It should be noted that the single most important impact of implementing a control system digitally is the lagging effect associated with the hold. First, copy the following commands to an m-file and run it in the Matlab command window.

num=[1];
den=[1 10 20];
numDz=[0.0107 0.0055];
denDz=[1 -0.8106 0.1353];

t=0:0.2:2;
step (num,den,t)	%plots continuous output response
hold
[x]=dstep (numDz,denDz, 11);
plot (t,x,'ro')		%plots discrete output response
hold off

From this plot, we see that the discrete response exactly matches with the continuous response at each instant of sampling time. This is true because the input was a time-invarying step function. Had the input been a continuously time-varying function, it would not have matched exactly with the continuous response. This can be seen by changing the input of the same system from the example above from the step to the impulse. In the above m-file, change step to impulse and dstep to dimpulse. Re-running this m-file gives you the following plot.

From this plot, you see the discrete output does not match the continuous output. The discrete response is lagging the continuous response in certain range of time.

Even if the discrete response matches the continuous response,

the average signal uBar(t) lags the continuous signal u(t) by T/2 sec. However, if you decrease the sampling time (Ts, in sec/sample), this lagging effect becomes small and uBar(t) closely matches the continuous signal u(t).

User feedback