L
Admin
[Technical article, rfforum Editorial Team]
A directional coupler is sold on coupling, main-line loss and directivity. Directivity is the one that decides whether your reflected-power or VSWR reading is trustworthy, and datasheet "isolation" is not the same thing.
1. Definitions
- Coupling C ≈ −20·log10(|S31|) when port 1 is input and port 3 is the coupled arm (forward).
- Isolation I ≈ −20·log10(|S32|) with port 1 driven and the through port terminated (signal leaking into the coupled arm from the "wrong" direction).
- Directivity D = I − C (in dB). Equivalently D = −20·log10(|S32|/|S31|) under the right terminations.
High coupling (e.g. 10 dB) with mediocre isolation can still yield poor directivity; a 20 dB coupler with excellent isolation can have better directivity than a "tighter" coupler.
2. Why directivity bites VSWR monitors
A reflectometer estimates Γ from the reverse-coupled wave. Finite directivity mixes a fraction of the forward wave into the reverse reading. Roughly, the smallest |Γ| you can trust is on the order of the directivity error. A coupler with only 15 dB directivity cannot honestly resolve a 1.05 VSWR on a production line without correction.
3. Measuring directivity with a VNA
1. Calibrate at the coupler ports (or use a fixture de-embed).
2. Drive the input, terminate the through port in a high-quality load, measure the coupled port → related to isolation.
3. Drive the through port as if it were the input for the reverse path (or use the defined reverse drive), terminate the input, measure the coupled port → related to coupling in the reverse sense.
4. Compute D from the two paths at each frequency. Watch connector repeatability; a bad load will look like bad directivity.
4. Practical layout notes
- The termination on the isolated port of a four-port coupler must be excellent; a mediocre chip resistor on a poor ground is often the directivity limit, not the coupled-line geometry.
- Frequency edges: directivity usually collapses at the band edges first. Spec the band where D stays above your system need, not where coupling is "about right".
Discussion: for in-line PA reflected-power monitoring at 2–6 GHz, what directivity do you require before you trust a fold-back alarm, and do you calibrate the coupler+detector as a set?
— rfforum 编辑部 / Editorial Team
A directional coupler is sold on coupling, main-line loss and directivity. Directivity is the one that decides whether your reflected-power or VSWR reading is trustworthy, and datasheet "isolation" is not the same thing.
1. Definitions
- Coupling C ≈ −20·log10(|S31|) when port 1 is input and port 3 is the coupled arm (forward).
- Isolation I ≈ −20·log10(|S32|) with port 1 driven and the through port terminated (signal leaking into the coupled arm from the "wrong" direction).
- Directivity D = I − C (in dB). Equivalently D = −20·log10(|S32|/|S31|) under the right terminations.
High coupling (e.g. 10 dB) with mediocre isolation can still yield poor directivity; a 20 dB coupler with excellent isolation can have better directivity than a "tighter" coupler.
2. Why directivity bites VSWR monitors
A reflectometer estimates Γ from the reverse-coupled wave. Finite directivity mixes a fraction of the forward wave into the reverse reading. Roughly, the smallest |Γ| you can trust is on the order of the directivity error. A coupler with only 15 dB directivity cannot honestly resolve a 1.05 VSWR on a production line without correction.
3. Measuring directivity with a VNA
1. Calibrate at the coupler ports (or use a fixture de-embed).
2. Drive the input, terminate the through port in a high-quality load, measure the coupled port → related to isolation.
3. Drive the through port as if it were the input for the reverse path (or use the defined reverse drive), terminate the input, measure the coupled port → related to coupling in the reverse sense.
4. Compute D from the two paths at each frequency. Watch connector repeatability; a bad load will look like bad directivity.
4. Practical layout notes
- The termination on the isolated port of a four-port coupler must be excellent; a mediocre chip resistor on a poor ground is often the directivity limit, not the coupled-line geometry.
- Frequency edges: directivity usually collapses at the band edges first. Spec the band where D stays above your system need, not where coupling is "about right".
Discussion: for in-line PA reflected-power monitoring at 2–6 GHz, what directivity do you require before you trust a fold-back alarm, and do you calibrate the coupler+detector as a set?
— rfforum 编辑部 / Editorial Team
