ISL55211_14 Datasheet, PDF (11/20 Pages) Intersil Corporation – Wideband, Low Noise, Low Distortion, Fixed Gain Differential Amplifier

ISL55211
+3.3V
115mW 35mA
10k
500 ^PD
50
1:1.4
Vi 1µF
RT
ADT2-
1T
or
ADT4-
1Wt
+
RG
200
VCM
0.1µF
-
RG
0.2pF
0.2pF
85
1µF ₃₅
VO
35
1µF 85
500
ISL55211
50
1:1
1µF ^Vm
ADT1-
1WT
FIGURE 29. TEST CIRCUIT 1
Working with a transformer coupled input as shown in Figure 29,
or with two DC blocking caps from a differential source, means
the output common mode voltage set by either the default
internal V_CM setting, or a voltage applied to the V_CM control pin,
will also appear as the input common mode voltage. This
provides a very easy way to control the ISL55211 I/O common
mode operating voltages for an AC-coupled signal path. The
internal common mode loop holds the output pins to V_CM and,
since there is no DC path for an I_CM current back towards the
input in Figure 29, that V_CM setting will also appear as the input
common mode voltage. It is useful, for this reason, to leave any
input transformer secondary centertap unconnected. The
internally set V_CM voltage is referenced from the negative supply
pin. With a single 3.3V supply, it is very close to 1.2V but will
change with total supply voltage across the device as shown in
Figure 27.
Most of the characterization curves starting with Figure 29 then
get different gains by changing the connections to the two pairs
of input R_G connections, as shown on the pin configuration
drawing on page 2. Two input turns ratios are intended for Test
Circuit 1; either a 1:1.4 turns ratio (ohms ratio of 2) or a 1:2 turns
ratio (ohm ratio of 4). The specific transformers shown in
Figure 29 are representative of broadband RF transformers but
alternate devices and manufacturers of these turns ratio devices
are certainly applicable. The output side of this test circuit
presents a differential 200Ω load while converting the
differential to single-ended through a resistive attenuator and a
1:1 transformer. This inserts approximately a 17dB insertion loss
that is removed to report the characteristic curves. For load tests
below the 200Ω shown in Figure 29, a simple added shunt
resistor is placed across the output pins. For loads > 200Ω, the
series and shunt load R's are adjusted to show that total load
(including the 50Ω measurement load reflected through the 1:1
output measurement port transformer) and provide an apparent
50Ω differential source to that transformer. This output side
transformer is for measurement purposes only and is not
necessary for final applications circuits. There are output
interface designs that do benefit from a transformer as part of
the signal path as shown in Figure 1. In that case, the 1:1:4
output side transformer becomes part of a filter design and
recovers the filter insertion loss from the amplifier output pins to
the ADC inputs.
Where just the amplifier is tested, a 4-port network analyzer is
used and the very simple test circuit of Figure 30 is
implemented. This is used to measure the differential S21 curves
vs gain of Figure 17 and as a simulation circuit for the differential
output impedance vs gain of Figure 18. Changing the gain is a
simple matter of adjusting the connections to the four input R_G
connections resistors, as shown in Table 1. This circuit depends
on the two AC-coupled source 50Ω of the 4 port network analyzer
and presents an AC-coupled differential 100Ω load to the
amplifier as the input impedance of the remaining two ports of
the network analyzer.
+3.3V
10k
RF
PD
50
1/2 of a 4-port
S-parameter
RG
+
VCM
RT
50
1/2 of a 4-port
S-parameter
-
50
50
RG
RF
ISL55211
FIGURE 30. TEST CIRCUIT 2 4-PORT S-PARAMETER
MEASUREMENTS
Using this measurement allows the small single bandwidth of
just the ISL55211 to be exposed. Many of the other
measurements are using I/O transformers that are limiting the
apparent bandwidth to a reduced level. Figure 17 shows the 3
normalized differential S21 curves for the possible internal gains
of 9dB, 14dB and 15dB. The small signal bandwidth is remaining
nearly constant at 1.4GHz due to the internal capacitive
feedback network.
The closed loop differential output impedance of Figure 18 is
simulated using Figure 30 in ADS. This shows a relatively low
output impedance (< 1Ω through 100MHz) constant with signal
gain setting. Typical FDA outputs show a closed loop output
impedance that increases with signal gain setting. The ISL55211
holds a more constant response due to internal design elements
unique to this device.
Common mode output measurements are made using the circuit
in Figure 31. Here, the outputs are summed together through two
100Ω resistors (still a 200Ω differential load) to a center point
where the average, or common mode, output voltage may be
sensed. This is coupled through a 1µF DC blocking capacitor and
measured using 50Ω test equipment. The common mode source
impedance for this circuit is the parallel combination of the
2-100Ω elements, or 50Ω. Figure 19 uses this circuit to measure
the small and large signal response from the V_CM control pin to
the output common mode. This pin includes an internal 50pF
capacitor on the default bias network (to filter supply noise when
there is no connection to this pin), which bandlimits the response
to approximately 30MHz. This is far lower than the actual
bandwidth of the common mode loop. Figure 20 uses this output
11
FN7868.0
June 21, 2011