Hello Fellow Yuneec Pilot!
Join our free Yuneec community and remove this annoying banner!
Sign up

Perfect level for Accelerometer calibration a myth?

Joined
Aug 6, 2016
Messages
362
Reaction score
206
Age
37
Location
Mexico
Hello guys,

So, I've been looking around some threads for some time, and I've seen that some users are on an endless mission to get the H perfectly leveled, even purchasing expensive tools to achieve an absolute surgical precision level in order to ensure a perfect accelerometer calibration.

Not only that, but also some people looking for open secluded areas very far from any buildings, metallic structures (I know these are mentioned on the manual I just mean some times its exagerated), cell phones, tablets, car keys, etc. for the compass calibration.

So I just made a video for you guys, and in a similar way as I did in the previous video I feel a disclaimer is due, please feel free to check the video description for the disclaimer.
(also similar to the one where I replaced the main power distribution board I tried to make it on a single take so you could see I was not editing or cutting anything from this process.)

Anyway, I'm very insterested in your opinion, I really hope you don't take this the wrong way, I know its very frustrating when things don't work the way they're supposed to (specially when things cost more than $1000dlls)

I just wanted to share this with you guys, hopefully it will help you will find out what's really causing drifting/TBowling problems; perhaps the perfect leveling and no devices nearby when calibrating is just giving false hope?


Greetings!
 
  • Like
Reactions: chiloschista
Accelerometer calibration does not require a "dead nuts perfectly level" condition to succeed. Everything has a tolerance and accelerometers are no different. As close to level as reasonably possible is good enough. Put it on the living room floor and don't move the darn thing during the calibration process until the system sends the light code message that's it's done. It's the easiest of any calibration to do. They need to establish a zero reference point, which is critical to have present for every boot up. During boot up the airframe can lean a little bit in any direction as long as the primary calibration established the zero reference point for the system to make comparisons with. Same applies to the gimbal. This is why permitting the H to be perfectly still for a period of time after power up is so important, if the aircraft is being moved about during the boot up self system checks there are no stationary position for the system to make comparative reference with the stored calibration data.

Lacking that reference a lot of things can happen, with one of them having the system "hunt" for a zero ref. If it was not permitted to establish on after power up you might anticipate some drifting and in severe cases "toilet bowling" if the compass calibration was not properly performed. We hear a lot about people having drift issues but what we don't read is what they were doing with the H after they pushed the power button. Did they take it for a walk, raise and lock the arms, decide they needed to clean the lens, or changing the direction it was facing? All of those things trash the initial self checks. You might also expect the gimbal to be off level because it has accelerometers too. Be where you want to be, everything ready to fly, and put it on the ground. Push the button and leave it alone. This is applicable to just about all multirotors. Leave them alone during the boot up cycle and you'll generally be rewarded with a very stable aircraft.

Those that are real smart will walk away from it a few paces, assuring that regardless of what flight mode it's in it will go where they command it to go when they push the throttle up. It's aviation, not a drag race. There's a great many reasons not to be in a hurry. On average I'd guess about 1300 of them.
 
We hear a lot about people having drift issues but what we don't read is what they were doing with the H after they pushed the power button. Did they take it for a walk, raise and lock the arms, decide they needed to clean the lens, or changing the direction it was facing?(...) It's aviation, not a drag race. There's a great many reasons not to be in a hurry.(...)

I definitely missed that point, which also makes me wonder why so many people (on youtube at least) compare drones about their set-up time, making videos about "which one takes less time from taking it out of the box to flying".
 
Agree with PatR...

When I was having issues with my first Typhoon H, I went to the trouble of buying a piece of perfectly level tile and would shim the tile till it was perfectly level. Then do the calibration, it made no difference on that Typhoon because it was defective.

On my working TH, I've done it on my slightly slanted sidewalk, my shimmed tile, on top of a plastic trashcan, eyeballing level, and its all been fine...
 
Accels don't care about level. The important thing is "DO NOT MOVE THE AIRCRAFT" Accels measure movement, Gyros measure angle.
 
Accels don't care about level. The important thing is "DO NOT MOVE THE AIRCRAFT" Accels measure movement, Gyros measure angle.
Not as simple. Gyros measure rate of rotation, accelerometers measure gravity too and therefore body attitude.
Gyros don't care about level, while accelerometers do.
 
JulesTEO,
thank you very much for that test. I agree with you (and wrote it several times ...). When people with limited knowledge are in trouble, they try to find any good motivation for those troubles, even invading esoteric fields.
The problem here is that no one knows exactly how that multirotor works.
How do they use IMU? Is it a real 10DOF?
How does the IMU drift? How are those drifts managed?
I have plenty of other questions, especially related to GPS and all the jokes around it.

By the way, lucky you, because if I should wait for 20 satellites I never go fly. I live between mountains and having more than 10 is already a joy.
More than the satellites number, there should be a 3D fix. What does the TH set a GPS fix? (lot of questions about that also).

Making it short I did want to do a similar test too, recording accelerometers offset with the TH off level by a lot: 10-20° and see how it behaves.
But the problem here is the inconsistency of the TH behaviour. It is not constant. I'm actually recording few good flights and then suddenly a crazy flight, another few very rock solid flights and then the "I don't want to land" flight. I know a little those chickens, so until now I was able to manage it, but that's annoying.

Best regards,
Ric
 
Last edited:
A larger concern with our GPS is the standards employed during the manufacture of our GPS systems. Are there any or are they dependent on the mood of the cheapest maker? One thing is certain, we are not using high quality, FAA certified, aviation grade equipment. To expect that kind of performance in what is essentially toy or somewhat elevated hobby grade grade equipment and their associated price point is ludicrious. That some of what we get are not as good as they could be should be expected, and owners should notify Yuneec of any issues as soon as they become apparent. Unless one is a certified avionics repair technician they will waste a lot of time and effort in troubleshooting something they are not qualified to do.


Sent from my iPhone using Tapatalk
 

New Posts

Members online

Forum statistics

Threads
20,954
Messages
241,586
Members
27,284
Latest member
csandoval