I've been building a 250 quad for the last few weeks, which is a lot of fun. However, when I'm trying to fly it, it doesn't seem to fly as it should. When I just raise the throttle, it takes off, but not in a clean vertical way. It seems to drift, mostly to the right and the back. To my understanding, when I just use the throttle, it should take off in a perfectly vertical way?

It's a 250 quad, with an SP Racing F3 flight controller. I've balanced the props, and calibrated the ESC's and the accelerometer of the FC. 

What am I doing wrong? What could be a solution?

There's a few ways how this could happen. I'll start with the simpler ones:

Did you adjust your transmitter subtrims so channels center exactly at 1500?

In case you're using angle mode, this could happen if accelerometer was not calibrated perfectly level. You could adjust acc trims to fix this.

do you have footage showing the issue?
1. check your COG is centred
2. make sure all channels on your TX are centred at 1500 when idle, trim it if not.
3. calibrate your ESC's
4. if you fly in self-level mode, also trim the ACC
5. you can try fly in acro mode see if it drifts the same way

(08-Jun-2017, 04:42 PM)fftunes Wrote: There's a few ways how this could happen. I'll start with the simpler ones:

Did you adjust your transmitter subtrims so channels center exactly at 1500?

In case you're using angle mode, this could happen if accelerometer was not calibrated perfectly level. You could adjust acc trims to fix this.

(12-Jun-2017, 01:06 AM)Oscar Wrote: do you have footage showing the issue?
1. check your COG is centred
2. make sure all channels on your TX are centred at 1500 when idle, trim it if not.
3. calibrate your ESC's
4. if you fly in self-level mode, also trim the ACC
5. you can try fly in acro mode see if it drifts the same way

Thanks for the tips,

1. I have build the quad symetrically as most as possible, so it should be centred pretty good. But how to test this precisely? Maybe suspend the quad on a string?
When the quad is drifting to the back, is the COG too much to the back? But in my understanding the FC should compensate this by running the rear motors harder?

2. All channels are centred at 1500, I'm using an ibus serial connection between RX and FC.

3. I have calibrated my ESC's, using BLHeli. But I'm reading some contradictory material here and there; is this the right way to calibrate?

4. How should this be done? I have calibrated the ACC various times. One thing that surprised me though, is that the Gyro isn't completely stable when the quad is stationary (motors not running). The XYZ values change very fast, constantly. But the variations are small, never beyond 1. The ACC seems to be stable. Of course, on the Z-axis there is 1G measured when stationary, which seems right to me  However, when the motors are armed, the ACC and Gyro go quite crazy. Probably due to the vibrations. 

5. I have only tried Angle and Horizon modes. In Angle mode the drifting seems worse than in Horizon mode. In Acro mode, it seems to drift mostly the same way, but I haven't been able to test this very well due to time constraints. Lastly, I was wondering whether a quad maybe never takes off completely clean, because of turbulance when hovering just above the ground (first 30 cms)?

Add; I also observed the quad is bouncing, when I'm landing it. Now, I have used flexible landing skids, but is that normal?

When I have some time in the coming weeks, I will rebuilt the quad and install fresh props. Also, I think I need to find a way to dampen the vibrations of the motors to the FC. I will also make a video then, if the problem still persists.

(20-Jun-2017, 09:58 AM)reynard80 Wrote: Thanks for the tips,

1. I have build the quad symetrically as most as possible, so it should be centred pretty good. But how to test this precisely? Maybe suspend the quad on a string?
When the quad is drifting to the back, is the COG too much to the back? But in my understanding the FC should compensate this by running the rear motors harder?

2. All channels are centred at 1500, I'm using an ibus serial connection between RX and FC.

3. I have calibrated my ESC's, using BLHeli. But I'm reading some contradictory material here and there; is this the right way to calibrate?

4. How should this be done? I have calibrated the ACC various times. One thing that surprised me though, is that the Gyro isn't completely stable when the quad is stationary (motors not running). The XYZ values change very fast, constantly. But the variations are small, never beyond 1. The ACC seems to be stable. Of course, on the Z-axis there is 1G measured when stationary, which seems right to me  However, when the motors are armed, the ACC and Gyro go quite crazy. Probably due to the vibrations. 

5. I have only tried Angle and Horizon modes. In Angle mode the drifting seems worse than in Horizon mode. In Acro mode, it seems to drift mostly the same way, but I haven't been able to test this very well due to time constraints. Lastly, I was wondering whether a quad maybe never takes off completely clean, because of turbulance when hovering just above the ground (first 30 cms)?

Add; I also observed the quad is bouncing, when I'm landing it. Now, I have used flexible landing skids, but is that normal?

When I have some time in the coming weeks, I will rebuilt the quad and install fresh props. Also, I think I need to find a way to dampen the vibrations of the motors to the FC. I will also make a video then, if the problem still persists.

Hey buddy, 

What you say about the quad bouncing, would be nice to see a video, but I had a very similar problem, and it was caused by having the barometer enabled (this is crap, disable that as it doesn´t help much). especially when you are close to the ground, you will have lot of extra wind coming and going around the FC affecting the barometer, and it can behave erratically. I struggled on that for a week, tried to cover the FC with foam, helped but did not resolved the problem. then I disabled this feature and never again had to deal with this.

This is something similar that can also affect professional drones like Mavic and  Inspire when flying over water, the humidity can vary and your bird can turn into a non-nuclear submarine :-)
Have a try and see if at least this problem gets resolved by doing this.

Thanks

(22-Jun-2017, 03:17 AM)flabombarda Wrote: ...
What you say about the quad bouncing, would be nice to see a video, but I had a very similar problem, and it was caused by having the barometer enabled (this is crap, disable that as it doesn´t help much). especially when you are close to the ground, you will have lot of extra wind coming and going around the FC affecting the barometer, and it can behave erratically. I struggled on that for a week, tried to cover the FC with foam, helped but did not resolved the problem. then I disabled this feature and never again had to deal with this.
...

I have been too busy with other things for a while, but have tried some stuff today. Disabling the barometer didn't help unfortunately. 

I've added a which shows the problems. The quad does not take off neatly but immediately starts drifting. In the video, when I lower the throttle, the quad starts bouncing heavily.

The bouncing on landing is quite normal... If it hits ground while not being perfectly levelled, the opposing motors will try to correct for the movement caused by this, which can result in an upwards push. The effect will be more noticeable with higher PID. Best approach for landing is to hover as low as possible above ground and then disarm and let it fall.

For the acc drift: there's the acc trim option in the gui to correct this. However, it could also be caused by vibrations, in that case you probably have to softmount the FC.

(04-Aug-2017, 05:08 PM)fftunes Wrote: The bouncing on landing is quite normal... If it hits ground while not being perfectly levelled, the opposing motors will try to correct for the movement caused by this, which can result in an upwards push. The effect will be more noticeable with higher PID. Best approach for landing is to hover as low as possible above ground and then disarm and let it fall.

Ok, didn't know that. The effect seems quite 'scary'.

(04-Aug-2017, 05:08 PM)fftunes Wrote: For the acc drift: there's the acc trim option in the gui to correct this. However, it could also be caused by vibrations, in that case you probably have to softmount the FC.

I already softmounted the FC. The strange thing, is that the drifting seems to be in a different way every time I run the quad. It seems to be much worse in Angle and Horizon modes.

Maybe you also to careful with pushing the throttle down. If you almost make the first bounce just cut the throttle and it will perfectly land. It's just practice.

(07-Aug-2017, 09:51 AM)reynard80 Wrote: I already softmounted the FC. The strange thing, is that the drifting seems to be in a different way every time I run the quad. It seems to be much worse in Angle and Horizon modes.

How much have you tried to adjust the PID? Theoretically, in acro mode it shouldn't drift at all. Maybe try acro mode first with some higher I-value.

Not sure if somebody else mentioned it before, but maybe you need to increase deadzone a bit as well - even if channels look well centered in receiver tab, there can be small jitter that results in drift.

Auto-level mode?
Just trim your accelerometer should fix that.