Academia.eduAcademia.edu

Figure 11 – uploaded by Muhammad Akmal Chaudhary

See full PDFdownloadDownload figure
Fig. 9. The measured dynamic RF and IF load-lines for baseband reflection coefficients (Ty=1180°) terminating the baseband components with short circuit impedance resulted in static IF voltage, consequently, no variations in ACPR, and ACPRy were observed.

Figure 9 The measured dynamic RF and IF load-lines for baseband reflection coefficients (Ty=1180°) terminating the baseband components with short circuit impedance resulted in static IF voltage, consequently, no variations in ACPR, and ACPRy were observed.

Related Figures (10)

Fig. 1. Refined multi-tone measurement system with the capability of broadband active baseband (IF) and RF load-pull in [7], the injected signal is generated independently. This avoids the oscillation problem but now, to maintain a constant emulated load impedance (reflection coefficient) under changing drive conditions, results in a slow iterative process; a number of magnitude and phase adjustments, manual or automated, of the injected signal are needed to maintain the target reflection coefficient, the current MTA system is set up in this way and has proved prohibitive to certain measurements, most importantly it should be mentioned that such systems are unsuitable for multi-tone measurements as the number of signal sources required makes the measurement set-up far too expensive and physically impractical.
Fig. 1. Refined multi-tone measurement system with the capability of broadband active baseband (IF) and RF load-pull in [7], the injected signal is generated independently. This avoids the oscillation problem but now, to maintain a constant emulated load impedance (reflection coefficient) under changing drive conditions, results in a slow iterative process; a number of magnitude and phase adjustments, manual or automated, of the injected signal are needed to maintain the target reflection coefficient, the current MTA system is set up in this way and has proved prohibitive to certain measurements, most importantly it should be mentioned that such systems are unsuitable for multi-tone measurements as the number of signal sources required makes the measurement set-up far too expensive and physically impractical.
Fig. 2. Representation of sampled spectrum in the frequency domain What we end up with in the frequency domain is a number of sections each containing all relevant data related to that particular harmonic of the carrier frequency as shown in the diagram in figure 2, where a waveform with order of 4 is illustrated.
Fig. 2. Representation of sampled spectrum in the frequency domain What we end up with in the frequency domain is a number of sections each containing all relevant data related to that particular harmonic of the carrier frequency as shown in the diagram in figure 2, where a waveform with order of 4 is illustrated.
Fig. 3. Complete modulation cycle of on oscilloscope averaged 25-Tones signal at a carrier frequency of 2GHz with 500KHz tone separation. average the waveforms. The software takes the first waveform collection as a reference data set; all subsequent waveforms are then mathematically realigned to the reference data set to allow for averaging of the waveforms. When multi-tone waveforms are aligned correctly and synchronized with a multi-tone signal source, complex modulated waveforms can be accurately averaged, as depicted in Fig. 3.
Fig. 3. Complete modulation cycle of on oscilloscope averaged 25-Tones signal at a carrier frequency of 2GHz with 500KHz tone separation. average the waveforms. The software takes the first waveform collection as a reference data set; all subsequent waveforms are then mathematically realigned to the reference data set to allow for averaging of the waveforms. When multi-tone waveforms are aligned correctly and synchronized with a multi-tone signal source, complex modulated waveforms can be accurately averaged, as depicted in Fig. 3.
Fig. 3. Baseband load-pull characterization of 10W GaN HEMT under multi-tone stimulus The significant observation to note is that there are pronounced ACPR variations and asymmetry around the measurement’ grid. These measurements suggested that the optimum baseband load for minimum asymmetries is indeed located at point-A and then terminating the baseband components at an experimentally determined optimum point-A, a reduction of approximately 10dB in ACPR was observed in comparison with point-B and it highlights the strong dependence of ACPR on the baseband impedance termination
Fig. 3. Baseband load-pull characterization of 10W GaN HEMT under multi-tone stimulus The significant observation to note is that there are pronounced ACPR variations and asymmetry around the measurement’ grid. These measurements suggested that the optimum baseband load for minimum asymmetries is indeed located at point-A and then terminating the baseband components at an experimentally determined optimum point-A, a reduction of approximately 10dB in ACPR was observed in comparison with point-B and it highlights the strong dependence of ACPR on the baseband impedance termination
caused a ripple to appear on the DC drain voltage present at the device output plane, which effectively resulted in additional modulation of the RF signal. This, then, lead to the change in the ACPR levels and can be seen there exists a significant memory effects with more than 6dB ACPR asymmetry between upper and lower ACPR products. For the case when optimum impedance is maintained to the baseband components at modulation frequency of 1MHz there are no baseband electrical memory effects, so the spectral regrowth shows no asymmetry for optimum baseband impedance termination. The measured output RF voltage envelopes obtained for the two baseband impedance conditions, are shown in Fig. 5. Fig. 4. Measured output spectrum, when different impedances were provided to baseband components with nominal 50Q RF termination.
caused a ripple to appear on the DC drain voltage present at the device output plane, which effectively resulted in additional modulation of the RF signal. This, then, lead to the change in the ACPR levels and can be seen there exists a significant memory effects with more than 6dB ACPR asymmetry between upper and lower ACPR products. For the case when optimum impedance is maintained to the baseband components at modulation frequency of 1MHz there are no baseband electrical memory effects, so the spectral regrowth shows no asymmetry for optimum baseband impedance termination. The measured output RF voltage envelopes obtained for the two baseband impedance conditions, are shown in Fig. 5. Fig. 4. Measured output spectrum, when different impedances were provided to baseband components with nominal 50Q RF termination.
Fig. 5. Measured RF output voltage envelopes for different baseband impedance environment for 2MHz modulation frequency.
Fig. 5. Measured RF output voltage envelopes for different baseband impedance environment for 2MHz modulation frequency.
Fig. 6. Measured IF1, IF2, IF3 and IF4 for modulation frequency ranging from 0.5 MHz to 7 MHz. greater magnitude. Whilst terminating the baseband components at point-A, which has been demonstrated to be the optimum load termination, an undistorted output envelope is produced. The broadband baseband short circuit (je =12180°) impedance termination was utilized as shown in Fig. 6, for modulation requency ranging from 0.5 MHz to 7 MHz, in order to achieve the symmetrical ACPR response. The magnitude of the IF reflection coefficient could not be brought precisely to short for frequency components above 10 MHz due to bandwidth limitation of baseband load- pull PA. For example, the third harmonic of 6 MHz at 18 MHz could not be adequately controlled.
Fig. 6. Measured IF1, IF2, IF3 and IF4 for modulation frequency ranging from 0.5 MHz to 7 MHz. greater magnitude. Whilst terminating the baseband components at point-A, which has been demonstrated to be the optimum load termination, an undistorted output envelope is produced. The broadband baseband short circuit (je =12180°) impedance termination was utilized as shown in Fig. 6, for modulation requency ranging from 0.5 MHz to 7 MHz, in order to achieve the symmetrical ACPR response. The magnitude of the IF reflection coefficient could not be brought precisely to short for frequency components above 10 MHz due to bandwidth limitation of baseband load- pull PA. For example, the third harmonic of 6 MHz at 18 MHz could not be adequately controlled.
Fig. 7 Measured fundamentals and ACPR power for different modulation frequencies. below 1 MHz are not related to variations in baseband impedance and attributable to other sources of memory, e.g. trapping and thermal effects.
Fig. 7 Measured fundamentals and ACPR power for different modulation frequencies. below 1 MHz are not related to variations in baseband impedance and attributable to other sources of memory, e.g. trapping and thermal effects.
Fig. 8. Measured baseband loads for a class-J emulated RF impedance environment. under complex tones (i.e. more than two and three tones), however, every effort was made to minimize the spread to demonstrate the system capability in emulating the impedances when the device was subjected to complex multi-tone excitation. Plotting the measured (RF, IF & DC) time-domain voltage vs. time-domain current yields a complete dynamic load-line which is used to describe the trajectory of the modulated envelope relative to the device’s DC characteristic. This intuitive view shows the actual voltage and current behaviour at the calibrated reference plane, and importantly, the nature of any interaction between the envelope and device boundary conditions.
Fig. 8. Measured baseband loads for a class-J emulated RF impedance environment. under complex tones (i.e. more than two and three tones), however, every effort was made to minimize the spread to demonstrate the system capability in emulating the impedances when the device was subjected to complex multi-tone excitation. Plotting the measured (RF, IF & DC) time-domain voltage vs. time-domain current yields a complete dynamic load-line which is used to describe the trajectory of the modulated envelope relative to the device’s DC characteristic. This intuitive view shows the actual voltage and current behaviour at the calibrated reference plane, and importantly, the nature of any interaction between the envelope and device boundary conditions.
Fig. 10. Measured dynamic power sweep for 9-Tone stimulus for 2MHz modulation frequency. Fig. 10 shows the measured RF 9-tone power performance as a function of input drive level for 2MHz modulation frequency. The behavior of ACPR,; and ACPR,y is clearly identical when short circuit impedance was presented to all the significant baseband components under class-J mode of operation. It is important and interesting to note that measurement system is effective in presenting constant impedances to baseband components when RF components are terminated with reactive impedances. This highlights the effectiveness of measurement system in engineering comprehensively out-of-band impedance environment.
Fig. 10. Measured dynamic power sweep for 9-Tone stimulus for 2MHz modulation frequency. Fig. 10 shows the measured RF 9-tone power performance as a function of input drive level for 2MHz modulation frequency. The behavior of ACPR,; and ACPR,y is clearly identical when short circuit impedance was presented to all the significant baseband components under class-J mode of operation. It is important and interesting to note that measurement system is effective in presenting constant impedances to baseband components when RF components are terminated with reactive impedances. This highlights the effectiveness of measurement system in engineering comprehensively out-of-band impedance environment.
Related topics:
EngineeringElectronic EngineeringComputer Science
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved