F
Member
【Editorial Discussion】
Two established noise-figure methods on the same amplifier, at the same frequency, should not be treated as interchangeable until the reference planes and the instrument corrections are written down. When Y-factor and a cold-source (or cold-attenuator) result disagree, the usual reflex is to “believe the quieter number.” That hides the term that actually moved.
1) Confirm that both numbers were reduced to the same reference plane and the same input match. A Y-factor measurement that includes a noise-source mismatch, or a cold-source measurement that was not corrected for the source Gamma presented to the DUT, is not the same experiment as a 50 Ω data-sheet row. If the DUT is poorly matched, the two methods can diverge without either instrument being “broken.”
2) Check the hot and cold ENR (or the cold-source noise temperature) against the actual path, not only the cal sticker. Excess loss between the noise source and the DUT input, an unaccounted adapter, or a source that is no longer at the temperature assumed in the correction will tilt Y-factor. The cold-source method has its own analogous term: the kTB reference and any residual noise from the source module or the attenuator.
3) Watch receiver linearity and averaging, not only the NF readout. A noise-figure analyzer or a spectrum-based setup that is compressing on the hot state, or that is still averaging a changing DUT gain, will move one method more than the other. Record the DUT gain measured in the same setup as the NF; a gain discrepancy is often the first sign that the noise measurement is not at the condition you think.
4) Frequency, bandwidth, and image response must be in the note. A broadband noise source and a receiver with a poor image reject will not tell the same story as a narrow cold-source measurement around one tone. If the two methods were not taken in the same bandwidth and with the same image treatment, the disagreement is not yet a DUT problem.
Open question: the last time Y-factor and a cold-source NF disagreed on your bench, what actually closed the gap — mismatch / Gamma correction, ENR or path loss, receiver compression, or bandwidth and image? Please share the frequency and which term you think was dominant; no customer names or unverified numbers needed.
Two established noise-figure methods on the same amplifier, at the same frequency, should not be treated as interchangeable until the reference planes and the instrument corrections are written down. When Y-factor and a cold-source (or cold-attenuator) result disagree, the usual reflex is to “believe the quieter number.” That hides the term that actually moved.
1) Confirm that both numbers were reduced to the same reference plane and the same input match. A Y-factor measurement that includes a noise-source mismatch, or a cold-source measurement that was not corrected for the source Gamma presented to the DUT, is not the same experiment as a 50 Ω data-sheet row. If the DUT is poorly matched, the two methods can diverge without either instrument being “broken.”
2) Check the hot and cold ENR (or the cold-source noise temperature) against the actual path, not only the cal sticker. Excess loss between the noise source and the DUT input, an unaccounted adapter, or a source that is no longer at the temperature assumed in the correction will tilt Y-factor. The cold-source method has its own analogous term: the kTB reference and any residual noise from the source module or the attenuator.
3) Watch receiver linearity and averaging, not only the NF readout. A noise-figure analyzer or a spectrum-based setup that is compressing on the hot state, or that is still averaging a changing DUT gain, will move one method more than the other. Record the DUT gain measured in the same setup as the NF; a gain discrepancy is often the first sign that the noise measurement is not at the condition you think.
4) Frequency, bandwidth, and image response must be in the note. A broadband noise source and a receiver with a poor image reject will not tell the same story as a narrow cold-source measurement around one tone. If the two methods were not taken in the same bandwidth and with the same image treatment, the disagreement is not yet a DUT problem.
Open question: the last time Y-factor and a cold-source NF disagreed on your bench, what actually closed the gap — mismatch / Gamma correction, ENR or path loss, receiver compression, or bandwidth and image? Please share the frequency and which term you think was dominant; no customer names or unverified numbers needed.
