F
Member
【Editorial Discussion】
A load-pull map is often treated as a property of the transistor or PA cell. In practice the contours can move when the bias tee, the baseband terminations, or the fixture low-frequency path changes, even if the RF tuner setting and the DC bias point look the same on paper. Before you rewrite a matching network, it is worth asking which plane actually changed.
1) Write down the low-frequency termination presented to the device, not only the RF Γ at the fundamental. Envelope and bias-line impedances at kHz-to-MHz offsets affect AM/AM, AM/PM, and the apparent optimum load. A “better” PAE contour that appears after swapping a tee may simply be a different baseband impedance, not a better RF match.
2) The fixture and the bias tee are part of the measurement. A different series resistance, a different bypass network, or a slightly different electrical length to the nearest decoupling capacitor will move both the RF and the envelope loops. If the cal or de-embedding stops at the RF connector and does not include how the bias network was mounted, two tees are two different fixtures.
3) Watch stability and low-frequency oscillations before trusting a shifted contour. A bias-loop resonance that is only marginally damped can look like a smooth efficiency improvement in one load region and like a sudden jump or a ragged contour in another. A spectrum check at the drain or collector, and at the RF output, belongs in the same folder as the load-pull file.
4) Keep the same reference plane, the same drive, and the same thermal condition when you compare tees. A contour set taken at a different available power, a different pulse duty, or after a different warm-up is not a fair tee comparison. Record Idq or the equivalent bias current, case or flange temperature, and whether the sweep was CW or pulsed, with every plot you intend to overlay.
Open question: when a load-pull contour last moved after a bias-tee or decoupling change, did you eventually attribute it to the DUT, to fixture de-embedding, or to the low-frequency termination? What did you record as the baseband impedance, and did a stability check change the conclusion?
A load-pull map is often treated as a property of the transistor or PA cell. In practice the contours can move when the bias tee, the baseband terminations, or the fixture low-frequency path changes, even if the RF tuner setting and the DC bias point look the same on paper. Before you rewrite a matching network, it is worth asking which plane actually changed.
1) Write down the low-frequency termination presented to the device, not only the RF Γ at the fundamental. Envelope and bias-line impedances at kHz-to-MHz offsets affect AM/AM, AM/PM, and the apparent optimum load. A “better” PAE contour that appears after swapping a tee may simply be a different baseband impedance, not a better RF match.
2) The fixture and the bias tee are part of the measurement. A different series resistance, a different bypass network, or a slightly different electrical length to the nearest decoupling capacitor will move both the RF and the envelope loops. If the cal or de-embedding stops at the RF connector and does not include how the bias network was mounted, two tees are two different fixtures.
3) Watch stability and low-frequency oscillations before trusting a shifted contour. A bias-loop resonance that is only marginally damped can look like a smooth efficiency improvement in one load region and like a sudden jump or a ragged contour in another. A spectrum check at the drain or collector, and at the RF output, belongs in the same folder as the load-pull file.
4) Keep the same reference plane, the same drive, and the same thermal condition when you compare tees. A contour set taken at a different available power, a different pulse duty, or after a different warm-up is not a fair tee comparison. Record Idq or the equivalent bias current, case or flange temperature, and whether the sweep was CW or pulsed, with every plot you intend to overlay.
Open question: when a load-pull contour last moved after a bias-tee or decoupling change, did you eventually attribute it to the DUT, to fixture de-embedding, or to the low-frequency termination? What did you record as the baseband impedance, and did a stability check change the conclusion?
