Application Information (Continued)
Choosing External Component Values
There are three key parameters to consider in the design:
Noise, signal bandwidth, and gain. Refer to
Figure 2
.
The best noise performance for a given transmit/receive
switch R
TR
is obtained by: choosing RSbetween 50Ω and
200Ω; selecting the matching terminationresistance R
in
; and
by reducing I
BIAS1
(by increasing R
BIAS1
) to increase r
e
which optimizes the Noise Figure (NF). For this circuit, with
R
TR
=
6Ω, the optimum NF isachieved at R
s
∼ 95Ω when R
in
is set to 50Ω and Rs∼ 145Ω when Rinis set to 200Ω.
The signal bandwidth is determined by the selection of L
1
,
L
2,RBIAS1
and R
BIAS2
which set the open loop gain roll-off of
the first stage. R
g
and C2form a desirable signal path filter
that introduces an additional highpass pole. The filter values
can be chosen to create a sharper high frequency roll off of
the closed loop gain. For R
g
=
1k, C2 ∼ 470pF the small sig-
nal (V
in
<
25mW), wide bandwidth performance can be ob-
served. By increasing C
2
(to>1500pF) and increasing output series resistance with a small resistor, the stability and
harmonic distortion performance can be improved for large
signals. This filter can also be designed as a multi-pole Butterworth filter but care must be taken to ensure stability with
the desired load over the operating temperature range.
The passband gain is customer selected by setting R
g
and
R
in
. Note that using R1to reduce or increase the gain allows
for minimal interaction with other parameters.
Capacitor C
C
and resistor RCare used for local compensa-
tion of the gm input stage with values of C
C
=
0.1µF and R
C
=
1k for the applications described below.
Calculating and Measuring the Noise
The circuit input referred noise is best calculated using a
SPICE model where the external components can be optimized for the transducer source impedance and
transmit/receive switch impedance. The SPICE model for
the CLC5509 is available on the NSC web site. Refer to the
figures for total noise performance over temperature and
supply at 3mA. Once the noise is modeled and circuit parameters chosen the evaluation board can be used to measure
actual noise performance.
To measure the CLC5509 input referred noise vs. other
noise sources, several key steps should be followed. The
bench setup is fairly simple using the evaluation board and a
spectrum analyzer. (If a noise figure meter is available that is
even easier yet.) The procedure requires calibrating out the
spectrum analyzer background noise, and other noise
sources from the CLC5509 noise. Since the thermal noise of
a resistor is well known, add a series resistor R
4
between the
signal source V
in
and the L1,R
BIAS1
bias network for these
noise measurements. Several R
4
resistor values are used as
’’reference’’ noise sources. The values chosen depend on
the R
s
of the system. For R
s
=
50Ω, resistor (R
4
) with values
of 0, 12.5, 25, 50Ω should be used. If R
s
=
200Ω, resistors
(R
4
) with values of 0, 25, 50, 100, 200Ω should be used.
Start by connecting the analyzer input to the evaluation
board output. Remove R
4
from the signal source and con-
nect R
4
to GND. Now take at least 10 measurements and av-
erage them for each R
4
reference value. Be sure to divide
the result by the analyzer and circuit gain to make the noise
input referred. Subtract the R
4
=
0Ω results from the data for
each value. Compare the result to the theoretical noise values. They should agree closely over the R
s
=
0toR
in
range.
This verifies the test method.The CLC5509 noise is the R
s
=
0Ω data point. A similar procedure can be used to remove
the T/R switch noise by varying the T/R bias current I
BIAS1
.
The total circuit noise performance can now be optimized for
R
in
as described above.
Evaluation Board
Evaluation boards are available for customer product evaluation for the 8-pin SOIC. Evaluation kits that contain an
evaluation board and CLC5509 samples can be obtained by
calling National Semiconductor’s Customer Service Cen-
ter. The evaluation kit number is CLC730101. The evaluation board utilizes surface mount components. The corner
frequencies are set to ∼ 0.9MHz to 12.5MHz with a passband gain set at 20dB. The highpass filter is set at R
g
=
1k,
C
2
=
470pF to view small signal (V
in
<
25mV) performance.
Increasing C
4
(to>1500pF) reduces the bandwidth and im-
proves distortion for large signals. R
9
, is a back match resistor that terminates the output and isolates cable capacitance, for minimum distortion, over the frequency band of
interest. An R
in
∼ 50.2Ω was chosen for R
s
=
50Ω (this
source resistor R
s
is open ) and R
TR
=
0, with I
BIAS1
set to
3mA. The expected input referred noise based on bench
measurements on similar boards is ∼ 0.6nV
oraNFof
2dB. The noise can be optimized with slight variations in R
2
,
R
g
and R
BIAS1
. If transmit/receive switches are added to the
evaluation board both the voltage drop and R
TR
should be
compensated for. The R
in
, gain and noise will be affected by
the addition of the T/R switch. The V
TR
drop can be removed, to a first order, by adding a second switch in series
with the feedback gain setting resistor R
1
to ground. This will
restore the input DC level to ∼ 0V. This T/R switch diode
should be biased with a resistor (∼ R
BIAS2
)toVCCand bypassed with a 0.1µF cap to maintain the same AC performance as the evaluation board without the switches.
CLC5509 Applications
The signal path for a typical ultrasound transceiver is shown
in
Figure 3
.
DS101304-32
FIGURE 2. Complete Circuit
DS101304-33
FIGURE 3.
CLC5509
www.national.com9