** Do not move the solo by hand after powering on **

[Pedals2Paddles]

CRITICAL UPDATE

The brains at 3DR along with several volunteers found a bug in the firmware that was causing these incidents. That bug has been removed from a firmware update released yesterday evening. Firmware version 2.4.1-6 is now out and every solo should be updated prior to your next flight.

important firmware update

The issue was unrelated to moving the solo around prior to takeoff. You still need to leave it still while it is powering up and initializing, which is stated in the manual as well. But once it is powered up and initialized, you can safely move it around,.

Everything below this line is superseded by the above.
---------------------------------------------------------------------------------------------------------------------------

This topic has been coming up in the FB groups after several recent crashes. They had a very similar pattern of events and some common denominators. What initially looks like a solid hardware failure is actually a series of events that weren't explicitly wrong, but led to a this pattern of malfunctions. So I'm posting this as caution to everyone, hopefully to avoid further issues.


Short version: If magnetic interference or false calibration data is introduced between powering on the solo and arming for takeoff, you will be taking off with bad compass data. In flight, as you move around and make turns, that compass variance will increase. Eventually, the compass data will be so out of whack, the Solo can't make heads or tails of it's horizontal position anymore. It will end up a fly-away situation, taking off in some random direction and not responding to any user inputs. A+B+Pause kill switch might still work since it is commanding a flight maneuver, but nobody has ever had the opportunity to test that. But basically nothing else will work.

Conclusion: Find your suitable taking location, put the solo down, and power it on at that location. For example, do not turn it on over by the car or the house to give the GPS a head start, then carry it out to the middle of the yard. If you must move it, it apparently would be best to power cycle it. I know it's annoying, but it's better than having to retrieve the solo from your neighbor's tree or living room. This all goes for any Arducopter based drone, not just the solo.

RTFM: The solo manual from 3DR says not to move the solo during startup while the sensors and flight controller are initializing. There is no duration or indicator associated with that instruction. IMO, most reasonable people would take that to me a very short period of time while it's fir powering up and beeping/blinking. But after seeing what several people are experiencing, this should include the entire during from power on to arming.

Disclaimer: This isn't an official instruction from 3DR. But several very smart people with intimate knowledge of the pixhawk flight controller have reviewed the videos and log files and come to this conclusion. Some in DIY community have known this best practice for years, but it's not something "everyone but you" knew. So, call it a serious recommendation backed up by lots of evidence.

Any of you smart dudes, feel free to chime in if I missed anything.

 

[Jubalr]

I just looked at the logs of another crash into a tree posted on FB a few hours ago. Same thing. He powered on and messed around with it then carried it out to the spot to take off. After a few minutes and during an Orbit, he started getting compass errors then EKF before hitting a tree.

 

[Rich]

Or try to power it on or arm it on a boat!

 

[Pedals2Paddles]

I just looked at the logs of another crash into a tree posted on FB a few hours ago. Same thing. He powered on and messed around with it then carried it out to the spot to take off. After a few minutes and during an Orbit, he started getting compass errors then EKF before hitting a tree.

Yep, I'm the Matt in those posts that put up the log screenshots. Did you notice the turning similarities? Both of them had the compass finally go bananas in the middle of a long sweeping turn. Almost completely identical.

 

[Jubalr]

Yep, I'm the Matt in those posts that put up the log screenshots. Did you notice the turning similarities? Both of them had the compass finally go bananas in the middle of a long sweeping turn. Almost completely identical.

Yep, good call. This could be something related to the latest FW. When Solo was released we saw a few of these, then nothing until the last month or so. Always good practice to power on and leave it, but something new could be making the issue worse.

 

[Pedals2Paddles]

Ya that's what bugs me. People have been powering on and moving these things around forever. Myself included. Usually cursing trying to get a GPS lock... And the compass variance issue at hand has always been possible. Randy posted about it on DIYD in 2014 before the solo even existed. But it's not until solo 2.4 rolled out that it reared it's ugly head repeatedly like this. I suppose it could be coincidence. But it bugs me.

 

[Ken Jupiter]

Thanks for the heads up P2P. I had a similar experience the other day when Solo (2.4) gave me a little excitement when it went into fly:manual mode during a MPCC. I was able to recover and land safely. The tlog showed compass and horizontal variance alerts. I always let Solo remain still when powering up, so my problem may be something else. I believe that 2.4 shows a sensor calibration alert as Solo powers up. 3DR tech support told me "these are safety measures implemented on Firmware version 2.4.0, due to the fact that some customers kept moving Solo after it was being turned on, causing both of these sensors to have bad readings, unless the message remains after that, you will be good to go"

 

[Pedals2Paddles]

Thanks for the heads up P2P. I had a similar experience the other day when Solo (2.4) gave me a little excitement when it went into fly:manual mode during a MPCC. I was able to recover and land safely. The tlog showed compass and horizontal variance alerts. I always let Solo remain still when powering up, so my problem may be something else. I believe that 2.4 shows a sensor calibration alert as Solo powers up. 3DR tech support told me "these are safety measures implemented on Firmware version 2.4.0, due to the fact that some customers kept moving Solo after it was being turned on, causing both of these sensors to have bad readings, unless the message remains after that, you will be good to go"

The plot thickens. So 3DR customer support acknowleges moving the solo around after powering on is a problem. But far be it from 3DR to, you know, TELL ANYONE ABOUT IT. This also unfortunately shows that all the pre-arm checks in the world may not catch every single little glitch. Some problems may have no signs/symptoms until it is too late.

It also shows this problem apparently occurred in other firmware versions as well, not just a 2.4 thing. Not noticing a rash of incidents until now could just be coincidence.

 

[User Name]

When "normal" APM/PX4 flight controllers power on, after a few seconds they go through their initial calibration, indicated by the status LED cycling through yellow/red/blue for a second or two. I'd it's moved during this time, it won't complete the calibration and needs to be power cycled. I would assume the Solo is basically the same, although there is no outward indication of the calibration sequence.

Perhaps the movement in these cases is occurring during the calibration routine? I don't see why movement after calibration would effect flight later on as how would that be different from movement during flight?

Not that I'm disagreeing with the advice not to move the Solo after powering on, especially post 2.4 firmware. I've always tried not to do this generally, but I'll make it a practice going forward. I'm still on 2.3 and these stories have me reluctant to update.

 

[Pedals2Paddles]

That was always my experience with APM and Pixhawk as well. Perhaps we're just lucky? Perhaps in the past it was just more forgiving of compass variance? It doesn't make a whole lot of sense to me, but I didn't build or code it so IDK. It could also be that turning it on near a structure that interferes with the compass is the crux of the problem. Moving it to another location just happens to be what takes place next. IDK.

 

[SPP]

Thank u P2P .... thank U this Forum.. and thanks to all the nice people.

 

[SteveReno]

Never seen or heard of anyone trying to locate GPS lock, while moving Solo around. LOL....

Common sense to "not to move" Solo while initializing.

If errors occur while initializing, I turn off Solo then "Depending on error for fix" i would then restart Solo & try again.
initializing is the most important procedure & the last tell tail sign that there is something wrong or everything is right before flight.
I live in the city, I get "Lost connection" often & depend on RTH for Solo's safe return.
This is another important reason to make sure that initializing Solo goes without errors!
If initializing fails & not for sure what to do, CONTACT TECH SUPPORT.

 

[Pedals2Paddles]

@SteveReno, The manual does indeed state not to move it around during startup while it is initializing. Nobody is doing that. They're moving it around long after startup is complete. That is not what the manual says, and common sense to a consumer user would dictate there is nothing wrong with that.

Picking it up and moving to another location 10-20 ft away is very common when it refuses to get a GPS lock. I don't mean to actually carry it around searching for a signal like people do with a phone. That would be funny to watch though.

 

[Ed Beck]

This is probably obvious but it just dawned on me. After you do a compass or level calibration REBOOT Solo before sending it up. Those calibrations certainly involve moving Solo around after startup.

 

[Ed Beck]

I live in the city, I get "Lost connection" often & depend on RTH for Solo's safe return.

Hey Steve,

I do a lot of my flying in WiFi rich environments and had a the same problem. I finally got this and haven't had any problems since. I've also considered getting extra power to amplify my signal.

It's a good thing I'm already married...

 

[Pedals2Paddles]

This is probably obvious but it just dawned on me. After you do a compass or level calibration REBOOT Solo before sending it up. Those calibrations certainly involve moving Solo around after startup.

That right there is an issue. The first of two people who had this problem discussed yesterday (the one that went through the window) in fact did a level calibration just before taking off because the app wanted him to do so. Here's the problem... when the level calibration completes, the app says GO FLY!. So that's what users do. There is no 3DR, solo, or arducopter instruction to do a power cycle after doing a level or compass calibration. There also is no instruction to turn it off if you have to move it around. But there are various people in the arducopter development groups and at 3DR that say you should as a good practice.

This whole thing is somewhat convoluted and really shouldn't be.

 

[twitchyzero]

i dont touch Solo immediately after bootup, but I guess one should not move it at all either when it's searching GPS? So if there's no lock after couple min, turn it off before moving to new location?

more battery juice is required!

 

[SARDG]

I have, on a couple of occasions, armed in the shade, under a tree, and then moved to launch. I did not notice any problems during those 2 or 3 flights. The reason I was booting/arming in the shade, and then moving Solo over, is because the heat in the sun lately is ridiculous - not only for me, but particularly Solo and GoPro. I was just trying to get a few flights without GP shutting down. Yesterday was 95 temp and 105 heat index and I didn't even try to fly.

I'm mostly using Rev B/cardboard bird and don't have any real problems getting a lock under the typically lonely shade tree.

 

[SteveReno]

That right there is an issue. The first of two people who had this problem discussed yesterday (the one that went through the window) in fact did a level calibration just before taking off because the app wanted him to do so. Here's the problem... when the level calibration completes, the app says GO FLY!. So that's what users do. There is no 3DR, solo, or arducopter instruction to do a power cycle after doing a level or compass calibration. There also is no instruction to turn it off if you have to move it around. But there are various people in the arducopter development groups and at 3DR that say you should as a good practice.

This whole thing is somewhat convoluted and really shouldn't be.

the Difference Between Moving Solo to find better GPS & Level or Compass calibration, is Both Controller & Solo resets some registers, & then completes initializing.
Do Not move Solo after initialization. It will cause a "failure to communicate"

Dropbox - failure to communicate_CLIPCHAMP_480p.mp4​

 

[PdxSteve]

Sorry, had to lol:

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies