I've recently purchased a Spark and noticed that the camera yaw seems to drift when the drone is yawed and then stopped. I realise that there is no mechanical yaw on the Spark - but there is a digital one.
I've uploaded a video to show what is happening:
You'll see in the vid, that I yaw the drone and then stop, however, the camera continues to yaw, even whilst the drone remains pointing in one direction. I've double checked this by removing the props from the drone, lifting it by hand, yawing it, and then putting it back down (so the drone is definitely stationary), however, the video continues to yaw drift.
I've even taken my original spark back to the shop I bought it from for a replacement, but the new unit does exactly the same!
I've spoken with DJI support a couple of times and followed their instructions to run, re-run, re-re-run, re-re-re-run, re-re-re-re-un, re-re-re-re-..........all of the calibrations (compass, IMU, controller, etc.) but this hasn't helped. Now the only response is to send it back - which is a pain, as I'll be without the drone for some time by the sounds of it!
I'm wondering if anyone else has had this issue. It makes it VERY tricky to get a decent forward climbing shot, or similar, when the camera continues to yaw on it's own! Certainly seems odd that it would do this on two completely seperate brand new units - maybe a software issue.....
I have that same "issue" but only noticed it when sparky is sitting still, ie. on my table or the back of the truck. Haven't noticed it mid flight, yet
i have the same thing since the beginning.
but its only real bad in low light situations. ill guess the image stabilistation isnt set right and cant process darker images.
in bright sunshine its isnt that visible, may be a slight pan after stopping the yaw input.
definitly isnt a bad calibrated controller. just set your spark up at night on your table and move it a bit around. no need for flying to see the problem.
Ah, I had a feeling that it was a bug - seemed pretty weird to be happening on two separate, brand new units.
Fault on 1 unit = possible ID10T error
Fault on 2 units = fault on unit error
Hopefully the next update will come out soon and solve the problem. I was considering an upgrade to Mavic whilst I have chance to return my Spark (again), but I do really like the spark!
Since sparky is using the exact same sensor as the mavic (4K) but only has a 2axis gimbal - the sensor is "cropped" to compensate the 3axis - hence the excellent 1080p.
I suspect and issue with the compass/IMU that triggers a slight "shift" on the sensor because it's trying to compensate, resulting in the drift even though it is not moving at all.
Just a thought.
Just to verify, had you tried another place to fly? Please try to fly in another place and make sure that you have calibrated the compass and IMU before this flight. If the drone still drifts the same, could you please export the data on the drone and upload it to Dropbox so we can check more about this? Appreciate it.
I have exactly the same thing going on, but i was thinking that it should be normal due to the 2-Axxis gimbal an the yaw stabilisation. I would like to have a FW update that corrects this.
Sometimes it does this and sometimes not. Saw this on different locations and on different light situations.
DJI Thor Posted at 2017-8-9 22:18
Just to verify, had you tried another place to fly? Please try to fly in another place and make sure that you have calibrated the compass and IMU before this flight. If the drone still drifts the same, could you please export the data on the drone and upload it to Dropbox so we can check more about this? Appreciate it.
https://www.youtube.com/watch?v=ngEHdAlAh9A
why ?
the AC doesnt drift, only the picture pans/drifts without input. is the Image stabilistaion connected to compass/ IMU ? If yes why does this affect only the picture not the AC ?
ill bet your engenieers can emulate this problem with ease. just make a few tests in low light, no need for flying, props or anything except an spark/RC/phone...
the AC doesnt drift, only the picture pans/drifts without input. is the Image stabilistaion connected to compass/ IMU ? If yes why does this affect only the picture not the AC ?
The gimbal of the Spark will be affected by its gyro and the AC's action.As you can see, it owns 2-axis gimbal for spark, and gimbal yaw follows the yaw of the Spark. So it is necessary for calibrating it. And I'll forward your advice to our engineer to test this case.
I've tested at several locations, with re-calibrations - still has the same issue. Looking back over my earlier footage, it is also still present (although I didn't notice it at the time, as I wasn't looking for it).
I haven't had chance to do the firmware update yet, or get back out to fly it. Has this been fixed in the new FW update?
Spookymike Posted at 2017-8-10 02:31
I have exactly the same thing going on, but i was thinking that it should be normal due to the 2-Axxis gimbal an the yaw stabilisation. I would like to have a FW update that corrects this.
Sometimes it does this and sometimes not. Saw this on different locations and on different light situations.
Please follow Thor's advice. Make sure that no interference around your spark. You can export the data for us to analyze. Thanks.
It looks to me like it is the electronic image stabilization trying to re-center itself after some movement. The same exact issue can be seen in this thread with the Spark sitting on a desk:
In my opinion, DJI needs to update their algorithm for how it does this because a drift over multiple seconds after the aircraft stops doesn't seem reasonable to me.
I also see this same behavior both in flight and when setting the Spark on a solid surface. My Spark is updated with all the latest and greatest firmware and software updates available at this time.
I noticed it after upgrading from .500 to .600. Maybe I didn't recognize the yaw problem in the .500 version because I was too busy playing around with my new Spark. What I can say is that my spark does not only have the "gimbal-yaw" issues, but also has some problems flying a straight line (even when there's no wind). It's always slightly drifting to one side. Furthermore I see a quite high Gyro bias as long as the Spark is cold (sometimes the indicator bar in the DJI go app even turns yellow). Calibration / factory reset / firmware refresh did not solve the issue. Maybe there's a bad Gyro IC lot out there? My friend's Spark is behaving as expected...
So, I've completed all updates, and finally taken my drone back out for a fly at the weekend. It's still doing the same "yaw drift". This makes it incredibly difficult to get a decent shot. How do we fix this?!?!? At the moment, I'm thinking about returning my spark for good.
Mine has done that on a couple of flights.. Certainly not all of them. It may well have something to do with the lighting conditions. When I did notice it it was starting to get dark.
I have had the same issue on my Spark which flew away from me yesterday due to a yaw error which inexplicably occurred in flight, which put it into ATTI mode, causing me to be unable to use RTH to bring it back.
Disregard my last comment. I'm new on here as well, and published it prematurely and it wouldn't let me change it. Anyway, I had the same issue with my Spark which flew away from me yesterday due to a yaw error which inexplicably occurred in flight, which put it into ATTI mode, causing me to be unable to use RTH to bring it back. Anyway, before that happened, I often noticed that the yaw would move even when the drone was still, and the camera often tilted at a funny angle.
Is there any new update on this issue has it been fixed in newer firmware? I ask because i have kept my spark on 0.600 firmware and djigo app at 4.1.18 as a lot of people had reported problems with more recent firmwares. My spark has this drift on yaw as well and if you try to go forward in a straight line the drifts to one side despite recalibrations of imu compass and transmitter. However as i do not want to introduce a more serious problem with a faulty firmware problem I thought i would ask
.
Dug up this thread as my Spark, less than 12 months old and with the latest firmware, has the same issue with the camera panning for a few seconds after the Spark stops yawing. I've done a quick bench test to show the problem (but it won't let me post the youtube link here?!?!). Using the simulator on DJI Assistant 2 I can even 'turn the camera' using the left stick with the drone sat on the bench, yet the image on my phone pans left and right slightly, maybe 5 degrees according to the data on Assistant. Does anyone, especially DJI, have any update on this? It's far from ideal when I have to try and 'correct' what the drone does.
I've come to the same conclusion having spent an evening searching the web for similar issues. I had kind of fell out of love with the Spark as I thought it was an issue, but having taken it out today and burnt through three batteries in winds gusting upwards of 20mph, I've reminded myself just how capable this little thing is! It's a case of adapting to the new feature - and panning/yawing certainly produces smoother footage than with older firmware.
agree with everyone here that thinks this is connected with 'software yaw stabilization' .. I think that they have changed the algo to be more aggressive with a simulated follow yaw and allow the 'image capture window' to float further in the 4k sensor frame and it does seem to take a while for the image to really recenter to the middle of the sensor... I find that the problem is exacerbated by doing turns all in the same direction .. like flying in a loop around a field vs a mix of left and right turns.. also since the algo for image stabilization is very sensitive .. any 'perceived' yaw by the aircraft is going to be 'corrected' by the software camera yaw damper algo so if / when we start seeing the yaw vs magyaw separation in values I believe the FC will command some of this 'correction' as it thinks the aircraft is yawing..
Strobing_NYC Posted at 11-16 15:34
It's not a "Bug" it's a feature so you get smooth transitions when ending a Yaw movement (i'll bet that's DJI answer lol)
MY SPARK IS 1 YEAR OLD BOUGHT NEW
Your right that’s exactly what it is, there are many who just can’t work these things and will put it down to FW/SW , no matter what’s wrong with their drone it’s always somebody else’s fault.