Drone stays up in the air at zero throttle sometimes (even climbs).

[dark holly]

I’m running into a strange problem. Motor #2 seems to be way more “aggressive” than the others — it feels like it’s producing much more thrust. Because of that, even at zero throttle, the other three motors are spinning up hard to maintain attitude, and the quad ends up holding altitude (or even climbing) at zero or very low throttle.

When it’s in this “won’t drop” state, the RPM reported from motor #2 is only around 4600 rpm. This makes me suspect an ESC issue.
And it seems Motor #2 tends to stay at a lower RPM during flight.

In Betaflight, the “motors” tab curve is the FC’s power request to the ESC, right? Could it be that the FC isn’t requesting much power, but that motor is actually spinning much faster and sending back incorrect RPM data? Seems unlikely, but uhhh.. I can’t rule it out.

Has anyone seen something similar, or have ideas on how to troubleshoot this?

[tepid prairie]

THis appears to be a move you did.
The quad is flat and you pushed throttle, so it climbed.
While climbing, you chopped throttle to zero.
The quad then needed to increase some motors in order to stay stable. There was a strong buildup in iTerm that was trying to stabilise the quad. That stabilisation required more of some motors than others, so one stayed at idle.
The RPM is consistent with the motor drive signals.
So figure out why it doesn;t drop symmetrically.
either some imbalance or drag problem.

[dark holly]

Well, that’s what got me thinking… I’m pretty confident the frame is balanced — the CG is almost exactly at the center of the quad. Nothing was dragging it down either. When I pulsed the throttle, the quad shot straight up into the sky, and there were definitely no obstacles around that could have been “dragging” it.

[queen lion]

This can definitely be due to your ESC settings. What ESC firmware are you running? And what's the size of the quad and motor KV?

[dark holly]

I’m currently using Bluejay 0.21.0, the latest version. It originally came with 0.19 from the factory, but after I received the drone I upgraded it and restored all the parameters.

[queen lion]

Well on bluejay there's fewer things it can be. Your motors may not have enough braking force causing them not to be able to slow down adequately which can be the cause of why your quad doesn't want to come down.

For me, it was the Complimentary PWM setting on an AM32 based ESC that when turned off, would cause my quad to not want to come back down, especially on lower PWM values like 24khz. On bluejay, I believe the setting is Brake on Stop. You'll have to turn it on and see if it fixes your quad not wanting to come back down.

[dark holly]

The “Brake on Stop” option has the description “Brake when motors are stopped”. I think this means the ESC will short the motor windings to actively brake when disarmed (instead of just letting them freewheel)?

Oh, and I was running at 48 kHz PWM frequency. Do you think I should roll back to Bluejay 0.19 and give it a try? I haven’t really tested much on 0.19 yet 🤔

[queen lion]

I stand corrected, I'm thinking of Active Braking. Not sure if that setting can be changed on the current versions of Bluejay though, as I don't recall seeing a setting for that.

[dark holly]

ah maybe there's no that option

[queen lion]

If it didn't have this issue before updating, then I'd go back.

[dark holly]

I'll try😂

[dark holly]

oh after that there a new setting (or older you may say ) called Maximum braking strength (Complementary PWM duty cycle).

[queen lion]

You'll want to leave that maxed out for quads