DJI S800 yaws in flight

[Javajack]

Will let you know. I have yet another V2 IMU on the way. Going to try that. It seems like we've invested more than we've ever made back with this thing, but we keep advancing our set-up, recently to 4K GH4.. the best move I've made. No more line dropping in images, so totally pro image now. Anyway, thanks again for your help.

 

[enak]

What weight is your S800? I ask because I started having intermittent loss of yaw control when I tried a heavier payload and although my EVO took off and flew fine I noticed every so often it would just step out for a bit and then fly fine. Once I noticed what I believe was one of the ESC lights changing to yellow briefly but since there is no log on the ESC it could have just been the angle. I am flying at <7Kg now and have not yet had a repeat. My theory is that the ESC cuts out briefly (for whatever reason) but is working again by the time that I've realised what is going on.

The symptoms you describe are exactly that of a motor/ESC/prop failure. Since the flight controller will happily deal with one motor out you would only notice the craft turning and home-lock/ course-lock would compensate the yaw drift automatically.

This could also be a bad lead from flight controller to frame or dry joint on the fight controller or frame connectors resulting in loss of signal to one motor. Difficult to find without fitting a servo data logger near the esc.

 

[R_Lefebvre]

I've had this problem from day one. On the first version of A2 software I was a able to reach speeds of up to 27m/s (tail wind + using a decent to pick up speed)

Sorry for the hijack but, 100 km/h from an S800? That's incredible.

 

[flappybird]

Yep, and totally possible on the first version software. All you had to do is start into a decent to gain speed with gravity and push the stick as far as it could go. pretty simple really. Its a shame A2 version 2.3 has a 20m/s limit because the S1000 could easily handle going faster. They also need to kick the maximum decent rate back up to -6m/s because 4m/s is too slow.

 

[Benjamin Kenobi]

Best I got out of mine was 40mph (60kph) on a windless day down a runway.

 

[flappybird]

40mph or 18m/s is consistent with the maximum speed you can achieve with 2.3. My S1000 will accelerate right up to 20m/s which is about 45 and will maintain that speed as long as I want with no issues and plenty of power reserve to ascend. the fact that 2.1 allowed for a faster decent and flight speed is kind of a mute point as no one is probably flying that old software anymore.

 

[Javajack]

Okay gang. Everything was perfect yesterday. I recalibrated the compass and flew a big job without an issue. The S800 operated like it should, like old times. Today, I took off early in the morning and the unit started to spin (toilet bowl) and I couldn't gain altitude. Obviously a motor is under stress or not performing as it should. I managed to land safely but this unit is grounded until I can figure out what's going on. My technician is away on holidays and he usually takes care of all hardware and software issues. I would pay someone to help me with reading the flight data and figuring out what is going on. If someone is interested in assisting me, please contact me at jack@javapost.ca. I am on a Mac computer so it is difficult for me to do anything with the data. I use the iphone app as my assistant for all other set-ups, but reading the date is another kettle of fish. Thx.

 

[Benjamin Kenobi]

If you post some of the files on the forum we can all have a look. Dropbox is good for sharing.

 

[Javajack]

Ok, I'm driving in from outside the city. I'll try to post something later. I need this unit operating this week. I am shooting for some commercials and a major movie in 4K. Thanks for the quick reply.

 

[Javajack]

How can I get the data off the iosd on a Mac? I don't really have access to a pc. I just plugged in the usb cable hoping it might pop up as a drive that I could get the data off of.

 

[Benjamin Kenobi]

It needs to be powered up, then use the iOSD assistant to put it into data view mode, then it will pop up as a drive.

I know, they don't make it easy eh? And it takes ages to download.

 

[Javajack]

Okay, I've put the data from the last flight, flight 75 from the iosd onto our ftp site at ftp.javapost.ca. I've placed it in the public folder/S800. There is no password for this folder as it is public. Let me know if you can figure anything out from this data. Thanks, gang.

 

[Benjamin Kenobi]

For some reason when I look at that data I just get flat lines for everything. Sure it's the correct file?

Any chance you know what version of the iOSD software you're using?

 

[Javajack]

Benjamin: I think I figured this all out myself. I've been looking over the flight data and here are the issues I see and some of the reasons I assume things went horribly wrong. First of all we had a gimbal malfunction. The back tube on the zenmuse was out of alignment and the gimbal was not swinging freely. I was just meeting the client for the first time and they were very curious and wanted to know what was wrong, of course. We had large cell towers towering over us a few hundred feet away. I believe this reduced my number of satellites on take-off. It looks like there are only 4 satellites according to the flight data. I always wait for at least 7, but as I say, I was probably distracted by other things and may have taken off with 4, which increased to 8 but then looks like it sank back to 4 again, according to the flight data. The copter was powered up a long time, because I had to mess with the gimbal. By rights we should have had lots of satellites, so my guess is that the cell towers were causing interference for us. That's my best guess at this point. The good news is I was able to land it safely but I don't want that happening again. Will recalibrate my compass again and do a flight test and let you know how it handles. I still think there is a weak motor. A half hour before this flight I did a flight where the S800 yawed around on me in a light wind. This shouldn't be happening. I should pull the data for that flight as well and upload it. It would be flight 74. See if anything jumps out.