N. Baćac*, V. Slukić*, M. Puškarić*, B. Štih*, E. Kamenar**, S. Zelenika**

* University of Rijeka, Faculty of Engineering, Rijeka, Croatia

** University of Rijeka, Faculty of Engineering and Centre for Micro and Nano Sciences and Technologies,

Rijeka, Croatia

Comparison of different DC motor positioning control

algorithms

Faculty of Engineering (riteh.uniri.hr) and Centre for Micro and Nano Sciences and Technologies (www.cmnzt.uniri.hr),

Precision Engineering Laboratory (precenglab.riteh.uniri.hr): People: Saša Zelenika, David Blažević, Ervin

Kamenar Main activities:

Ultra-high precision positioning systems and laser interferometric measurements

Energy harvesting/scavenging systems Stereomicroscope measurements Laser Doppler vibrometer measurements

About us…

Our projects… On-going:

GOLDFISH - Detection of Watercourse Contamination using Sensor Networks in Developing countries (EU FP7 project)

Ultra-high precision compliant devices for micro and nanotechnology applications (Funded by Ministry of Science Education and Sports of the Republic of Croatia)

Center for micro and nano sciences and technologies (CMNST)

Finished: Wireless Autonomous Tire Pressure Sensor

(WATPS) (The Business Innovation Agency of the Republic of Croatia - BICRO)

System for automatic pressure regulation with a self-regulating autonomous valve (SAV) (From Ip To Business - FIDES)

Introduction

Experimental Set-up

DC motor positioning control algorithms PID Cascade State-Space

Experimental results and discussion

Conclusions

Content

DC motor finds application in many of today’s mechatronics systems: precision positioning machines, robotic arms, pick-and-place machine for production of Printed

Circuit Boards (PCB), …

When DC motors are used, a feedback sensor is needed in order to establish velocity/position control.

Introduction

Experimental system employed in this work is composed of a DC actuator with an embedded planetary gearbox and an incremental rotational (quadrature) encoder

Actuator and feedback sensor (1/2)

1. Gearbox2. Encoder

Actuator and feedback sensor (2/2)

Element Type

Manufacturer Parameters

Actuator

2342 DC Faulhaber

Gearbox

Planetary Faulhaber i = 64:1

Feedback

sensor

Rotational incrementa

l (quadrature) encoder

Faulhaber2 channel, 12 CPR, X4 encoding

used (48 edges/ revolution = 7.5˚ resolution)

7 2

3

3

5.7 10

13.4 10

1.4 10

m

e

J kgm

NmK

AV

Krpm

1.9

12

75

1

a

n

an

R

U V

I mA

mNmb

rpm

National Instruments PXI-1050 chassis, PXI-8196 embedded controller and PXI-6221 Data Acquisition Card (DAQ) are used as a control system.

Control alghoritms are programmed in the LabVIEW programing environment.

LM675T based Power operational amplifier is used to drive the DC actuator.

Control system

Experimental Set-Up block scheme

64:1

Control algorithms – PID (1/2) 1

0)()(

nyny

DK

n

kke

IKne

PKnu

Predefined LabVIEW PID block is used to implement the PID alghoritm.

The tuning of the PID controller parameters is conducted in two steps; by using Ziegler-Nichols method (rough estimate of the gains) and by an experimental method (fine-tuning).

Control algorithms – PID (2/2)

PID parameters are verified by using custom developed Maltab model.

Control algorithms – Cascade (1/2)

Cascade control is composed of two loops: the velocity and the position loop.

Velocity (PI) loop is constituted by a proportional gain (KP), which scales the velocity error, and an integral time constant (TI), which defines the integration time.

A velocity loop itself cannot ensure that the actuator stops in a certain position, hence a positioning loop in cascade (series) with the PI velocity controller is employed.

Control algorithms – Cascade (2/2) The cascade controler parameters are tuned by

using a custom developed Matlab Simulink model.

The resulting parameters are implemented in the LabVIEW environment where additional online fine tuning is performed in order to match better real system response.

Control algorithms – State Space (1/2)

State-space (SS) control derives from the state-variable method of representing differential equations.

The basic principle of the synthesis of the SS controller is to achieve desired closed-loop dynamics with proper values of the L vector.

Gain vector L is calculated by using Ackerman’s formula.

Control algorithms – State Space (2/2)

SS Matlab model

DYNAMIC RESULTS FOR THE STATE-SPACE CONTROL METHOD

PARAMETERS FOR THE CASCADE CONTROL METHOD

PARAMETERS FOR THE PID CONTROL METHOD

Experimental results (1/2)

MATLAB ExperimentProportional gain 0.576 0.6Integral time [ms] 0.313 0.3Derivate time TD [ms] 0.021 0.02

MATLAB Experiment

Positioning loop

Proportional gain 13.56 13.905

Integral time [ms] ? 4

Velocity loop

Proportional gain 0.508 0.508

Integral time [ms] 5 5

MATLAB ExperimentRise time [ms] 430 390Percent overshot [%] 0.35 0.5Settling time [ms] 580 550Steady-state error [%] 0 0.04

Experimental results (2/2)COMPARISON OF DYNAMICS FOR DIFFERENT CONTROL METHODS (EXPERIMENTS)

PID Cascade S.S.Rise time [ms] 425 1120 390Percent overshot [%] 0 6.35 0.5Settling time [ms] 800 3500 550Steady-state error [%] 0 0.1 0.04

An overview of different DC motor control approaches is given in this work.

A Matlab model of the used actuator is established in order to simulate different control approaches.

Three different controllers are implemented in the LabVIEW environment and experiments are conducted.

Conclusions (1/2)

It is concluded that positioning control via the state-space controller has the fastest response and the lowest settling time.

Cascade control can be efficiently used, although the tuning of its parameters can often be cumbersome and computationally more intensive due to the presence of two PI blocks and the needed velocity calculation.

PID control results in negligible steady-state errors and acceptable rise and settling times.

Conclusions (2/2)

Thank you for your attention!

Questions