L
Admin
[Technical article / FAQ, rfforum Editorial Team]
Coupler datasheets quote directivity in a connectorised fixture. On a real bench, adapters, torque, and the DUT port residual often eat more of your "effective" directivity than the coupler ageing.
1. What you are actually measuring
- Directivity is a ratio of coupled forward vs reverse paths; any mismatch or connector residual at the test ports converts into apparent directivity loss
- A "good" 30 dB coupler behind two mediocre adapters and a loose N torque can look like 20 dB in a reflectometer or ALC loop
- Frequency matters: edge-of-band directivity collapse is common; a single mid-band number is not a traveler
2. Freeze list for a coupler in a reflectometer / leveling / sampling path
- Adapter stack and connector sex kept identical between cal and use (same habit as VNA residual — see also /t/306, /t/315)
- Torque / gauge discipline on every reconnect that sits in the coupled path
- Whether you verify directivity with a known good load / short at the working band edges, not only at one spot frequency
- Thermal / flex: semi-rigid that moved between cal and DUT can look like coupler drift
3. Habit
Log coupler model + adapter stack + a quick directivity spot-check (or open/short asymmetry) on the same sheet as the measurement that depended on it. Arguing about 0.2 dB mismatch correction while the coupled path lost 10 dB of effective directivity is the wrong fight.
Discussion: when your last reflectometer or ALC loop lied, was the root cause coupler directivity, adapter residual, or a load that was not as good as its label — and do you re-check directivity after every adapter change?
— rfforum 编辑部 / Editorial Team
Coupler datasheets quote directivity in a connectorised fixture. On a real bench, adapters, torque, and the DUT port residual often eat more of your "effective" directivity than the coupler ageing.
1. What you are actually measuring
- Directivity is a ratio of coupled forward vs reverse paths; any mismatch or connector residual at the test ports converts into apparent directivity loss
- A "good" 30 dB coupler behind two mediocre adapters and a loose N torque can look like 20 dB in a reflectometer or ALC loop
- Frequency matters: edge-of-band directivity collapse is common; a single mid-band number is not a traveler
2. Freeze list for a coupler in a reflectometer / leveling / sampling path
- Adapter stack and connector sex kept identical between cal and use (same habit as VNA residual — see also /t/306, /t/315)
- Torque / gauge discipline on every reconnect that sits in the coupled path
- Whether you verify directivity with a known good load / short at the working band edges, not only at one spot frequency
- Thermal / flex: semi-rigid that moved between cal and DUT can look like coupler drift
3. Habit
Log coupler model + adapter stack + a quick directivity spot-check (or open/short asymmetry) on the same sheet as the measurement that depended on it. Arguing about 0.2 dB mismatch correction while the coupled path lost 10 dB of effective directivity is the wrong fight.
Discussion: when your last reflectometer or ALC loop lied, was the root cause coupler directivity, adapter residual, or a load that was not as good as its label — and do you re-check directivity after every adapter change?
— rfforum 编辑部 / Editorial Team
