P
Member
Getting 0.3dB discrepancy between Y-factor and cold-source NF measurements on a 2GHz LNA. Which one do you trust? My noise source ENR cal might be the culprit.
LNA noise figure measurement - Y-factor vs cold source
L
Admin
0.3 dB between Y-factor and cold-source at 2 GHz is in the zone where ENR / mismatch / receiver linearity usually dominate — not a mysterious LNA.
Trust order many labs use:
1. Cold-source (or cold attenuator) when the ENR table is old, the noise source VSWR is ugly, or you are near the DUT NF floor and Y-factor excess noise is large.
2. Y-factor when ENR is freshly calibrated, mismatch is corrected (or DUT Γ is low), and the analyzer’s NF uncertainty budget is smaller than your 0.3 dB gap.
3. Always check: noise-source ENR age and connector, DUT input match at the measurement frequency, second-stage correction / gain accuracy, and that neither method is compressing the receiver (too-hot Y-factor is a classic +0.2–0.5 dB liar).
Practical next step: re-seat the noise source, verify ENR vs the cal sticker at 2 GHz, and run both methods with a known good 3–6 dB pad on the DUT input — if the gap collapses, it was mismatch/ENR; if it stays, dig into cold-source reference temperature and analyzer second-stage terms. Which noise source model / last ENR cal date are you on?
— rfforum 编辑部 / Editorial Team
Trust order many labs use:
1. Cold-source (or cold attenuator) when the ENR table is old, the noise source VSWR is ugly, or you are near the DUT NF floor and Y-factor excess noise is large.
2. Y-factor when ENR is freshly calibrated, mismatch is corrected (or DUT Γ is low), and the analyzer’s NF uncertainty budget is smaller than your 0.3 dB gap.
3. Always check: noise-source ENR age and connector, DUT input match at the measurement frequency, second-stage correction / gain accuracy, and that neither method is compressing the receiver (too-hot Y-factor is a classic +0.2–0.5 dB liar).
Practical next step: re-seat the noise source, verify ENR vs the cal sticker at 2 GHz, and run both methods with a known good 3–6 dB pad on the DUT input — if the gap collapses, it was mismatch/ENR; if it stays, dig into cold-source reference temperature and analyzer second-stage terms. Which noise source model / last ENR cal date are you on?
— rfforum 编辑部 / Editorial Team
广告 Ad
Renesas Electronics微控制器、模拟与射频/时钟方案 · MCUs, analog, RF and timing访问官网 →免费展示 · 资料来自公开信息 · 如需撤下请联系 ads@rfforum.com
Renesas Electronics微控制器、模拟与射频/时钟方案 · MCUs, analog, RF and timing访问官网 →免费展示 · 资料来自公开信息 · 如需撤下请联系 ads@rfforum.com