Welcome DJI Spark Pilot!
Jump in and join our free Spark community today!
Sign up

Flyaways

After 39 years building H/W & S/W systems for the DOD and 15 years of those working with UUVs here are my observations:

These drones have a number of sensors that are used collectively to navigate. The IMU , GPS, or Compass can be used independently should other sensors fail. If a Flyaway occurs it is most likely the fault of the S/W not handling an abnormal condition. I think DJI makes some a really nice and impressive product for the price but their S/W seems to be poorly written. Example: read all the problem with the DJI GO App.

I hope the S/W engineers that work for DJI don't get a job with GM or Ford

Just my 2c worth
 
  • Like
Reactions: Hotwire
The imu gps and compass do operate independently. The imu continues to stabilize the craft without gps or compass, but the the gps of course cannot be used without compass.
However I have to say I was extremely disappointed when the other day, for no apparent reason I beheld my crystal sky displaying 'go4 app not responding , close?'
I had thought I'd left all that behind with CS , but alas no.
 
..... but the the gps of course cannot be used without compass.

Is that so?
With the right software, and with a minimum distance movement, the drone can determine in with direction it is moving, based on the change of the GPS coordinates. Or am I wrong?
 
  • Like
Reactions: REMUS1
Is that so?
With the right software, and with a minimum distance movement, the drone can determine in with direction it is moving, based on the change of the GPS coordinates. Or am I wrong?

This reminds me when, a couple of years ago, I used to play "geocaching" and my movements were based on the numbers of the LAT and LONG moving up and down while I was moving around until I finally figure out in which direction goes up and wich one goes down...

Sorry for that memmory but just came to my head...
 
Well it can be done I believe, but you would need gps obviously, and decide to ditch the compass instead, which would entail a fair bit of programming I imagine to deduce the way home. I think it's relatively easy on an aircraft which is almost guaranteed to be moving forwards.
Still I'm not an expert in any way, and just wish dji could put the money into my app not crashing.!
By the way I would love to know where all these compass errors in flight come from, except the obvious ones of flying too close to buildings, ships power lines etc.
 
PS I think all this would be out of kilter with the spark's brief. It's already pretty complicated I imagine, and each additional complication brings new challenges.
Where I am unhappy is the difference between the performance I expected and what I got, perhaps I saw what I wanted to see.
The go 4 app crashing on my crystal sky mid flight has pushed me very close to calling it a day.
 
Is that so?
With the right software, and with a minimum distance movement, the drone can determine in with direction it is moving, based on the change of the GPS coordinates. Or am I wrong?

Technically, yes. But without a compass, the drone is blind to which direction it is pointed relative to North. So, if the GPS coordinates indicate a movement to the west, and the drone is physically pointed east, it wont know to fly "forwards" to correct the movement. What ends up happening is the drone starts to hunt....it makes a correction in one direction.....then another....then another....and the result is a big spiral, also known at the "toilet bowl effect". DJI tries to eliminate this concern by ignoring GPS imput, and entering atti mode.If the drone is low enough, VPS should keep it stable (as it does indoors).

Just as you said with geocaching. Your handheld GPS probably doesnt have a compass to tell you north. So you walk a bit in one direction....nope. Then a bit more in another....nope. Repeat until you stumble on the cache. You have the advantage of being aware of your surroundings. The drone does not.
 
If you integrate the Acceleration Vector, you get the Velocity Vector, If you integrate the Velocity Vector, you get distance. Draw a line between two points and you get heading So Navigation via the IMU is possible. Assuming you have a initial GPS fix. I work with underwater "drones" and they have to navigate with the IMU since GPS is not available.
 
  • Like
Reactions: Martin Lines
Using 'hand gestures only', don't! Because it could do something weird and you are helpless to stop it.

Yea... mine did. I was practicing HG mode in the front of the house on our semi-rural dead end street. It has never worked well. This time I was doing pretty good. I got it to go left and right and up and down a little. Then I waved at it so it would back up and follow. After several waves, it backed up. I took a picture....seemed to work. Next I started backing up down the street while watching the drone to see if it would follow.... it did.... for a while. Then it stopped following and just hung there about 15 feet up. I tried everything in the book to get it to respond but it just kept hanging there. Eventually it ran out of juice and descended and I let it land in my hand. My though was.... Gee, if I was doing this from the deck of an anchored cruise ship or from the ledge of a cliff I would have been out of luck.
 
  • Like
Reactions: Victoria
Yea... mine did. I was practicing HG mode in the front of the house on our semi-rural dead end street. It has never worked well. This time I was doing pretty good. I got it to go left and right and up and down a little. Then I waved at it so it would back up and follow. After several waves, it backed up. I took a picture....seemed to work. Next I started backing up down the street while watching the drone to see if it would follow.... it did.... for a while. Then it stopped following and just hung there about 15 feet up. I tried everything in the book to get it to respond but it just kept hanging there. Eventually it ran out of juice and descended and I let it land in my hand. My though was.... Gee, if I was doing this from the deck of an anchored cruise ship or from the ledge of a cliff I would have been out of luck.
That is scary using this gesture mode. I think there is a specific distance it will follow you till it stops. But it shouldnt stop responding altogether. Argh!
 
  • Like
Reactions: RaspPiGuy
Yea... mine did. I was practicing HG mode in the front of the house on our semi-rural dead end street. It has never worked well. This time I was doing pretty good. I got it to go left and right and up and down a little. Then I waved at it so it would back up and follow. After several waves, it backed up. I took a picture....seemed to work. Next I started backing up down the street while watching the drone to see if it would follow.... it did.... for a while. Then it stopped following and just hung there about 15 feet up. I tried everything in the book to get it to respond but it just kept hanging there. Eventually it ran out of juice and descended and I let it land in my hand. My though was.... Gee, if I was doing this from the deck of an anchored cruise ship or from the ledge of a cliff I would have been out of luck.

you can still connect to via controller or phone, whichever it is setup for. I'd keep it handy in those cases.

I always use the controller, but always also to gesture control launch using the double tap on the battery, I then go into the app and change it from gesture to "normal"
 
From what ive read, the rc also uses wifi connection to the aircraft as well, so i assume the same types of disconnect may happen as well. I suppose the rc just has a stronger wifi connection which minimizes the disconnect chances?

Keep in mind the controller has big honkin' antennas.
 
Yea... mine did. I was practicing HG mode in the front of the house on our semi-rural dead end street. It has never worked well. This time I was doing pretty good. I got it to go left and right and up and down a little. Then I waved at it so it would back up and follow. After several waves, it backed up. I took a picture....seemed to work. Next I started backing up down the street while watching the drone to see if it would follow.... it did.... for a while. Then it stopped following and just hung there about 15 feet up. I tried everything in the book to get it to respond but it just kept hanging there. Eventually it ran out of juice and descended and I let it land in my hand. My though was.... Gee, if I was doing this from the deck of an anchored cruise ship or from the ledge of a cliff I would have been out of luck.
Hand gesture mode is not really any more than a gimmick. The money would have been better spent on descent video transmission.
I think Spark is a great idea poorly thought out.
 
Here's the thing. Spark gesture only follows you for 10 metres. Then it stops. No gestures then work. You can turn on rc recover or it will auto land on flat battery. But I don't think it's poorly thought out. It's more that it has taken me some months to explore it
After following fora for this time , I opine that virtually no flyaways are spark faults. People fly it at night in 50 mph winds behind things.
You can tell from the problem flight posts that most have not even downloaded the manual.
Even the true flyaways have probably been crashed or misused first.
Some knowledge of flight re ground speed vs airspeed is essential in my book.
But hey ho, can't help every one.
 
So my biggest concern with my two drones that I just got (spark and platinum mavic pro) is flyaways.

Honestly I'm still trying to figure out what really is the cause for flyaways other than the fact that it's due to Magnetic interference, loss of GPS signal, Wi-Fi interferences Etc.

I know many people have stated that it's usually pilot error and not having a thorough PreFlight checklist. I'm curious though that even with a thorough PreFlight checklist and ensuring that there is no magnetic interference and having a strong GPS signal and no wifi interference, before taking off. Does it fly away after the fact because one of the symptoms may have come out of nowhere while flying or is it because the pilot never checked all of the above before taking off?

I'm still unclear and unsure to have a comfortable feeling flying this thing anywhere.
i think it is mostly bad software programming by DJI engineers. Also to reduce costs, they do not test the software for quality and let the bugs come from buyer feedback.
 
Well after familiarising myself with it, I had some very surrprising results which after thought all turned out to be correct behaviour, many of which people are still contesting as faults.
When the public know that in the main spark flyers' and others' main interest is to blatantly ignore the drone regs by choosing to fly as far and as high while out of sight as they can, we'll be banned outright.
We can't claim responsible use since it isn't.
It will be known as a cowboy fraternity, rightly in my view.
 
Well after familiarising myself with it, I had some very surrprising results which after thought all turned out to be correct behaviour, many of which people are still contesting as faults.
When the public know that in the main spark flyers' and others' main interest is to blatantly ignore the drone regs by choosing to fly as far and as high while out of sight as they can, we'll be banned outright.
We can't claim responsible use since it isn't.
It will be known as a cowboy fraternity, rightly in my view.

Sorry I'm not understanding what you are saying. Are you saying that there is no such thing as flyaways and that it's all a users fault?

As long as you fly within your line of sight and within the limits it should never happen?
 

Members online

No members online now.

Forum statistics

Threads
14,601
Messages
118,824
Members
18,013
Latest member
JulieMyers