A-mode drift

[dwallersv]

Doing some skill development and practice flying around indoors in A-mode.

P4 consistently drifts slowly sideways to the left, requiring constant correction input to stay in place. switch to S or P mode, the positioning system kicks in, and it stays in place. Back to A, very consistently, very constant rate slide to the left.

Does this indicate a need for IMU calibration? Or are the 6-axis IMU sensors not active in A mode as well?

This would seem a bad design decision... the IMU sensors are always available, just like the barometer, which we know is used in A mode for maintaining elevation when hovering.

 

[N017RW]

Wind currents from fan or A/C?

Controller calibration (maybe a bit of left roll/aileron)?

Do you notice any body tilt when switching to A mode?

 

[dwallersv]

I'll scrutinize the controller, and also check the telemetry logs to see if there was any stick input. I doubt it, but need to check to be sure.

When switching between A and S mode, back and forth, I don't see any detectable shift o movement in the orientation of the aircraft; only that, as soon as the VPS is disabled, it starts to drift at a very constant (and very gradual) speed to the left. On the order of 0.5-1 ft/s (maybe even slower). I don't think it's propwash effects, but I didn't try spinning it 180 and then see if it continues to drift the same direction relative to the aircraft, or drifts the same absolute direction (which would imply a local turbulence effect from propwash).

I'll go test these things in the next hour and report back.

 

[F3honda4me]

This sounds pretty normal to be honest. By nature, in A mode there is nothing to hold the drone in place, so it's going to drift from things like prop wash, wind, etc.

 

[Erised]

When indoors the VPS sould hold position if the floor surface has a pattern for the cameras to lock on, if your floor is without pattern try putting something on the floor with a pattern. You may also need a VPS calibration using the assistant software. Make sure to follow the directions to the T. Also if the floor pattern is small and repetitive it may have difficulty holding and like No17RW said the prop wash can make you drift but with a strong floor pattern my P4 holds well. Make sure your VPS is turned on as we now have the option.

 

[F3honda4me]

When indoors the VPS sould hold position if the floor surface has a pattern for the cameras to lock on, if your floor is without pattern try putting something on the floor with a pattern. You may also need a VPS calibration using the assistant software. Make sure to follow the directions to the T. Also if the floor pattern is small and repetitive it may have difficulty holding and like No17RW said the prop wash can make you drift but with a strong floor pattern my P4 holds well. Make sure your VPS is turned on as we now have the option.

His VPS is working fine, not sure why he would want to calibrate it? He is asking about his drone doesn't stay in place with VPS OFF. Which again is 100% normal.

 

[F3honda4me]

Does this indicate a need for IMU calibration? Or are the 6-axis IMU sensors not active in A mode as well?

This would seem a bad design decision... the IMU sensors are always available, just like the barometer, which we know is used in A mode for maintaining elevation when hovering.

To my knowledge, the IMU's are used to determine attitude, not drifting.

 

[Erised]

Your right F3 he did turn off VPS and floating around is normal, dwall the VPS is for indoors if you leave it off you may hit THE WALL.

His VPS is working fine, not sure why he would want to calibrate it? He is asking about his drone doesn't stay in place with VPS OFF. Which again is 100% normal.

 

[N017RW]

To my knowledge, the IMU's are used to determine attitude, not drifting.

They sense inertia in 3 axis, drift would be sensed but just not acted upon.

 

[Erised]

The barometer determines altitude.

 

[dwallersv]

Sorry if I was too terse in explaining the "problem".

Yes, I expect it to drift around and need control inputs to stay put when in A mode. That's not what I observe as the "problem".

When in A-mode, indoors, quite a ways from any other objects (at least 20'), it consistently drifts left, at a very consistent, constant rate. Not fast, but always in the same direction.

It shouldn't, so long as everything (props, motor power, etc.) are symmetrical. Of course, in the real-world they aren't, and that's what all these sensors and the flight controller are for. I'm an RC helicopter hobbyist/pilot, so have lots of experience with trying to hover and keep in place.

The bird does exactly what I expect it to do in P and S modes, when the flight controller is doing all the work to stay on station. However, I thought IMU played a much bigger role in station-keeping, where it seems (by implication) that the accelerometers are being used only for orientation, not translation detection. The VPS apparently is the the only part of the system involved in positioning (in addition to GPS).

The drift only bothers me in that now I wish I had a trim dial to compensate for it, so I don't have to give constant tiny corrections on the stick.

In the end, it's all pretty academic, because I won't be flying under these sorts of conditions much, if at all (i.e. indoors, dead calm, A-mode).

 

[rocketflyer]

So have you calibrated the controller?

 

[dwallersv]

So have you calibrated the controller?

Controller? Not sure what you mean, could you please elaborate?

I'm only aware of calibration procedures for the compass, IMUs, and the VPS/OA.

 

[rocketflyer]

If the controller sticks are out there is an option in the go app to calibrate them. Hopefully this may solve your issue.

 

[Patrick]

They sense inertia in 3 axis, drift would be sensed but just not acted upon.

if the iMU is off kilter it will think its level when its really not , causing a drift. ever few months or so ill notice the phantom drifting more than usual and will be fixed by an IMU calibration.

 

[N017RW]

if the iMU is off kilter it will think its level when its really not , causing a drift. ever few months or so ill notice the phantom drifting more than usual and will be fixed by an IMU calibration.

If you mean mis-aligned sure, this is why I enquired about any visible tilt.
With the seeming absence of tilt it would point towards air currents.

 

[Patrick]

if the iMU is off kilter it will think its level when its really not , causing a drift.

If you mean mis-aligned sure, this is why I enquired about any visible tilt.
With the seeming absence of tilt it would point towards air currents.

if he's 20 ft off the ground i wouldn't think ground effects would be causing drift, might not notice a visible tilt if its only slightly drifting but should be interesting to see what the culprit was. i wonder if an out of balance prop would do this?

 

[F3honda4me]

The barometer determines altitude.

Except when you are close enough to the ground for VPS, then VPS maintains altitude. Within 30 feet of the ground roughly.

 

[F3honda4me]

if he's 20 ft off the ground i wouldn't think ground effects would be causing drift, might not notice a visible tilt if its only slightly drifting but should be interesting to see what the culprit was. i wonder if an out of balance prop would do this?

He said 20' away from objects, I wonder if that includes the ground?

 

[Pete Gianakos]

.5 to 1 ft/second seems like extreme drift for indoors. Are you testing when the fan in the house is not running?


Sent from my iPad using PhantomPilots