2.2 One-Tone Characterization Tests
2.2.4 One-Tone Characterization Setups
AM-AM and AM-PM characterizations are performed reading output signal com- ponents whose frequency is equal to the input excitation. Therefore, a usual ampli- tude controlled sinusoidal—or continuous-wave (CW)—generator connected to a vector network analyzer are sufficient for these tasks. The corresponding setup is depicted in Figure 2.4.
Since the network analyzer simultaneously measures DUT’s gain and phase, it is possible to characterize both AM-AM and AM-PM with a single amplitude power sweep. For that, relative gain is first converted into absolute output power, and then, that value, along with measured phase difference, is plotted against input drive level.
However, if a gain plot is directly used, it provides an immediate way for evaluating the DUT’s 1-dB compression point. Since this P1dB is nothing more than the output power at which the gain is already compressed 1 dB from its small- signal value, a gain plot inspection directly gives the corresponding input power level, which can be readily converted to output power, adding the actual measured gain. This procedure is exemplified in Figure 2.5.
Alternative, and less expensive, AM-AM characterization setups use a scalar network analyzer, or even a spectrum analyzer. Unfortunately, since neither of these pieces of equipment is able to measure phase, AM-PM characterization would no longer be possible.
Figure 2.4 AM-AM and AM-PM characterization setup based on a vector network analyzer.
THDcharacterization can only be performed with a spectrum analyzer, as the measured output includes frequency components that are different from the input excitation.
Figure 2.6 presents one such setup that can perform both AM-AM andTHD characterization. For that, the input generator is swept in amplitude and the output is measured at the fundamental frequency, for AM-AM, or at the harmonic compo- nents, forTHD.Obviously, thisTHDevaluation method relies on individual output power measurements at each harmonic, thus requiring a subsequent calculation according to (2.5).
Another simple AM-AM characterization setup relies on a power meter for measuring the DUT’s input and output powers. However, some care must be taken when using this setup because the power meter integrates all the power generated
Figure 2.5 DUT’s gain versus input drive level, showingP1dBevaluation.
Figure 2.6 AM-AM and/orTHDcharacterization setup using a spectrum analyzer.
by the DUT. Therefore, accurate AM-AM characterization requires that the DUT’s harmonic content be negligible in comparison to the fundamental output power, in the whole input power sweep span.
For the AM-PM characterization, the simple setup represented in Figure 2.7, which is based on a calibrated phase shifter and a spectrum analyzer, could also be used.
In this case, the output of the DUT—signal plus distortion—will be added to a sample of the input signal shifted by ␣ degrees. The objective of the bridge network is to cancel the output fundamental signal. The ␣ degrees introduced by the phase shifter, for perfect bridge adjustment, correspond to o1 = ␣ + (2k+1), wherekis any integer number ando1is the DUT’s output phase. If this procedure is repeated for an input power sweep, the resulting ⌬o1 versus Pinfunction constitutes the sought AM-PM characterization. The main problems associated with this setup are due to the phase shifter’s finite resolution, and to
IMDCharacterizationTechniques Figure 2.7 AM-PM characterization setup using a calibrated phase shifter and a spectrum analyzer.
the possible variable phase shift introduced by the attenuator present in the DUT’s branch. Furthermore, since the auxiliary branch is supposed to provide a signal that is an exact replica of the input, it must be guaranteed that its phase shifter does not generate any distortion components.
Other one-tone characterization setups, relying on dedicated or special labora- tory equipment, are possible. From these, the use of the microwave transition analyzer deserves to be mentioned. This modern piece of equipment not only combines the vector network analyzer operation with a spectrum analyzer, as some of its software options directly allow AM-AM, AM-PM, and THD automated measurements. Indeed, its two-port high-speed sampling oscilloscope, with built- in Fourier transform software, turns it into a revolutionary spectrum analyzer with phase measurement capabilities.
A final remark on these setups should assert that excessive signal generator phase noise or long-term frequency instability, as well as reduced signal analyzer dynamic range, can create severe impairments on the quality of the results. These difficulties, which are shared by almost all distortion measurement methods, and will be discussed later in greater detail, are especially notorious when large signal to distortion components ratios are involved.