atti Mode"
Very good Dji's Ace wavepoint but DONT TAKE 10 £$**$*£ SECONDS TO DERIDE TO DO THIS especial when heading towards a VERY large tall tree.
And within this 10 seconds I flipped over to manual but got NO response at all.

Well you can imagine what a mess was made when the Heli plowed into the tree.
I sent a full report back to Dji with the ACE and COST of damage.
There response was "The system takes 10 seconds decide weather or not the GPS data is usable or not, meanwhile the system locks the outputs in there current state". (full steam ahead) "We will replace the Ace wavepoint as a gesture of good faith." !!!!

Thanks Dji.

Anybody want to buy 5X A2 and 3X Ace wavepoints?
 
Last edited by a moderator:


JLO

Member
Hi ACP thats a lot of damage omg........
just to clarify could you specify more "Flip of death" in your case first what did the components on the copter do, or stop doing, like if it was a motor #1 than the others etc. so to try to isolate the fiscal chain of events BEFORE you hit the ground
very important your flight mode or settings and if you where trying to do something at the controls at the time of the flip of death
thanks for sharing and really sorry for your loses
 


JLO

Member
Guys if I may sujest we really need to use this thread to try to document the incidents in detail like a CSI's hoping to find a pattern or something specific to the A2 failures, yes we know the sad and frustrating part after a crash wether it was minor or mayor damage, I bet A2 users need answers, not so much condolences, in order to move forward with this, lets concentrate in the facts BEFORE the damage, we as a group in the apparent absence of DJI could hopfully come up with solutions and recommendations, remember this forum has several big DJI dealers that are also HL copter builders, pilots, and users of DJI stuff, and the A2, like in the USA, the folks at Aerial Media Pro, or in Europe the folks at Kopterworx and several others, it is to the best interest of the dealers for so many reasons that the issues of the DJI flag ship multirotor flight control A2 system are resolved, it would be great if they can comment in a objective way and help to find the REASONS of why the A2 is having this issues.
 
Last edited by a moderator:


ACP

Member
A2 Vibration Test results.

HI all, on the “A2 FC Firmware 2.3 - Here We ALL go Again” thread I suggested a test.
I thought this is a better place to post the results.
I wanted to see if I could reproduce the IMU lockup I had with my Cinistar 8.
What I mean by IMU lock is, I was lucky enough to be watching my OSD at the time my A2 went nuts. I noticed that... I put in BOLD as this is important...
The pitch and Roll data at the top centre of the OSD went blank, seconds before the A2 went NUTS (OK 0.8 seconds apparently but felt like seconds) and shut my motors off. “


So I took a brand new A2 out it’s box, mounted it to a controllable vibration plate using the sticky pad provided.
I connected up 8x ESCs and motors (no Props) and so on.
Essentially a typical setup for a Multirotor.
For simplicity sakes I imported the settings from my Cinistar, powered up and performed a calibration.
All the wires got in a mess so re-setup the rig on a ply board (like I should have done in the first place) tacked everything down and tried again.
Ok so to make this as realistic as possible I “flew” around the garden in atti mode. Then up onto the workshop roof. Into lied it down then into GPS. Tilting and moving the board the motors reacted as you would expect.
Back in the workshop I watched the ODS and all was as expected.
I started the Vibe table up at 100hz and nothing to observe. (it’s a shame you can’t get live data).
So I setup the frequency to slowly ramp up to 6,000 hz over a 120 seconds.
Again nothing to observe, I can still here the motors running above me.
So I setup a random 4,000 to 6,000 hz varying frequency for 60 seconds.
This time I did here the motors change pitch erratically. I still have Pitch and roll data on the OSD, I stopped the Vibe board, went up and tilted the board around and still had appropriate response from the motors.
Maybe the frequencies are way off. So I opened the video file from the FPV DVR and listened to the sounds seconds before the crash. Ah an idea. So after a coffee I fashioned up a way to play the auto back from my PC to the Vibration board. (I guess a speaker would have worked and been a lot easier).
So I played the video back with the A2 running in GPS and watched the waveform on the oscilloscope. And….
There it is. The pitch and role data vanished from the OSD…….. Just like it did when the Cinistar crashed.
I ran up to the roof but the motors are still running un-like the crash BUT NO RESPONSE from the A2.
VIBRATIONS!!!!!!!!!! The A2 does not like vibrations at certain frequencies!!!!!


I’m now studying the data log. The think that instantly jumps out is large spikes in the can voltage.
The other data looks remarkable like the data from the crash. Now I wonder, in electronics when frequencies increase there is a higher current demand. Could this be the cause of the spikes in the Can voltage??
Dose this get so bad the A2 or IMU falls over???
It’s 04:15 now so maybe not the best time for my little brain to analyse my findings.
But I hope this test sheds some light on the issue.
Incidentally I can’t get the A2 to work again now. I think I have killed the IMU. But it was worth it if this helps solve the problem.


attachment.php
 

Attachments

  • Can Volt.JPG
    Can Volt.JPG
    84.1 KB · Views: 555

ACP

Member
Now when I think about it, the current demand for higher frequencies is unlikely to be the issue. Since Micro processors oscillators are are not dynamic, they always operate at the same speed.
 

JfvCarbon

New Member
Hello everybody, first post and unfortunately to report a first crash!

We are developping a new drone based on carbon fiber moulded pieces to get a very strong cell to protect stuff from shocks and rain and other problems.
It' is a quad X with the new 8108 motors and big 28 inch propellers, i have installed an A2 with the IOSD mark II to perform some flight tests.
Everything went perfectly fine since the first take off, and i we have performed many test in various conditions during about more than 10 houres without having any problems.

But suddenly as the drone was hovering at 120 metres high, i was checking my fpv screen to setup a descending speed of 0.5 m/s when i heard a roar over my head.
The drone suddenly started to spin quickly and 2 seconds later it started to oscillate and tumble down to the ground.
As it was at 120 meters high it took more than 15 seconds to hit the ground. In the data logs i saw that it hit the ground at 14 m/s and yawing at 1 turn per second.
I had plenty of time to go from GPS to Attitude then manual, witout any changes, i also had the time to try the failsafe and the IOC, but no chance.
Fortunately as we build it very strong there was almost no damage, the drone stayed pinned on one arm, the propeller sticked in the ground up to the motor.
I had to give a lot of energy to take it out, but miracle !! the propeller was not broken, just some little scrtaches!
So it ended up with just a little video cable that had been ripped of the connector when the battery moved form its place.

We have double checked everything, all motors work perfectly also the esc, we have made some tests in the bench to see if there was a ploblem, but nothing.
We also have checked the A2, but everything seemed ok, all setting were still correct.
But it's when we saw the data that we saw what did happen.
From the data logs the drone was hovering very smoothly, no acceleration, not gyro movement, no input from the transmitter, but suddenly the motor M4 went to full spedd and stayed blocked like this for 10 seconds, two seconds after the M4 full speed the opposite motor went to full idle, so you can imagine the result.
It was like the A2 was locked for a while giving wrong order to the motors.
It the logs we could see that without any reason the drone started to yaw at 1 turn per second for more than 15 seconds until it hit the ground, the cause was the motor 4 that went to full speed.

After a complete dissassembling and reassembling we made a new test, but it was totally impossible to even take off. The A2 is giving crazy orders to the motors, you power up then give some gas and it seems to take off but with no reason suddenly one or two motor go to idle or accelerate.

We are actually building a well suspended plate to filter vibrations as we suspect that this is the problem. But when you check the logs from the impossible take off you can see that there are no erroneous data from the acceleration sensor or gyro, but the A2 motor out are giving erratic orders.
If you set the fly mode to manual the motors seem to respond correctly, in attitude or GPS it's totally impossible to even take off, so it's not an esc or motor problem, but certainly an A2 failure.
 
Last edited by a moderator:

ACP

Member
This looks like a very clear,clean log showing the motors being given strange commands. I initial wondered if the A2 browned out or locked, the watchdog timer then should reset the MCU, but if the MCU locked I would suspect the PWM
to drop out thus stopping the motors. The PWM is software generated not hardware as the internal hardware PWM is far too high a frequency. Unless the ESC holds the last known good PWM signal?
 

ACP

Member
I posted this on another thread and thought it would be sensible to add it to this.
We were discussing anti-vibration.

It is common scene that an IMU would suffer from excessive vibrations. these are sensitive MEMs. You can't expect them to perform perfectly when they are being vibrated.
I have tested this with the A2 as posted in http://www.multirotorforums.com/showthread.php?21630-A2-Problems-Catalogue
the A2 can't cope with certain frequencies and simply gives up.
I wont be doing this test with the WKM as they are the only unity I have left I can semi trust.
I't is unclear whether the A2 IMU has built in anti-vibration but I presume not.
There is also the amplitude of the vibrations to take into account. Dji CAN'T be expected to manufacture a bullet proof system at these prices.
We have to look after the hardware so they look after us.

I add though if we do look after the hardware and give it a chance and it still fails then Dji simply has to withdraw there product and fix the problems.
We users can't afford to keep testing them for Dji.

If I can get the A2 working again I will re-test as I did before with anti-vibration measurements.

It would be nice for somebody from Dji to comment on this this but I suspect they won't.
 

soler

Member
Interesting analysis, I hope you get it working to see if you can reproduce the test.


Sent from my iPad using Tapatalk
 

JfvCarbon

New Member
After last week crash we have desassembled and rebuild everything, we have mounted all the A2 system on a well suuspended plate, we have mounted the motors on silentblock also.

I also have downgraded the A2 to 2.1 and re-upgraded to 2.3 to be sure that every part has been flashed to the last firmware.

We have tested it today, it's flying correctly, but we had some little problem especially when you are yawing, it can happen that it starts to wobble a little, but not every time.
As we are not really confident in that setup, tonight i have installed a wookong in it and we will make some tests tomorrow.

But today we have learnt something very usefull, our copter is really goog at trimming trees !!
My friend was a bit to optimistic and tried to cut some branches on my pear tree, the copter was sticked in the tree cutting foliage and branches with those large 28 inch propellers without seeming to be much disturbed, thanks god no damage, after having cut his way out it couuld go on as nothing had happened:
My god those carbon propellers are indestructible, they could certainly cut you head off without noticing it!
They are Foxtech Supreme C/F Propeller(28x8.0), they give 3.8 kg thrust at 2500 rpm with the 8108 foxtech Motors KV135 for 15 amp on 6 cells. So for a weight of 7.5 kg you can fly 25 minutes with a 12 000 Ma/h.
 

snurre

Member
You (and DJI) might find this info utterly interesting.
Keyword: Sensor aliasing

Following is an extract from the user forum at Ardupilot.com and the post is written by Andrew "Tridge" Tridgell, the lead developer of Ardupilot/Plane.

.....

Dual Sensors

The second really big change for this release is support for
dual-sensors. We now take full advantage of the dual accelerometers
and dual gyros in the Pixhawk, and can use dual-GPS for GPS
failover. We already had dual compass support, so the only main
sensors we don't support two of now are the barometer and the airspeed
sensor. I fully expect we will support dual baro and dual airspeed in
a future release.

You might wonder why dual sensors is useful, so let me give you an
example. I fly a lot of nitro and petrol planes, and one of my planes
(a BigStik 60) had a strange problem where it would be flying
perfectly in AUTO mode, then when the throttle reached a very specific
level the pitch solution would go crazy (sometimes off by 90
degrees). I managed to recover in MANUAL each time, but it certainly
was exciting!

A careful analysis of the logs showed that the culprit was
accelerometer aliasing. At a very specific throttle level the Z
accelerometer got a DC offset of 11 m/s/s. So when the plane was
flying along nice and level the Z accelerometer would change from -10
m/s/s to +1 m/s/s. That resulted in massive errors in the attitude
solution.

This sort of error happens because of the way the accelerometer is
sampled. In the APM code the MPU6000 (used on both the APM2 and
Pixhawk) samples the acceleration at 1kHz. So if you have a strong
vibrational mode that is right on 1kHz then you are sampling the "top
of the sine wave", and get a DC offset.

The normal way to fix this issue is to improve the physical
anti-vibration mounting in the aircraft, but I don't like to fix
problems like this by making changes to my aircraft, as if I fix my
aircraft it does nothing for the thousands of other people running the
same code. As the lead APM developer I instead like to fix things in
software, so that everyone benefits.

The solution was to take advantage of the fact that the Pixhawk has
two accelerometers, one is a MPU6000, and the 2nd is a LSM303D. The
LSM303D is sampled at 800Hz, whereas the MPU6000 is sampled at
1kHz. It would be extremely unusual to have a vibration mode with
aliasing at both frequencies at once, which means that all we needed
to do was work out which accelerometer is accurate at any point in
time. For the DCM code that involved matching each accelerometer at
each time step to the combination of the GPS velocity vector and
current attitude, and for the EKF it was a matter of producing a
weighting for the two accelerometers based on the covariance matrix.

The result is that the plane flew perfectly with the new dual
accelerometer code, automatically switching between accelerometers as
aliasing occurred.

Since adding that code I have been on the lookout for signs of
aliasing in other logs that people send me, and it looks like it is
more common than we expected. It is rarely so dramatic as seen on my
BigStik, but often results in some pitch error in turns. I am hopeful
that with a Pixhawk and the 3.0 release of APM:plane that these types
of problems will now be greatly reduced.

For the dual gyro support we went with a much simpler solution and
just average the two gyros when both are healthy. That reduces noise,
and works well, but doesn't produce the dramatic improvements that the
dual accelerometer code resulted in.

Link to the full version of Tridge´s forum post:
http://ardupilot.com/forum/viewtopic.php?f=98&t=7129

Tell me what you think, guys.
/ Tom
 

gtranquilla

RadioActive
Snurre,

That seems like a good info find to me. Yes doubling up on the sensors might help but seems overly complex.
The anti-aliasing concept is still a bit vague too. Hope someone can explain it better.


You (and DJI) might find this info utterly interesting.
Keyword: Sensor aliasing

Following is an extract from the user forum at Ardupilot.com and the post is written by Andrew "Tridge" Tridgell, the lead developer of Ardupilot/Plane.



Link to the full version of Tridge´s forum post:
http://ardupilot.com/forum/viewtopic.php?f=98&t=7129

Tell me what you think, guys.
/ Tom
 

min0nim

Member
That is really interesting.

I just ordered an IMU vibration isolation mount - one of these: http://www.readymade3d.com/anti-vibration-mounts - because of something funky happening with the suability in certain conditions. I was thinking some kind of standing wave in the frame.

Anyone using something like this?

Just trying to eliminate everything one step at a time...

edit: gtranquilla: The aliasing issue - imagine the IMU samples the level of the bird by looking at a wave pattern. The IMU is reading slices of the wave, and for most of the time these slices match the wave just fine. But you can produce a wave that matches the sample time of the IMU, so that when it reads a slice, it only reads the tops of the wave. So it thinks everything looks like a straight line, not a wave.

Two IMU's would potentially solve this by always reading different pieces of the wave.

This is a really simplistic analogy, but hope it helps.
 
Last edited by a moderator:

Tahoe Ed

Active Member
That is really interesting.

I just ordered an IMU vibration isolation mount - one of these: http://www.readymade3d.com/anti-vibration-mounts - because of something funky happening with the suability in certain conditions. I was thinking some kind of standing wave in the frame.

Anyone using something like this?

Just trying to eliminate everything one step at a time...

edit: gtranquilla: The aliasing issue - imagine the IMU samples the level of the bird by looking at a wave pattern. The IMU is reading slices of the wave, and for most of the time these slices match the wave just fine. But you can produce a wave that matches the sample time of the IMU, so that when it reads a slice, it only reads the tops of the wave. So it thinks everything looks like a straight line, not a wave.

Two IMU's would potentially solve this by always reading different pieces of the wave.

This is a really simplistic analogy, but hope it helps.

Use this at your own peril. It is not recommended by DJI. DJI recommends a hard mount using their thin 3M double stick tape. The isolation confuses the IMU and it does strange things. I have always used the recommended mount and have never had problems from Naza to A2. It is your choice.
 

MombasaFlash

Heli's & Tele's bloke
... I just ordered an IMU vibration isolation mount - one of these: ...

Whilst not that heavy the IMU does have a certain mass, so you would want to guard against it being able to flop about. If the little dampers on those isolating platforms are too soft that is what could happen. I do not use DJI recommended double backed tape. It is far too thin and does absolutely nothing to shield the IMU sensors. I use Kyosho green gel (two 10-15mm strips, not a whole slab) with a small Hook&Loop security strap that also inhibits any large movement of the IMU. It's worked so far.
 


Top