[HELP]Quad drifting backwards

[OPO]

Hi Guys,

After finishing my build I tried to fly my 250 quad but first it tried to fly to the left all the time, Jhitesma suggested me to flash my CC3D with the TauLabs software, after doing this the quad is a bit quiter but when I try to fly this happens:

(This also happens when I am higher in the air)

The CG is right where the FC is and I did the 6 point calibration in the Taulabs CGS. I tested it with leveling and Acro mode, in both modes it still happens.

I tried using some trim but that won't since I gave it full trim and it still tried to fly backwards.

 

[makattack]

Does it do that in rate/acro mode? You may need to do a level trim type setup if Taulabs supports that. It could be that the accelerometer calibration wasn't done perfectly level or still if it doesn't do this in rate/acro mode.

 

[OPO]

It does the same in all modes, I have done the calibration like 2 times and it keeps doing the same. I have no idea what the issue could be

 

[Ace2317]

Where is your board placed on the quad? If it's too far back, that could be causing it.

 

[OPO]

Where is your board placed on the quad? If it's too far back, that could be causing it.

It is placed a bit to the back yes, I am going to rebuild the quad and see if that solves and issues, I have my PDB on the bottom side of my quad so I have to change that also.

Do you really think that would make it fly like that?

 

[Ace2317]

Potentially. Your control board is designed to think that it's at the center of the quad. I would definitely start there because it will definitely always affect the flight in Auto-level. Get it as centered as possible and see. It may not be it, but I would start there if it were me.

For a quick fix, you can push your battery further forward and maybe it will change it too. Get it a little ahead of the board. Keep in mind, though, that in auto-level the system will constantly be fighting itself, thinking some outside force is pushing it forward so it will be trying to push back.

 

[RichB]

If it does that in acro mode, then your transmitter center points needs to be adjusted.

If in auto-level mode, then you need to calibrate accelerometers, then trim the board for level flight.

Location of the flight controller on the frame is of no consequence here, so don't move it or rebuild the frame.

Get the flight controller connected to the computer so you can view your RC inputs to make sure all channels are centered when your are hand off the sticks. Once you have that sorted, do the flight trimming.

NEVER adjust transmitter trims for a multirotor. Get them centered and leave them there. If your transmitter can only adjust center points using trims, then you can do that but then NEVER touch them again, even when troubleshooting problems like this.

 

[joshuabardwell]

I maintain that FC position in rate/acro mode won't cause issues like this, because if you assume a rigid body, the amount and rate of angular change should be the same for all parts of the body. Take the simple example of a straight line. Choose an arbitrary point on the line. Cause that point to undergo a certain angular change--five degrees CCW, for example. All other parts of the line must also undergo the same angular change, over the same amount of time, or the line will bend.

Since copters are not perfectly rigid, it is more likely that vibrations will affect the FC if the FC is at a point far away from the CG, such as at the end of a motor boom. But nobody puts their FC at the end of a motor boom.

I assert that, in a self-level mode, having the FC away from the CG will cause coupling of inputs. For example, you yaw, and the copter also rolls to the side and begins translating. But having the FC away from the CG should not cause the FC to fail to return to level, assuming that the accelerometer is correctly calibrated.

I agree with Rich that the first thing you need to do is check your inputs using whatever monitor software is correct for your FC.

 

[PhenomPilot]

What RichB said is true for my naze32 board. Trim the transmitter in acro mode, then if it is drifting in autolevel use the trim on the board. On the naze32 there is a stick command for this. Disarm, full throttle, then flick the right stick a few times in the direction you want to trim. Don't know if it is the same for cc3d but you might have something similar.

 

[OPO]

The switcks are centered when I look in the software, but I am using a PPM rx I couldn't get my orange rx to work, could this be the cause of all my issues?

 

[makattack]

What does the ground station software show for your rcchannels in the taulabs software?

 

[OPO]

Those are correct right?

 

[makattack]

You got those values after running the configuration wizard? If so, it does seem like the pitch channel is the same as the roll and yaw, so that does look good.

Puzzling. I need to defer to other experts.

Incidentally, what tx/rx combo do you have? Seems like some channels were reversed by the setup, but maybe that's how taulabs worKs.

 

[RichB]

My operating theory, after seeing your screenshots, is that your FC has a pretty good idea of your transmitter's center points and end points, and therefore this all comes down to accelerometer calibration.

Try again to calibrate your accelerometers while the craft is sitting still and level.

After that, if it still drifts backward, then try this (works on MW-based FCs. not sure what yours is).


Disarm
Push throttle stick to full throttle.
Push right stick all the way forward for 10 seconds
Pull throttle stick to zero throttle.
Arm and test.

If that helps, then repeat it. If you go too far, then pull the right stick backward for 5 seconds, or 1 second, whatever.

 

[makattack]

Have your tried setting the level trims as well?

https://github.com/TauLabs/TauLabs/wiki/Level-triming

It's odd this occurs in rate/acro mode, but I don't know enough about how that's implemented in taulabs.

 

[OPO]

You got those values after running the configuration wizard? If so, it does seem like the pitch channel is the same as the roll and yaw, so that does look good.

Puzzling. I need to defer to other experts.

Incidentally, what tx/rx combo do you have? Seems like some channels were reversed by the setup, but maybe that's how taulabs worKs.

I am using a Dx6i with the Lemon 6 channel DSM2 Rx with PPM output, I cannot get my Orange rx from hobbyking to work on my FC. I had to reverse some channels myself.

My operating theory, after seeing your screenshots, is that your FC has a pretty good idea of your transmitter's center points and end points, and therefore this all comes down to accelerometer calibration.

Try again to calibrate your accelerometers while the craft is sitting still and level.
...

I will see if this works, thank you for the suggestion.

Have your tried setting the level trims as well?

https://github.com/TauLabs/TauLabs/wiki/Level-triming

It's odd this occurs in rate/acro mode, but I don't know enough about how that's implemented in taulabs.

I have not done this yet all I did was auto level and the 6 axis calibration

 

[jhitesma]

Those settings look reasonable...but I'm curious what your RX is actually seeing. This is one thing I'm not big on about Tau is that there isn't a quick easy way to see your actual real-time RX values without going into the object browser (unlike BF/CF/MW where it's right there in the main GUI screen)

You'll want to go to a screen that has the object browser on the right side. I believe all of the configuration screens have it by default, but there are different ways you can configure GCS and I haven't played with all of them yet so I'm not sure just which screens will have it for you.

In the object browser you'll want to expand:
Data Objects -> Manual Control Command -> Channel

In there you should see something like:


Things will be flickering orange as they're updated from the FC. The values there are the values the FC is reading from your RX in real time. So with your throttle at rest and sticks centered you should see values very close to those you posted on your input setup screen.

If not...then that's the problem.

If they are very similar...then something really goofy is going on and I'd probably suggest posting on the TL Forums (http://forum.taulabs.org/) and seeing if someone there has any ideas.


You shouldn't have to do the level trimming yet...if it's drifting this bad in acro mode then something bigger is going on still. Level trimming is really for making minor changes to fine tune the autolevel leveling. If it won't hold attitude in acro then something else is going on.

 

[OPO]

Okay so... I rebuild my quadcopter because I got mad that I couldn't fix this issue thinking it was because I didn't build it properly. I am 100% sure it's build like it should be, I placed the FC in the middle using nylon standoffs, I used a 1000 3cell battery I used new props, I redid the calibration I calbirated my ESCs again. But it did not solve the issue.. With the OpenPilot software it only drifted to the bottom left, could it be that there is something with my FC?

Here is a video of the data objects:

(I am not touching the sticks at all)

The first takeoff is level mode, the other ones are all acro I couldn't get the quad to go up stable enough it would just drift around, but you can see it's also drifting to the back.

 

[cranialrectosis]

You need to pop the copter up off the floor 1'-2' to get it out of its own wash so we can see what it is doing.

I don't see anything out of the ordinary. If I fly 4" off a hard tile floor, my copter will be uncrontrollable, or nearly so, too.


Take it outside. Get it up in the clean, free air where it wants to be.

No matter what, you will have to fly the copter. You will never be able to take your hands off the sticks and not have it drift off. To do stuff like that, you need a good GPS and barometer for altitude and position hold.

 

[OPO]

You need to pop the copter up off the floor 1'-2' to get it out of its own wash so we can see what it is doing.

I don't see anything out of the ordinary. If I fly 4" off a hard tile floor, my copter will be uncrontrollable, or nearly so, too.


Take it outside. Get it up in the clean, free air where it wants to be.

No matter what, you will have to fly the copter. You will never be able to take your hands off the sticks and not have it drift off. To do stuff like that, you need a good GPS and barometer for altitude and position hold.

I will try to fly it outside as soon as the weather plays nice, it has been raining and snowing a lot dont want to short out my quad first time I fly it outside haha. But even if I get it high enough from the floor it still drifts backwards

EDIT:

Okay so I tried to record it as good as I could. When the quad is in the air it only drifts a little bit to the back, is it okay if I use trims to solve this?

In acro mode I cannot get the quad high enough in the air to see if the drift also happens, it just flops around.