Hardware Reference

In-Depth Information

Since the microactuator model is assumed stable, so is the error dynamics

(equation 3.177). Thus we have

k→∞
y
M
(k) − y
M
(k) = lim

lim

k→∞
H
M
x(k)=0.

(3.178)

P
M
= P
M
,

Hence, we have that if P
M
is stable and

k→∞
y
M
(k)−y
M
(k)=0.

lim

(3.179)

It should be noted that for MEMS based microactuator with relative po-

sition error signal [125], y
M
is available and thus can be used in the control

scheme directly instead of y
M
.Incasey
M
is not available, its estimation y
M

can be used as in Figure 3.96. In general, an open loop estimator does not

guarantee asymptotic stability of the error dynamics [8].

Next, we prove using an example that even with a reduced order PZT

actuated suspension model and parameter uncertainty, y
M
is still a close esti-

mate of y
M
, and hence the modified algorithm is effective in dealing with the

microactuator saturation.

Figure 3.96: Dual-stage control with microactuator saturation taken into con-

sideration.

3.7.5 Experimental Results

Althoughthedesignprincipleshowninthe previous section applies to various

types of dual-stage actuators, experiment using an actuated suspension is used

here to show the effectiveness of the proposed nonlinear control scheme. against

the original linear control scheme.

The nominal control loops, i.e., VCM path, microactuator path and overall

dual-stage open-loop are shown in Figure 3.97 with

P
V
(s)=
2.04
×
10
21

s +3.14 × 10
4

1

s
2
+ 301.6s +2.58 × 10
5

s
2
+ 2073s +4.3
×
10
8

s
2
+ 1508s +3.55 × 10
8