Benefits of Direct Drive Motors

This technical note describes concepts of geared and direct drive motors. Inertia ratio and gear equations are provided. Geared motor challenges are discussed and the solution to those challenges, direct drive motors, are detailed. This paper also provides a basic example comparing the motor output power of a geared motor to a direct drive motor.

Introduction

Direct drive motors offer several benefits over geared motors for a variety of motion systems and applications like Delta bots, robotic arms, healthcare robotics, mobile robotics, and factory automation systems. While gears can help achieve load torque requirements for servo motors, direct drive motors offer a significant advantage to system integrators through reduced cost of ownership, faster settling times, and design simplicity.

Why Gears are Used – Torque/Speed Translation

Gears are used for torque and speed translation – typically to achieve higher torque from high speed, low torque servo motors. When paired with a gearbox, torque output increases in proportion to the gear ratio. However, output speed is reduced by the same ratio.

Gear ratios are determined based on the mechanical design of the gearbox. Equations for the gear ratio can vary. When selecting a gear, the minimum gear ratio needed for an application can be calculated based on the output torque requirement of the system and the available input torque from the servo motor.  Gear efficiency must also be factored into this equation.

Gear Ratio equation

Alternatively, the equation above can be used to determine the input torque that is required from the servo motor based on a selected gear ratio and output torque requirement. For example, if 50 Nm of torque is required to drive the load with a 50:1 gear ratio and 80% gear efficiency, the input torque to the gear from the motor must be 1.25 Nm.

When choosing the proper gear ratio, the maximum load speed must also be considered as the addition of a gear requires the motor to operate faster than the required output speed. Gear efficiency does not impact the speed translation.

Gear efficiency equation

Using the same gear ratio as in the example above, if 100 RPM is required at the output of the gearbox, then the input speed must be 5000 RPM.

Reduce Inertia Mismatch for Over-Constrained Systems

When the bearings in a gearbox are not sufficiently sized for an off-axis load, a second set of load bearings must be added. This creates an over-constrained system that requires a flexible coupling to account for any eccentricity between the load shaft and the gear shaft.

The use of a flexible coupling introduces resonance to the plant transfer function of the system. When load and motor inertias are significantly mismatched, the frequency of the resonance decreases. Gears can be used to improve unbalanced load to motor inertias and increase the resonant frequency such that it does not interfere with the control loop. To better understand this phenomenon, an understanding of the inertia ratio is critical.

Inertia ratio

If the inertia ratio is high, then the control system will be more susceptible to instability and overshoot. The ideal inertia ratio is typically between 10:1 and 3:1 with the lower inertia ratios being used in precision applications.

Figure 1 below shows that a higher inertia ratio shifts the resonant frequency to the left in the plant transfer function. To avoid instability caused by the lower frequency resonance, the bandwidth should be reduced – also reducing the performance of the control system.

Inertia Ratio
Figure 1

When a gear is paired with a motor to drive the load, the load inertia is reduced by the square of the gear ratio (GR). The gear inertia (JG) is added to the load inertia but is relatively small compared to the gear reduction. The load inertia is replaced by the reflected inertia (JR) in the inertia ratio equation.

The following simple, unitless example compares the inertia ratio with and without a gear reducer.

1. The load is driven by a direct drive motor. The load inertia is 100 and the motor inertia is 1.

inertia ratio with and without gear reducer

This high inertia ratio will require a balance between performance and stability from a control perspective. To avoid the interference of resonances and maintain stability, the bandwidth of the control loop must be reduced.

2. The load is coupled to a gearbox which is driven by a servo motor. The gear ratio is 5:1, and the gear inertia is 1.

gear ratio 5 to 1

The inertia ratio has been significantly reduced by the addition of a gearbox from 100:1 to 5:1. In this case, the bandwidth and performance of the control loop are less restricted by the resonances.

The gear effectively reduces the reflected inertia, reducing the inertia ratio and increasing the stability of over-constrained systems. For more information on resonances introduced to the control system by mechanical components, read our technical note on Frameless Motor Advantages.

Achieve Higher Resolution

The distance the gearbox output moves relative to the input is dependent on the gear ratio. With a 20:1 gear ratio, for example, the input shaft must turn 20 revolutions for the output to turn one revolution. This effectively increases the resolution of the input encoder by a factor of 20. Due to the gearbox backlash, however, it is not possible to realize the increased resolution for small moves within the dead-band.

Challenges When Integrating Gears

While gears are effective at translating torque for servo motors, there are several challenges that need to be considered before integrating a gearbox into a motion system.

Backlash

All gearboxes contain some amount of backlash. Backlash is a mechanical gap that occurs in gears when changing directions due to the mechanical play in the gear teeth. When the input shaft changes direction, the output teeth become physically disconnected until the input moves the distance of the mechanical play. At that point, the output teeth come back in contact with the input and follow the movement of the input, as shown in figure 2 below.

Backlash
Figure 2

Backlash introduces a time delay and position dead-band between the input and output of the gearbox and can cause instability in servo loops. Backlash is specifically undesirable for applications that require direction changes like pick-and-place systems used in packaging and assembly and web-line systems that utilize phase-advance and phase-retard routines for high synchronicity.

While there are techniques to minimize backlash, such as bearing preload, this can increase the friction in the gearbox and reduce the life of the system. Component integrators can also utilize dual encoders on the input and output of the gearbox to close the velocity loop on the input and close the position loop on the output, shown in figure 3. While this improves accuracy and repeatability when changing the direction of rotation, it does not fix effects of backlash. It is important to note that backlash increases over time as gear teeth wear due to direct contact during rotation.

Velocity loop
Figure 3

Settling Time

Gearboxes are not torsionally stiff components, meaning they have some amount of compliance. Flexible couplings, which are typically used with geared servo motors, are also compliant. These mechanical components add resonances to the system which interfere with the servo drive control loops, adding vibration or instability depending on where the position loop is closed.

The time it takes the load to settle after meeting the commanded set point is called the settling time. During this time, the load oscillates until it comes within tolerance of the set-point, figure 4.

Set Point
Figure 4

If only a single encoder is mounted to the input of the gearbox, any backlash and resonances at the load could be undetected at the input where the position loop is closed. Therefore, the settling time as detected at the input would be faster than the actual settling time of the load.

If two encoders are used, one at the input and one at the output of the gear, the position loop can be closed at the load to track the true position within the dead-band. However, a resonance introduced by a flexible coupling could increase the settling time of the load. Longer settling times lead to reduced throughput which is detrimental to applications such as delta bots, indexing tables, and semiconductor wafer handling systems.

Inefficiency

Gearboxes are not perfectly efficient, meaning that there is power lost as torque is increased and transmitted to the load. The efficiency is dependent on mechanical components such as the gear teeth, bearings, seals, and grease which introduce friction to the gearbox. To overcome this friction and achieve the required load torque, the motor output torque must increase. This also means that the input power to the motor must increase, placing a bigger burden on the servo drive. A geared motor is not an ideal solution for mobile robotics, exoskeletons, and battery-powered applications that have a reduced power budget.

Size

The addition of a gear increases the axial length and mass of the assembly, posing a challenge for robotic arms and other robotic systems that operate in constrained spaces. This is particularly challenging for lower payload mobile robotics systems and unmanned aerial vehicle (UAV) applications that have restrictive weight allowances.

Reduced Lifetime

When operating continuously or at high duty cycles, temperature within the joint increases. High temperatures reduce the viscosity of gear lubrication, increasing friction in the gear and reducing its lifetime. The gear lubrication also poses a risk of contamination which requires additional mechanical components like labyrinth seals or IP rated housings increasing the size, mass, and cost of the geared motor assembly. Additionally, the friction from the mating gear teeth causes wear and increases backlash over time.

Cost

In addition to these technical concerns, the addition of a gear to the mechanical system increases the overall cost. Not only does system cost increase due to the additional component or components (if flexible couplings for constrained loads are also used), but due to increased service costs from component degradation/failure.

What is a direct drive motor?

A direct drive motor is a motor that directly drives the load. It consists of a stator, the fixed component, and a rotor, the rotating component that is directly coupled to the load typically through mounting screws or adhesive bond. Direct drive motors typically have large through-holes and high torque density when compared to servo motors of similar form factor.

Why use a direct drive motor?

Faster Settling Times
When using a direct drive motor, the load is directly mounted to the rotor and bypasses other transmission components, such as gearboxes, belts, pulleys, and flexible couplings for maximal stiffness. This enables significantly faster settling times than geared motors. When the load settles faster, the motion profile is completed faster, offering improved throughput compared to geared actuators.

Reduced Size
Direct drive motors generally have higher torque density than servo motors of a similar size, so a gear is not always needed to achieve the load torque requirements. This means that size of the actuator is reduced to only the motor and load. In certain cases where the load is constrained, bearings and flexible couplings may be required. But the removal of the gearbox alone significantly reduces the size and weight of the actuator.

Lower Cost of Ownership
While they typically have higher up-front costs, direct drive motors offer longer mean time before failure (MTBF) to the simplicity of the design. The reduction of components reduces the failure points in the system and the cost of ownership.

Removes Burden of Inertia Mismatch
Since the load is directly mounted to the motor, removing the need for flexible couplings, the inertia ratio can be higher without sacrificing control loop performance. However, load inertia does impact the achievable acceleration of the motor and should be considered during motor selection. A high load inertia reduces the angular acceleration while the torque is maintained.

high load inertia acceleration

The blue dots in figure 5 summarize the advantages of both direct drive and geared servo motors.

advantages of geared and non geared motors
Figure 5

Motor Output Power Example

Consider the simple, unitless example below comparing motor output power from direct drive application to a geared application.

Load Inertia (JL) = 4
Load Torque (TL)  = 2
Load Speed (WL)  = 1
Gear Ratio (GR)  = 5:1
Gear Efficiency (Geff)  = 80%

For the direct drive application:

Motor Output Power Equation

Since the motor is coupled to the load, the motor speed is equivalent to the load speed and the motor torque is equivalent to the load torque.

Torque Equivalent to torque load

For the geared application:

Motor output power geared

where the motor torque is now reduced by the gear ratio and motor speed is increased. The gear efficiency must also be considered.

Motor output power geared efficiency equation Motor output power geared power equation

This example shows that the motor output power is higher for the geared application than for the direct drive application. But this is not the whole story. What also matters in discussions about motor efficiency is the motor input power requirement from the servo drive which is dependent on the power losses in the motor.

Motor output power loss equation

There are two key contributors to power loss in a motor – copper and core. Copper losses are caused by heating in the motor as current flows through the copper windings. Copper loss is equivalent to the square of current multiplied by the line-to-line resistance.

Power loss copper equation

Core losses are caused by eddy currents in the stator stack and hysteresis of magnetic flux lines from the rapidly changing magnetic field. These losses are speed dependent – the higher the electrical frequency, the larger the contribution of core losses to the total power losses in the motor. Core losses are difficult to model but are typically equal to copper losses at 80% of the no-load speed and 80% of the continuously rated torque.

When using a gear, the torque requirement from the motor is reduced which generally means that the motor can operate at a lower current. Using our copper loss equation above, we see that a lower current reduces the copper losses in the motor windings. However, the motor must now operate at a faster speed, which we now know will increase the core losses in the motor.
If a direct drive motor is used, the motor must be able to produce enough torque to drive the load. This means that it will need more current than the geared motor, and copper losses will increase. However, now the motor will run at a slower speed, resulting in lower core losses than the motor that is paired with a gearbox.

The conversation around efficiency is not straightforward and heavily dependent on the motion profiles and gear efficiency. Generally, geared solutions consume more power strictly based on the gear efficiency. However, each application will require a different motion profile and torque-speed requirement. Output power, power loss, and input power should be modeled to determine the true power savings of a direct drive motor over a geared motor.

While gearboxes may be needed to achieve higher torque from traditional servo motors, direct drive motors offer zero backlash and faster settling times which can significantly improve throughput. While upfront costs are higher, direct drive motors require fewer mechanical components and experience lower friction losses than geared servo motors. This saves the integrator over time on maintenance cost and system down-time.