[APF]

Hello everyone,

I've build a Quadcopter with a CC3D board and everything works quite fine but for a small issue: every 30 seconds or so the engine power increases for some 2-3 seconds (together with a tendency to turn right) until it settles down on the former level. This is probably not noticable in outdoor flying, but as I'm an attic flyer right now (weather is too bad and surrounding is too clogged up with buildings etc.) I have to feather the throttle until the power surge settles back to normal again.

Is there any possibility to send telemetry data of control inputs and PWM outputs to check wether it's an issue of the inputs?

Greetings and happy new year!

APF

[TheOtherCliff]

That is the kind of thing that is best tackled by starting with a look at telemetry when it happened.

Generally speaking, you need Revo+OpLink or two OpLinks for the RF telemetry.

I would look at ManualControlCommand.Throttle to see what the receiver is saying, and ActuatorDesired.Thrust to see what the FC has changed that to.  Also, make sure it is not entering RC failsafe or FC failsafe.

[sasquad]

Give us some more information. Motors, speed controllers, radio, receiver used, Battery and cell count. This will help us understand your set up and how to assist you. Also, when setting this quad up did you calibrate the speed controllers ? I've seen this before and it turned out to be a faulty speed controller. Not saying that is the issue here, but we need more info.


Thanks,

-Bill

[APF]

Sorry for the late reply but I first had to organize a serial->BT transmitter for telemetry and a new set of LiPos (the old ones were several years old and got too warm for my liking).

So, as to the setup:

Motors: Multistar Elite 2204-2300KV
Props: 5x3
ESC: Afro ESC 12Amp Ultra Lite Multi-rotor Motor Speed Controller (SimonK Firmware) Version 3
Radio: MPX Royal Evo 9 (2.4 GHz upgrade)
Receiver MPX RX-9 DR
Battery: 3s 2.2Ah LiPo 30C

Speed controllers are calibrated.

... and now with telemetry I see that #2 some motor frequently revs up. I'd like to put some telemetry data here but I didn't succeed with saving anything. Shouldn't the data be saved if settings->scope->[any scope] log data to csv file is checked? But I can't find any file(s) afterwards.

Oh, and I guess it's much nore frequent - like any 5 seconds or so. But I'd like to get some telemetry files to check befor I say anything more specific.

Greetings

[TheOtherCliff]

For starters, if you have a good looking scope, you could post a screenshot.  That is easier to look for initial issues.

About 3 comments come to mind:
- I hope it isn't hardware (or maybe I hope it is so that it isn't a bug)
- Are you aware that when it is sitting on the ground with motors idling, and in attitude mode, one or more motors will get faster and faster, trying to level it out, when it can't because it is "stuck" on the ground.
- I have a similar problem and the work around that I found hasn't helped anyone yet, but you can try it.  I'm not even sure it is valid for the CC3D or whether it will ignore it.  Go to System -> Settings -> MPUGyroAccelSettings -> FilterSetting and set it to LowPass_42_Hz.  Click FilterSetting to turn it blue.  Press the red up arrow at the top of the page to save it permanently.  Power off and on and carefully try to fly it and see if it happens.  Put it back where it was if it doesn't help.  You can also play with Configuration -> Attitude -> Accelerometer (I recall that this should be 0.05, but is 0) and Configuration -> Stabilization -> Expert -> Gyro Noise Filtering

[APF]

Turned out that the attic where I tested the thing produced some really long-lasting vertices or so against which the Quad tried to fight (note to self: rudder authority is much smaller than with a copter). The issue is much less pronounced (if it exists at all) outside the house.

Oh well, I added a screenshot where you can see that #0 and #2 both rev up repeatedly for quite some time while I tried to keep the control inputs steady.

.. and finally a photo of the quad.

Yes, the special features are on behalf of my son

EDIT: as to the questions:
- looks as if the HW is ok.
- yes, I saw the thing with a single engine rev'ing up if the Quad sits idling on the floor, but as you said I put this one to the I-term of the PID controller.
 - I changed the filter settings but this didn't help. But those air streams are so long-lasting you'd probably need a mHz filter which would send the Quad into the ground before it'd react.

[TheOtherCliff]

That picture of the scopes is very helpful.  You obviously put some time and thought into this issue.

Any time you see a pair of diagonally opposite motors speed up while the other pair slows down, you are seeing a yaw command.  I can see from the picture that you get about a 5 second period where one pair speeds up and the other pair slows down, yet you did not give it a yaw command.

Edit:  Oh, was this flying or were you holding it in your hand?  If you were holding it in your hand, that is probably the whole yaw issue discussed below.  Even a fairly loose USB cable attached to GCS on ground can pull a bit of yaw into it in flight.  In a quad, yaw is a very underpowered control, and even holding it with a light touch will cause it to yaw hard.  That is what fits the facts the best of all.  What does it do when actually flying with no cable attached?

Does it drift much in yaw?  What stabilization mode are you using for yaw?  We generally recommend "Axis Lock" for yaw to add some stability to yaw.  Edit: And axis lock tries even harder than rate to fight your holding hand.

I can see that it consistently needs to run (zero based numbers) motors 0 and 2 at a higher level than 1 and 3.  Two possibilities come to mind:
- You are using a different brand/model/size of CW prop vs CCW prop (or different motors or ESC's)
- Loose props.  This one sounds a little more likely if it only happens intermittently.

Other notes.  For best performance, you should be running ESC's that can handle at least PWM 490Hz updates (and have it configured that way too) and with active braking (light damping) turned on.  Without light damping, the motor speed up happens more quickly than the motor slow down, and directional commands (caused by you or stabilization) cause the quad to rise because one pair of motors is commanded to slow down 20% and the other pair to speed up 20%.  The speed up happens much faster, and what you have is e.g. -2% and +20% for a net increase in thrust of 9% when the controller thinks it should be getting a 0% net increase.  This is very noticeable when you command yaw back and forth.  Try it. 

So maybe we have at least narrowed down your issue to stabilization causing yaw commands and the lack of light damping in the ESC's making that into an aLtitude gain.

[APF]

[...]
Edit:  Oh, was this flying or were you holding it in your hand? 

Nope, it was during free flight (telemetry via BT-serial adapter).

Does it drift much in yaw?  What stabilization mode are you using for yaw?  We generally recommend "Axis Lock" for yaw to add some stability to yaw.  Edit: And axis lock tries even harder than rate to fight your holding hand.

Right now I use Rate for yaw, but I've a setting which would switch to axis lock. Will try tomorrow.

I can see that it consistently needs to run (zero based numbers) motors 0 and 2 at a higher level than 1 and 3.  Two possibilities come to mind:
- You are using a different brand/model/size of CW prop vs CCW prop (or different motors or ESC's)

Not, they are both form the same package and look the same. Well, like mirrored twins, obviously.

- Loose props.  This one sounds a little more likely if it only happens intermittently.

No, the props are locked tight. And the spinners are mounted to provide auto-tightening as well.

The speed up happens much faster, and what you have is e.g. -2% and +20% for a net increase in thrust of 9% when the controller thinks it should be getting a 0% net increase.  This is very noticeable when you command yaw back and forth.  Try it. 

I'm not so sure about that. The engines are braked by the props after all. But I will try to make a video tomorrow and see whether I can show the effect on film.

Greetings

EDT: sorry, no video today: the first try was miserable and we had an invitation to visit usually off-limit airport repair facilities today which had priority.

[APF]

Apologies for the late reply but I needed several attempts until I figured out how to make a video which showes both the telemetry data and the quad:



At 0:51 and 0:57 the sudden speedup of the motors 1 & 3 and the yaw of the quad can be seen.

Greetings