2nd Order Transient Response

Start with a second order circuit (typically an inductor and a capacitor, or two of each).

We get a second order differential equation.

d 2 vo (t) dt 2

R dvo (t) L dt

1 vo (t) LC

1 vi (t) LC

First order method from last lab is not helpful, but methods from class are!

Assume homogeneous response is of the form Aest.

Find characteristic values (s); for second order circuit we expect two. For constant input (this lab) assume particular solution is a constant.

Homogeneous Response
Rewrite differential equation (with no input):
d 2 vo (t) dt 2 d 2 vo (t) dt 2 R dvo (t) 1 vo (t) 0 L dt LC dv (t) 2 2 0 o 0 0 v o (t) dt

Assume homogeneous response is of the form Aest. 2 s 2 2 0s 0 0

s1,2

2 2
2

2 0

2 0

Three cases of interest with passive circuits.

1: 1: 0

two distinct real roots, overdamped two distinct real roots, critically damped 1: two distinct complex roots, underdamped

We will consider only first (overdamped) and third (underdamped) cases.

Overdamped Case (>1)

This is the easy case (for calculations, but not in lab) Find roots.

s1,2

2 0 0

Write homogeneous response (constants are unknown).

vo,h (t) A1es1t A1et/ 1

1
1,2 0 0

A 2 es 2 t A2et/ 2
2

Underdamped Case (0<1)

This is the easy case (in lab, but not for calculations) Find roots.

s1,2

2 0 0 0

1
2

j j
d

(book 's notation)

Write homogeneous response. There are several forms, use whichever is convenient (constants are unknown).

vo,h (t) A1es1t e Ce

0t

A 2 es 2 t B1 cos
0t

(A1 , A 2 complex)
0

1
2

B2 sin

cos

Particular, Complete Response

For unit step input, particular response = constant.

vo,p (t) K
Complete response (overdamped):

vo (t) vo,p (t) K

v o,h (t) A 2 es 2 t

A1es1t

Complete response (underdamped):

vo (t) vo,p (t) vo,h (t) K Ce

0t

cos

Note: K is determined by input and differential equation. Constants of homogeneous response are determined by initial conditions.

Initial Conditions
For this circuit, and step input:

d 2 vo (t) dt 2
vo (0 ) ...?

R dvo (t) L dt

1 vo (t) LC

1 vi (t) LC

vo (0 ) 0

dv0 (t) dt t

...?
0

dv0 (t) dt t

0
0

Effect of
Increasing denotes increasing damping ( determines nature of response)
Step response, 2 =0 1.8 1.6 1.4 1.2 =0.2 =0.7 =1 =2 varies

Amplitude

1 0.8 0.6 0.4 0.2 0

10

15 Time (sec)

20

25

30

Increasing 0 denotes increasing speed (0 determines speed of response)

Step response, 2 1.8 1.6 1.4 1.2
0 0 0 n

Effect of 0
varies

=0.5 =1 =2

Amplitude

1 0.8 0.6 0.4 0.2 0

10

15 Time (sec)

20

25

30

In-Lab Measurements (d)

(underdamped case only)
Measuring 2 1.8 1.6 1.4 1.2
d

= 2 /T; T is one period

Amplitude

1 0.8 0.6 0.4 0.2 0

Why use zero crossings instead of peaks?

10 Time (sec)

12

14

16

18

20

In-Lab Measurements ()
(underdamped case only)
y1 y2 a b e a b e
t1 t2

y1 a y2 a

b e b e e

t1 t2

t1 t 2

log

y1 a y2 a

t1 t 2

This is called the logarithmic decrement.

log

y1 a y2 a t 2 t1

Note: we could also do a curve fit for more accuracy (and use all the data).

Switching parameters
(underdamped case only)
0 d 0

1
2 0

2 2 d

2 2 0

1
2 2 2 d

2 d

2 0 2 0

2 0

If you do a curve fit, you can solve directly for and n.

The Overdamped Case

For overdamped (and critically damped) case, there is no easy way to find the two time constants from the experimental data unless you do a curve fit.
Second order, overdamped 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0

y(t)

A1e

t/ 1

A 2e

t/ 2

A curve fit is made difficult because the unknown coefficients (A1, A2, 1, 2) are hard to make initial guesses for by looking at the graph but you can use theoretical values as your initial guesses.

Amplitude

10 12 Time (sec)

14

16

18

20

If you do a curve fit, you can solve directly for and n.