Can a worm gear keep up with the response time needed for drone stabilization?

[Jarfi]

Hello, I am an amateur designing a drone.

The mechanism which stabilizes it, has a worm gear connected to a gear which moves the motor, i.e in the system we have

1: Servo(with built in gears) 2: worm gear which servo moves 3: gear which worm gear moves 4: another gear which gear number 3 moves.

Gear number 4 changes the location of the Brushless motor to stabilize the drone in case it starts leaning on either site, or to make it turn etc.

Now my question is, 3 gears connected to a servo which reads feedback from an IMU sensor too much for the response time?

When the drone tilts, the servo needs to jerk/move the motor to ajust the drone, quickly, but I am worried about the Worm gear being slow and inefficient in doing this fast enough.

 

[berkeman]

Hello, I am an amateur designing a drone.

The mechanism which stabilizes it, has a worm gear connected to a gear which moves the motor, i.e in the system we have

1: Servo(with built in gears) 2: worm gear which servo moves 3: gear which worm gear moves 4: another gear which gear number 3 moves.

Gear number 4 changes the location of the Brushless motor to stabilize the drone in case it starts leaning on either site, or to make it turn etc.

Now my question is, 3 gears connected to a servo which reads feedback from an IMU sensor too much for the response time?

When the drone tilts, the servo needs to jerk/move the motor to ajust the drone, quickly, but I am worried about the Worm gear being slow and inefficient in doing this fast enough.

That does sound like a pretty slow way to try to stabilize a vertical lift platform. Do you know of any existing examples of drones or helicopters that use such a mechanism?

Is there a reason that you don't just modulate the drive power to 4 vertical lift pods reaching out from the main body of the drone? There's a reason that most vertical lift drones use that configuration...

 

[rcgldr]

Gear number 4 changes the location of the Brushless motor to stabilize the drone in case it starts leaning on either site, or to make it turn etc.

I'm not aware of any conventional radio control model that moves mass (the brushless motor) as a means to control the model.

Most drones simply alter the speed of the motors driving the propellers. Some aerobatic drones have propellers with variable pitch (positive and negative), so they also vary the pitch in addition to varying the speed. For aerobatic maneuvers, it's works out better if the motors run at near constant speed and only change the pitch of the propellers. This requires adjusting the voltage (and current) to compensate for increased drag that occurs with increased pitch. In the case of conventional helicopters, this is done via an adjustable curve that maps the "throttle" versus "pitch" outputs generated by the transmitter. For a drone, the mapping would be performed by circuitry in the drone.

Most servos used for radio control models have specification that state how quickly they can move and how much torque they can generate (sometimes both moving torque and holding (stalled) torque are specified) for given voltages, which used to be 4.8 volts (4 cell) or 6 volts (5 cell), but now some servos operate at the higher voltages available with some LiPoly battery packs.

 

[Jarfi]

Most servos used for radio control models have specification that state how quickly they can move and how much torque they can generate (sometimes both moving torque and holding (stalled) torque are specified) for given voltages, which used to be 4.8 volts (4 cell) or 6 volts (5 cell), but now some servos operate at the higher voltages available with some LiPoly battery packs.

Correct, I am using servos and I have lithium batteries. The servos have a reported "Operating speed: 0.14sec / 60 degrees (6.0V no load)". But that is fast enough, my concern was not how fast the servo spins(I might as well get a casual motor if speed is a problem), the problem was the Backlash and response time of the gears/mechanism. If that is too "laggy" the stabilization mechanism will be inefficient.

That does sound like a pretty slow way to try to stabilize a vertical lift platform. Do you know of any existing examples of drones or helicopters that use such a mechanism?

Is there a reason that you don't just modulate the drive power to 4 vertical lift pods reaching out from the main body of the drone? There's a reason that most vertical lift drones use that configuration...

No I have nothing else to compare it too, it's not a conventional quadcopter so drive power modulation alone is not enough to keep it stable.

 

[berkeman]

So you are just going to build it and see if it crashes on its first flight? That's now how good engineering usually works (hopefully).

 

[rcgldr]

problem was the Backlash and response time of the gears/mechanism. If that is too "laggy" the stabilization mechanism will be inefficient.

The final output is going through several stages of step down gears. Any slop in the intial worm gear drive would be divided down by the overall gear ratio after the the worm gear and end up very tiny. Usually there's very little slop in the final output stage of a servo, since the slop would end up allowing control surfaces to flap.

It's still not clear to me what you are using a servo for. A drone normally uses a separate speed controller for each rotor and adjusts the speeds of the rotors to control the drone. Variable pitch propellers use a servo to drive the collective pitch (positive, zero, or negative).

 

[Baluncore]

You will be unable to control quickly using a worm and reduction gear train. Apart from the speed limitation you will have hysteresis due to play in the gear train that will permit flutter and delay any control response.
If you can eliminate the gear backlash it might work as a slow trimming system. But you will need a faster active control system to maintain orientation.

 

[Jarfi]

You will be unable to control quickly using a worm and reduction gear train. Apart from the speed limitation you will have hysteresis due to play in the gear train that will permit flutter and delay any control response.
If you can eliminate the gear backlash it might work as a slow trimming system. But you will need a faster active control system to maintain orientation.

That is my fear. But let me emphasize on a few things.

The airframe will be fairly balanced, it will simply tip slowly to either side due to wind or structural imbalances I imagine. And I am thinking(hoping) that the motor will need to be ajusted only a little bit to achieve a large amount of turning force(due to the nature of the mechanism) which is why I was thinking that despite the slowness of a worm gear(gear reduction) that it would allow for precise balancing, but the flaw would be that it would not be able to ajust the craft fast enough in case of a hard cust of wind, say it were to swiftly be pushed 60 degrees, then the mechanism would be too slow and the craft would tip over and crash.

But I am guessing that maybe this: http://www.stepanlunin.com/Art1.html would prove much faster and actually work fairly great. Because when I think of it it does a similar thing(turns the force 90 degrees), the upside is that it's faster, downside being loss of precision and added backlash(more sloppy).

I read a bit about the worm gear and it says it is extremely precise, so now the only real flaw is how fast it can move the motor to balance the drone.

Thanks again for the help and patience everyone.

 

[Baluncore]

Air turbulence has high frequency components, especially near the ground or downwind of flow obstructions.
The control system must be very active to counter those disturbances.
There will be accurate gyro signals available every ten or so milliseconds.
The sooner the correction is made the less correction will be required.
Control surface accuracy is in no way as important as the speed of response.
The accurate gyro system will quickly attenuate all previous control loop errors.