Datasheet AD8350 (Analog Devices) - 9

AD8350
For the output matching network, if the output source resis- The same results could be found using a Smith Chart as shown tance of the AD8350 is greater than the terminating load in Figure 7. In this example, a shunt capacitor and a series inductor resistance, a step-down network should be employed as shown are used to match the 200 Ω source to a 50 Ω load. For a fre- on the output of Figure 3. For a step-down matching network, quency of 10 MHz, the same capacitor and inductor values the series and parallel reactances are calculated as: previously found using the resonant approach will transform the 200 Ω source to match the 50 Ω load. At frequencies exceeding R 100 MHz, the S parameters from Tables II and III should be S RLOAD R X LOAD = × where = × S XP RS (2) X used to account for the complex impedance relationships. P RS – RLOAD For a 10 MHz application with the 200 Ω output source resistance of the AD8350, RS = 200 Ω, and a 50 Ω load termination, RLOAD = 50 Ω, then XP = 115.5 Ω and XS = 86.6 Ω, which results in the following component values: CP = (2 π × 10 × 106 × 115.5)–1 = 138 pF and LS = 86.6 × (2 π × 10 × 106)–1 = 1.38 μH The same results can be obtained using the plots in Figure 5
LOAD SOURCE
and Figure 6. Figure 5 shows the normalized shunt reactance versus the normalized source resistance for a step-up matching
SHUNT C
network, RS < RLOAD. By inspection, the appropriate reactance
SERIES L
can be found for a given value of RS/RLOAD. The series reactance is then calculated using XS = RS RLOAD/XP. The same technique can be used to design the step-down matching network using Figure 6.
2 1.8
Figure 7. Smith Chart Representation of Step-Down Network
R AD SOURCE X LO S 1.6
After determining the matching network for the single-ended
/R R P X LOAD P
equivalent circuit, the matching elements need to be applied in a
1.4
differential manner. The series reactance needs to be split such
1.2
that the final network is balanced. In the previous examples, this
ANCE – X CT 1
simply translates to splitting the series inductor into two equal halves as shown in Figure 3.
0.8 Gain Adjustment 0.6
The effective gain of the AD8350 can be reduced using a num-
0.4
ber of techniques. Obviously a matched attenuator network will
NORMALIZED REA 0.2
reduce the effective gain, but this requires the addition of a
0
separate component which can be prohibitive in size and cost. The attenuator will also increase the effective noise figure resulting
0.01 0.05 0.09 0.13 0.17 0.21 0.25 0.29 0.33 0.37 0.41 0.45 0.49 0.53 0.57 0.61 0.65 0.69 0.73 0.77 NORMALIZED SOURCE RESISTANCE – R
in an SNR degradation. A simple voltage divider can be imple-
SOURCE/R LOAD
Figure 5. Normalized Step-Up Matching Components mented using the combination of the driving impedance of the previous stage and a shunt resistor across the inputs of the AD8350 as shown in Figure 8. This provides a compact solution but
3.2
suffers from an increased noise spectral density at the input
R
of the AD8350 due to the thermal noise contribution of the
SOURCE X
shunt resistor. The input impedance can be dynamically altered
AD S 3 LO R X LOAD P
through the use of feedback resistors as shown in Figure 9. This
/R P
will result in a similar attenuation of the input signal by virtue
2.8
of the voltage divider established from the driving source imped-
ANCE – X
ance and the reduced input impedance of the AD8350. Yet
CT 2.6
this technique does not significantly degrade the SNR with the unnecessary increase in thermal noise that arises from a truly
2.4
resistive attenuator network.
2.2 NORMALIZED REA 2 2 4 6 8 2.4 2.8 3.2 3.6 4.4 4.8 5.2 5.6 6.4 6.8 7.2 7.6 8.4 8.8 NORMALIZED SOURCE RESISTANCE – RSOURCE/R LOAD
Figure 6. Normalized Step-Down Matching Components REV. C –9–