[kenneth]

I am using Revolution in a 600mm drones.
I finished all the settings and tried to take off.
The drones rise as they move sideways, not vertically.

I have tried Calibration several times and tried changing the motor output.
What changes would make normal takeoff and hovering possible?

I have already checked the following points.
- FC mounted upright and with arrow pointing forward.
- Each ESC plugged into the correct numbered ESC connector on the FC.
- Each ESC to motor connection done correctly so that the motor spins in the correct direction
- The correct type of prop (CW=R or CCW) on each motor.

[Mateusz]

• Is that your first build ?
• Are propellers balanced ?
• Is flight-controller mounted without vibrations ?

Start with simple Complementary (Basic), calibrate your accelerometer by having Revo in the case on a desk, and rotating it accordingly to instructions (align with surface of flat desk/table).
Gyro you can calibrate after mounting FC in the frame and having it motionless. Temperature calibration is used for two sensors simultaneously baro and gyro. Although in the first step you won't be using baro yet, it's good to have it done properly. So when you do temp calibration, cool down (~20min) board in fridge, then close in the dark box and keep board in darkness (baro is light sensitive) and board without any motion (gyro is motion sensitive). Let temp calibration complete took ~40min in my case.

When you take off, first few centimetres it will go sideways but once you lift-up it should stay in hover with very small drift if everything is well balanced and there are no severe vibrations.

[kenneth]

It is the first build.
The propellers are well balanced.
The FC was mounted without vibration.

Accelerometers and gyros were calibrated.
However, the temperature was not calibrated.
Is there a possibility of temperature calibration being a problem?

If it takes off quickly, it is only a few centimeters, but if it takes off slowly, move a few meters. Is this normal, depending on the speed of takeoff?
I have seen many hovering up vertically without movement in other drones.
So it does not make sense whether this is normal.

And if the movement radius is wide after taking off, is it due to vibration?

Thank you for your reply.

[Mateusz]

It is normal that when you take off too slow, copter will be unstable when close to the ground. You have to take off immediately.
When you take off too slow, flight controller is trying to keep copter level but it can't so it's applying more correction and it's becoming wobbly.

[kenneth]

I upload a test video. Is it possible to check what status it is?



The beginning of the video is when I slowly raise throttle.
Move sideways along the ground.

The second half of the video is when I raise throttle fast.
The quad rises to a tilted position.

Thank you for your advice.

[TheOtherCliff]

The second one looks good.  It looks flyable. 

Are you flying in Stab 1 (Attitude mode)?  And have you done Attitude -> Calibrations -> BoardLeveling?  If so, and if you have not changed the trims since you ran the transmitter wizard, it should be controllable.

(No Wind!)  Maybe take it outside to a grassy area with a piece of flat cardboard to take off from and soft grass to land in if there is a problem.  Your first task is to maintain altitude, not let it get more than 1m high or less than .5m high.  You may have to set it down if it drifts away.  Once you can maintain altitude, you need to start correcting the drift.

I would try to arm and take off even quicker than that, but what you did is OK and not so bad. 

[kenneth]

The Quad Copter flew in attitude mode.

After board leveling calibration, I confirmed that it was leveled. It seems a little better.

I tried to fly outside.
However, the quad-copter still does not rise vertically.
I upload additional videos.
I capture the flight data and upload it together.



What am I supposed to do more?
Thank you for your kind comments.

[TheOtherCliff]

To explain, expecting a quad to rise vertically (without GPS) is like expecting a car to stay in the center of the lane (without autopilot).

What I see there actually looks fairly normal.  There is wind blowing down in the middle that rises at the walls and that makes it even more unstable.

Did you try and steer it away from the wall before it crashed?

Outside over grass with no wind is the best place to learn to fly. 

Once you get it hovering outside (no wind), and determine that for instance it always drifts forward and left, you can fine tune it with Attitude -> Settings -> "Rotate virtual...".  Wiki should tell you how.  You probably need to use search.

[TheOtherCliff]

Adjust Configuration -> Attitude -> Settings -> RotateVirtual

if it is drifting forward you subtract from pitch
if it is drifting left you subtract from roll

With the amount it is drifting I would probably start with 2 degrees each.

[kenneth]

I managed to avoid hitting the wall.

As you explained, I thought there was a problem with the test inside.
So I tested it in a wide outdoor area.

The flight proceeded without any major problems.
Thank you very much.



Vertical rise or drift problem, as you explained, tried to use GPS.

Before I went outside, I proceeded to magnetometer calibration as follows.

https://librepilot.atlassian.net/wiki/display/LPDOC/Sensor+calibration

However, when I set the GPS outside, it was not displayed in Flight Data. I thought it was caused by an alarm on MAG.

So I did the magnetometer calibration again, but there was still an alarm on the MAG.

I use the Ublox M8N GPS Module NEO-M8N.

What is the problem?
Upload relevant videos.

Thank you for your advice.

[TheOtherCliff]

GPS flight requires twisted power wires; battery to PDB, PDB to ESC, ESC to motor.  An argument could be made that it requires some form of external mag (typically built into the GPS unit).  LP supports DJI GPS/mag.  Authentic DJI GPS/mag can be bought for about $37 shipped (eBay, search the LP forum).  Clones about $10 less than that.

You should use INS13 Attitude Estimation.  You must set Home Location.  Take off in Attitude (Stab1) before switching to a GPS mode.  Use VelocityRoam for testing / playing with / using GPS.

You need to get Attitude flying well and get used to flying it before using GPS.  You may need to switch out of GPS and handle it yourself.  Just like you can't send an 8 year old out to the store with your car just because it has autopilot.

[kenneth]

I twisted the power line like the picture.
The alarm frequency was reduced, but the MAG alarm still occurred.

I used the INS13 Attitude Estimation.

I do not necessarily have to use external mag, right?
Is there any problem with "Ublox M8N GPS Module NEO-M8N"?

If the power wire is close to the sensor as shown in the picture, could it be a problem?
If so, can I connect only USB to FC and proceed with MAG calibration and reconnect ESC, RF, and receiver?

Please answer.
Thank you.

[TheOtherCliff]

It's the high current wires that are the issue, not the signal wires.

The wire pairs themselves need to be twisted together like strands make a rope, before mounting on the frame, not just wrapped around the arms. 

I see that the ESC's are mounted very close to the Revo, and that is a big problem.

I would guess that on that quad, you would have to have an external mag on a full length pole (about 16cm) plus twist the high current wire pairs correctly in order to get GPS flight working well.  About half of the brands of I2C mags that I bought did not work well with long wires.  Perhaps they would work better with pull up resistors (maybe 3.3k to 2.5k) (to 3.3v) added in the GPS/mag on the I2C SDA and SCL lines.  To avoid I2C signal problems, I am using DJI GPS/mag or OP GPSv9 GPS/mag on my new builds.  They do not use I2C.  You can't buy GPSv9 cheaply any more if you can find them at all.

Good luck!