0

As shown underlined in attached photo Overshoot is mentioned as 9% I am confused in understanding it It is 9% of error or size of jump(1)? enter image description here

DSP_CS
  • 1,910
  • 2
  • 29
  • 65
  • 1
    See the comment by Hilmar under this question: https://dsp.stackexchange.com/q/1144/41790 . It links to wiki. – Ed V Apr 03 '20 at 16:45

1 Answers1

1

It is the error with respect to size of jump. It has nothing to do with error at discontinuity. As per the explanation given, suppose the perfect square is having values $0$ and $1$. Just after the discontinuity of a rise time, the perfect square would take value $1$, while the reconstructed wave would take the value $1.09$ at the peak of overshoot. Here the size of discontinuity is difference between perfect low value $0$ and perfect high value $1$ of the square wave. In general if it were $A$ and $B$, the overshoot percentage would have been $$ \frac{(Ovs_{peak} - (B-A))}{(B-A)} \times 100 $$

The error of $0.5$ in the error graph is not related to overshoot. For perfect square wave of $0$ and $1$, at discontinuity it is either $0$ or $1$, but the reconstructed wave will be midway $0.5$ ($|0.5−0|$ or $|0.5−1|$). So max error will be at discontinuity and its absolute value will be $0.5$.

jithin
  • 2,263
  • 2
  • 9
  • 16
  • This book text only means +ve overshoot 1.09 or also - overshoot such as 0.91? – DSP_CS Apr 04 '20 at 06:54
  • There is nothing like -ve overshoot. Overshoot itself is a damped oscillation with subsequent peaks having value less than 1.09. After the peak of 1.09, the next low (or trough) will not be 0.91. It will less than that because it is getting damped as time progresses. – jithin Apr 04 '20 at 07:42
  • Ok thanks. You are saying in your answer that overshoot is 9% of jump size(1) , but if you see in figure/graph, highest point/value is 0.5,so it appears that the mentioned 9% is 9% of 0.5 – DSP_CS Apr 04 '20 at 09:51
  • The error of 0.5 in the error graph is not related to overshoot. For perfect square wave of $0$ and $1$, at discontinuity it is either $0$ or $1$, but the reconstructed wave will be midway $0.5$ ($|0.5-0|$ or $|0.5-1|$). So max error will be at discontinuity and its absolute value will be $0.5$. In summary, max error is measured at discontinuity, while overshoot is measured with respect to how much it exceeds the step size. – jithin Apr 04 '20 at 09:58
  • If comfortable, please kindly try to adjust your last comment in your answer – DSP_CS Apr 04 '20 at 10:16
  • No problem, I have added. – jithin Apr 04 '20 at 11:24