Unformatted text preview:

Massachusetts Institute of Technology Department of Mechanical Engineering 2.004 Dynamics and Control II Laboratory Session 4: Closed-Loop Performance of a Proportional Velocity Controller1 Laboratory Objectives: (i) Introduction to the 2.004 digital PID Controller. (ii) Detailed study of the effect of controller gain on transient and steady-state behavior (iii) Study of the ability of the closed-loop system to “follow” time varying commands. The 2.004 PID Controller In this session you will set up the closed-loop proportional velocity control, just as you did last week, but this time you will use the 2.004 computer-based controller. This is a digital, “discrete-time” controller, written in the LabVIEW language, that allows you to adjust the controller parameters by a set of virtual knobs on the front panel. This controller is a more general form than we have met so far, it is known as a PID (proportional + integral + derivative) controller, but in this lab we will only use the proportional function. In later labs we will examine the effects of the integral and derivative terms. The controller software runs in an infinite loop, reading the values of the command and feedback signals on two of the input channels, computing the value of the output to the 1December 14, 2006 1s e t amplifier, and writing that value as a voltage to a digital-analog converter (DAC). In the case of proportional control the computation is trivial: vout = Kp(vcom − vresp) where vcom is the command signal (set-point), and vresp is the system response (in this case the tachometer output). The loop is synchronized to an internal “clock”, and runs at a speed controlled by a knob on the front panel. The controller includes a chart-recorder that generates four traces: the input command (red), the system response (yellow), the error signal (green), and the voltage output (blue). As in the Chart Recorder software, the recorded data may be saved to a MATLAB or Excel file - with one difference: the data saved is the data displayed on the screen. You must use the pan and zoom tools to select the data you want before saving it. In both MATLAB and Excel the data is saved in arrays with four columns, where the column order is command, response, error, and output. To use the controller, you simply need to choose a sample/update rate - we suggest you use 50 samples/sec, and the Proportional Gain Kp. Make sure that the two other knobs KI and KD are set to zero. These invoke more advanced control algorithms that we will introduce in later labs. You start/stop the controller with the button at the lower left of the panel. The set-up is shown below: P o w e r A m p t a c h m o t o r - p o i n t C o m p u t e r A D 0 A D 1 D A 0 f i l t e r Notice that there is an additional element in the feedback path: a low-pass filter. The filter is a simple first-order RC circuit (see Appendix B) that attenuates the high frequency “noise” on the output of the tachometer. During Experiment 1 try your closed-loop system with and without the filter and observe it’s effect. Experiment #1 In the first experiment you will repeat the measurements that you made with the op-amp controller in Lab 3. The goal is to measure and plot 1) the closed-loop time-constant, 2) the fractional steady-state error, and 3) the peak motor current in response to a step input. Proceed as follows: (a) Connect the computer-based controller. (b) Start the controller, and select a sampling rate of 50 samples/sec. (c) Set the function generator to output a square wave with 1.0v amplitude, 0.5v offset and 0.2Hz frequency. 2(d) Install one magnet to act as a disturbance torque. (e) Set Kp = 2 on the controller. Start the system and record the waveforms. (f) Select a relevant section of the trace, save it, and then estimate (1) the time constant, (2) the fractional steady-state error, and (3) the peak motor current. Enter your data on the results sheet (attached). (g) Repeat the above procedure with Kp =3, 4, 5 (for Kp = 5, reduce the amplitude to 0.5v and the offset to 0.25v). Use the closed-loop transfer function (see the attached summary in the Appendix) to predict the variation of the time-constant and steady-state error as a function of Kp. Gen-erate a pair of plots comparing the theoretical values and your measured values. Also make a plot of the peak motor current as a function of controller gain. Experiment #2 Imagine that our rotational plant was part of a numerically controlled machine. The accuracy of the parts produced by the machine depends on how well the plant can follow a command signal. In other words, an important specification for many controlled systems is how they should respond to a time-varying input. Ideally, of course, the response should follow the command exactly - but that rarely happens. Our second experiment today is to investigate just how well the angular velocity of the plant follows a sinusoidal input command at several frequencies. Let the input command to the closed-loop system from the function generator be a sinusoid vin(t)= Ain sin ωt where A is the amplitude, and ω is the angular frequency. We will show in class soon that the response of a linear system to a sinusoidal input is another sinusoid with the same frequency ω, but differing in amplitude and with a shift in phase, that is vout(t)= Aout sin(ωt + φ) The ratio of the amplitudes Aout/Ain at any frequency ω is known as the gain at that frequency. Your task is to measure the gain at several frequencies, and to comment on the ability of this system to follow high and low frequency commands. With the closed-loop system set up as in Experiment # 1, proceed as follows: (a) Set the function generator to generate a sinusoidal output at a frequency of 0.2 Hz, with a peak-to-peak amplitude of 2 volts. (b) Set Kp = 2, start the controller, and record the response. (c) Either record the traces, or directly from the screen (adjust the y-scale), estimate the gain Aout/Ain at this frequency. Record your result on the results sheet. (d) Repeat for frequencies of 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0 and 2.0 Hz. 3Make a plot of gain as a function of frequency. Now we will let you in on a little secret. For a first-order plant such as this, at an angular √ frequency of ω =1/τ


View Full Document

MIT 2 004 - Study Notes

Documents in this Course
Load more
Download Study Notes
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Study Notes and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Study Notes 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?