L
Admin
[Technical article / FAQ, rfforum Editorial Team]
Y-factor with a noise source is still the workhorse NF method on many RF benches. Cold-source / gain-method setups show up when you need lower uncertainty or when a noise diode is inconvenient. Both fail in characteristic ways when the assumptions are not written down.
1. Y-factor traps
- ENR table uncertainty and connector / adapter loss between the source and the DUT reference plane dominate many "0.1 dB" arguments
- Receiver linearity: if the "hot" state compresses the analyser or NF meter, Y is wrong in a quiet, systematic way
- Mismatch: ENR assumes a stated source Γ; a poorly matched DUT input plus an uncorrected source mismatch converts into NF error, not just a return-loss footnote
2. Cold-source / gain-method traps
- You need a trustworthy gain measurement at the same reference planes; a bad S21 or a skipped de-embed becomes an NF fiction
- kTB reference and temperature of the cold termination must be real, not "room temp ≈ 290 K" handwaving in a hot rack
- Low-gain DUTs make the instrument NF correction critical — know when the meter is measuring itself
3. Habit
Write on every NF sheet: method (Y vs cold-source), ENR / cold temperature, adapter stack, reference plane, and whether mismatch correction was applied. Two labs disagreeing by a few tenths of a dB are often disagreeing about ENR + adapters, not about the LNA.
Discussion: when your last NF correlation failed between stations, was the root cause ENR/adapters, compression on hot, or a gain/de-embed error on the cold-source path — and do you require mismatch notes on the traveler?
— rfforum 编辑部 / Editorial Team
Y-factor with a noise source is still the workhorse NF method on many RF benches. Cold-source / gain-method setups show up when you need lower uncertainty or when a noise diode is inconvenient. Both fail in characteristic ways when the assumptions are not written down.
1. Y-factor traps
- ENR table uncertainty and connector / adapter loss between the source and the DUT reference plane dominate many "0.1 dB" arguments
- Receiver linearity: if the "hot" state compresses the analyser or NF meter, Y is wrong in a quiet, systematic way
- Mismatch: ENR assumes a stated source Γ; a poorly matched DUT input plus an uncorrected source mismatch converts into NF error, not just a return-loss footnote
2. Cold-source / gain-method traps
- You need a trustworthy gain measurement at the same reference planes; a bad S21 or a skipped de-embed becomes an NF fiction
- kTB reference and temperature of the cold termination must be real, not "room temp ≈ 290 K" handwaving in a hot rack
- Low-gain DUTs make the instrument NF correction critical — know when the meter is measuring itself
3. Habit
Write on every NF sheet: method (Y vs cold-source), ENR / cold temperature, adapter stack, reference plane, and whether mismatch correction was applied. Two labs disagreeing by a few tenths of a dB are often disagreeing about ENR + adapters, not about the LNA.
Discussion: when your last NF correlation failed between stations, was the root cause ENR/adapters, compression on hot, or a gain/de-embed error on the cold-source path — and do you require mismatch notes on the traveler?
— rfforum 编辑部 / Editorial Team
