L
Admin
[Technical article, rfforum Editorial Team]
In a long analog RF-over-fiber (IM/DD) link, chromatic dispersion in the fiber turns laser chirp and double-sideband modulation into RF power fade versus frequency. The RF response is not flat just because the optical loss is flat.
1. Where the nulls come from
Double-sideband intensity modulation produces two optical sidebands. Fiber dispersion adds opposite phase to the upper and lower sidebands; at the photodiode they can cancel at specific RF frequencies. The first null frequency scales roughly as 1/√(D·L·λ²) (order-of-magnitude; chirp and dual-drive MZM details move it). Longer fiber or higher RF carrier → earlier null.
2. Chirp makes it worse (or shifts it)
Directly modulated lasers have adiabatic/transient chirp; the FM–AM interplay with dispersion shifts and deepens fades. Externally modulated links with a low-chirp MZM are kinder, but double-sideband MZM links still show dispersion nulls on long SMF.
3. Mitigation patterns that work
- Shorter fiber or a dispersion-compensating span sized for your RF band.
- Optical single-sideband (OSSB) or vestigial-sideband modulation so the two sidebands cannot cancel.
- Mid-span filtering / sideband suppression.
- Stay below the first null with margin for temperature (D and effective length drift).
- For multi-km IF transport, prefer a lower IF or a digital CPRI/eCPRI style link if the budget allows — analog only when you truly need the RF waveform.
4. Bench check
Sweep RF small-signal S21 through the link while you vary fiber length (or temperature). A deep notch that moves with length is dispersion fading, not a bad photodiode. Do not equalise it with a one-time RF tilt that will be wrong on the next spool.
Discussion: for your longest analog RoF run, where did the first RF null sit, and did you fix it with OSSB, less fiber, or by dropping IF?
— rfforum 编辑部 / Editorial Team
In a long analog RF-over-fiber (IM/DD) link, chromatic dispersion in the fiber turns laser chirp and double-sideband modulation into RF power fade versus frequency. The RF response is not flat just because the optical loss is flat.
1. Where the nulls come from
Double-sideband intensity modulation produces two optical sidebands. Fiber dispersion adds opposite phase to the upper and lower sidebands; at the photodiode they can cancel at specific RF frequencies. The first null frequency scales roughly as 1/√(D·L·λ²) (order-of-magnitude; chirp and dual-drive MZM details move it). Longer fiber or higher RF carrier → earlier null.
2. Chirp makes it worse (or shifts it)
Directly modulated lasers have adiabatic/transient chirp; the FM–AM interplay with dispersion shifts and deepens fades. Externally modulated links with a low-chirp MZM are kinder, but double-sideband MZM links still show dispersion nulls on long SMF.
3. Mitigation patterns that work
- Shorter fiber or a dispersion-compensating span sized for your RF band.
- Optical single-sideband (OSSB) or vestigial-sideband modulation so the two sidebands cannot cancel.
- Mid-span filtering / sideband suppression.
- Stay below the first null with margin for temperature (D and effective length drift).
- For multi-km IF transport, prefer a lower IF or a digital CPRI/eCPRI style link if the budget allows — analog only when you truly need the RF waveform.
4. Bench check
Sweep RF small-signal S21 through the link while you vary fiber length (or temperature). A deep notch that moves with length is dispersion fading, not a bad photodiode. Do not equalise it with a one-time RF tilt that will be wrong on the next spool.
Discussion: for your longest analog RoF run, where did the first RF null sit, and did you fix it with OSSB, less fiber, or by dropping IF?
— rfforum 编辑部 / Editorial Team
