L
Admin
Yagi-Uda looks simple — one driven element, a reflector, a string of directors on a boom — until the first article shows a pretty S11 and a front-to-back that collapsed at the band edge you care about. A few habits that have saved rebuilds.
**1. Driven element is not “a dipole you already know”**
- Length starts near 0.45–0.49 λ in free space, then walks with boom diameter, element diameter, and the feed gap.
- A **folded dipole** driven element is common for a reason: ~4× the feed impedance of a thin dipole, so a 200–300 Ω folded feed + 4:1 balun lands nearer 50 Ω than a skinny centre-fed dipole that wants a different transform.
- Freeze driven length **after** the balun and boom are present. Tuning the dipole alone on a foam stand, then bolting it to the boom, is a classic “S11 moved 20 MHz” surprise.
**2. Reflector and directors are spacing games, not magic lengths**
Rough starting map (free-space λ of the design centre — then EM / range tune):
- Reflector ≈ 0.5 λ, ~0.15–0.25 λ behind the driven.
- Directors slightly shorter than driven, spaced ~0.15–0.35 λ; more directors buy gain slowly and narrow the bandwidth / F/B faster than the brochure implies.
- Element diameter and boom conductivity change electrical length — a thick tube on a fat conductive boom is not the same Yagi as a thin wire model.
Habit: change **one** variable per cut (one director length, or one spacing), and log gain + F/B + match together. Optimising S11 alone invents a match that radiates backwards.
**3. The balun / choke is part of the antenna**
A Yagi feed with common-mode on the coax outer is a monopole-plus-boom hybrid. Symptoms:
- Pattern tilts when you grab the feed cable.
- Front-to-back moves when you add a ferrite on the coax.
- Cross-pol fills in the nulls.
Fixes that actually stick: a real balun (folded-dipole 4:1, sleeve, or well-designed bazooka/choke) at the feed, plus a few clamp-on ferrites as a **sanity** check — if ferrites still move the pattern, the balun is not done. Do not rely on ferrites as the production balun.
**4. What usually fails first in a “finished” Yagi**
- **Front-to-back at the band edges**, not peak gain at f0.
- **Boom / mast interaction** when the mast is parallel to the elements or the clamp sits in the high-current region of the driven element.
- **Detuning by ice / radome / paint** on outdoor units — re-check match and F/B after the mechanical finish, not only on the bare metal prototype.
**5. A short acceptance gate**
1. Match across the *use* band with the production balun and mast clamp.
2. Relative F/B (or rear null depth) at centre and both edges — same range geometry.
3. Cable-dress / ferrite smoke test: if grabbing the coax or sliding a ferrite moves F/B or match >~1 dB, stop and fix CM before you publish a gain number.
**Discussion**
On your last Yagi (or Yagi-like board/boom array), what died first at the band edge — match, front-to-back, or pattern tilt from feed CM — and did you end with a folded-dipole + 4:1 or a different balun?
— rfforum 编辑部 / Editorial Team
**1. Driven element is not “a dipole you already know”**
- Length starts near 0.45–0.49 λ in free space, then walks with boom diameter, element diameter, and the feed gap.
- A **folded dipole** driven element is common for a reason: ~4× the feed impedance of a thin dipole, so a 200–300 Ω folded feed + 4:1 balun lands nearer 50 Ω than a skinny centre-fed dipole that wants a different transform.
- Freeze driven length **after** the balun and boom are present. Tuning the dipole alone on a foam stand, then bolting it to the boom, is a classic “S11 moved 20 MHz” surprise.
**2. Reflector and directors are spacing games, not magic lengths**
Rough starting map (free-space λ of the design centre — then EM / range tune):
- Reflector ≈ 0.5 λ, ~0.15–0.25 λ behind the driven.
- Directors slightly shorter than driven, spaced ~0.15–0.35 λ; more directors buy gain slowly and narrow the bandwidth / F/B faster than the brochure implies.
- Element diameter and boom conductivity change electrical length — a thick tube on a fat conductive boom is not the same Yagi as a thin wire model.
Habit: change **one** variable per cut (one director length, or one spacing), and log gain + F/B + match together. Optimising S11 alone invents a match that radiates backwards.
**3. The balun / choke is part of the antenna**
A Yagi feed with common-mode on the coax outer is a monopole-plus-boom hybrid. Symptoms:
- Pattern tilts when you grab the feed cable.
- Front-to-back moves when you add a ferrite on the coax.
- Cross-pol fills in the nulls.
Fixes that actually stick: a real balun (folded-dipole 4:1, sleeve, or well-designed bazooka/choke) at the feed, plus a few clamp-on ferrites as a **sanity** check — if ferrites still move the pattern, the balun is not done. Do not rely on ferrites as the production balun.
**4. What usually fails first in a “finished” Yagi**
- **Front-to-back at the band edges**, not peak gain at f0.
- **Boom / mast interaction** when the mast is parallel to the elements or the clamp sits in the high-current region of the driven element.
- **Detuning by ice / radome / paint** on outdoor units — re-check match and F/B after the mechanical finish, not only on the bare metal prototype.
**5. A short acceptance gate**
1. Match across the *use* band with the production balun and mast clamp.
2. Relative F/B (or rear null depth) at centre and both edges — same range geometry.
3. Cable-dress / ferrite smoke test: if grabbing the coax or sliding a ferrite moves F/B or match >~1 dB, stop and fix CM before you publish a gain number.
**Discussion**
On your last Yagi (or Yagi-like board/boom array), what died first at the band edge — match, front-to-back, or pattern tilt from feed CM — and did you end with a folded-dipole + 4:1 or a different balun?
— rfforum 编辑部 / Editorial Team
