When switched to the amplifier state, the

insertion gain is +21 [pm]1.5 dB with high linearity (+38 dBm typical TOI).

Insertion gain ([G.sub.i]) is the typical gain measured with a scalar network analyzer (or power meter plus source) and is generally defined as the ratio of the power received with the DUT in place to that received with a thru line in place (that is, using the same source-receiver pair).

Consider the difference between using insertion gain and available gain by first looking at the gain ratio

Assuming the primary interest is in amplifier measurement, the analysis of the ratio of available gain to insertion gain as a function of DUT output match and the DUT's noise figure sensitivity to gain uncertainty suggests the conclusion that poorly matched, low noise figure amplifiers are most in need of proper gain definition and accurate gain measurement.

Insertion gain uncertainty (in the case of direct noise figure extraction) is often limited by the calibration accuracy of the attenuators within the noise receiver and may be on the order of 0.10 to 0.15 dB.[4] These attenuators are utilized between calibration and measurement steps to ensure that the power detector is kept in a linear range.

Note that since all currents [I.sub.XXX] are normalized to the input current [I.sub.INP], the ratio 20log [absolute value of [I.sub.OUT]/[I.sub.INP]] is not identical although it is proportional to the power

insertion gain. This result is due to the fact that the input impedance of the amplifier [Z.sub.INP] is not equal to [Z.sub.o].

The total

insertion gain of the modulator is 0 dB [+ or -] 0.5 dB at 5 MHz.

Figure 7 shows the new via hole locations and the insertion gain predicted by the field solver.

Figure 9 shows the new via locations and the insertion gain predicted by the field solver.

The final via hole configuration, which eliminates all resonances up to 30 GHz, is shown in Figure 11 with its predicted insertion gain. The field solver predicts greater than 40 dB isolation below 20 GHz for the passive structure from input to output.

The network analyzer also was used to measure the

insertion gain or loss ~[S.sub.21]~[sup.2].

Since both source and load impedances are fixed to a real value of 50 [Omega], transducer gain and loss and

insertion gain and loss calculations give the same result.