0.4 Hz bandwidth of PID controller performence would quite danger ?

[ammjy]

Good day, Dear respected EEVblog seniors !
I am so sorry that always only asking...

If you do not mind, I would like to ask you about bandwidth of PID controller.

The control object which I'm looking at right now is a large multicopter weighing about a 1 ton.
And with the current PID gains, it has a 0.4 Hz bandwidth control performance from bode plot.

For missiles or small UAVs, I can surely say that 0.4 Hz is insufficient and can lead to dangerous situations.
But, I am not really sure because, this kind of large control object is first experience to me.
How much bandwidth would be fine for this kind of large object ?

As a rookie who first started job in this field, I would like to ask the opinions of the seniors.
Could I possibly get some valuable advice from you ? Please.. let me know.

[moffy]

Well if 0.4Hz is around 2 second delay, with gravity 9.8m/s^2 that is 19.6m fall, saying you go from a stall in mid air, but that isn't quite what you are talking about, because it has probably to do more with the ratio of the thrust of the lift motors to inertia of the vehicle. I would guess that winds and any type of impulse disturbance could be a real issue. Depends on conditions, but anything other than perfectly still air could be an issue. Nasty.

[ammjy]

I deeply appreciate for your advice, Mr. Moffy.

(I am so sorry, If my understand about your words are not correct, Please tell me.)

Yes, right !
If focusing on controller bandwidth problem about alttitude channel, affect of losing performance will be started from altitude error, so it is not a big deal I think..
Because, losing performance is start of gradual increase of error in altitude channel such as tracking 9m from 10m command.

But, I am still not sure what will be the result of the 0.4Hz bandwidth in attitude control including disturbance and if so how much..

[moffy]

You have to model the disturbances and introduce them into your control loop for analysis. The device has surface area, so to model wind loadings you need the area and expected wind gust which then exerts a force. You have inertias for the model, as well as thrust, you must have had something to come up with the PID values and then you introduce maximum wind loadings as a step force/disturbance to see how the model works. You would need to do that from various angles as the loadings would vary with surface area. Sorry I can only give general advice as I have no experience with designing drones.