r
[Cross-site technical note from rfforum 编辑部]
When an analog RF-over-fiber link looks "noisy" on a spectrum analyzer, teams often blame the laser RIN first. On the bench the floor is frequently a stack of three contributors, and the order of attack matters.
1. Laser RIN, transferred by the link gain.
Directly modulated links often show the laser RIN peak in the same offset region as close-in RF phase noise. Measure RIN (or a well-known optical attenuator substitution) before you redesign the RF LO.
2. Photodiode + TIA / post-amp noise.
After the optical hit on the PD, shot noise and the TIA's input-referred current noise set a floor that scales differently with received optical power than RIN does. A quick Popt sweep separates "RIN-like" (improves then flattens) from "TIA thermal" (keeps improving with more optical power until PD compression).
3. Fiber dispersion turning intensity noise into phase / amplitude structure.
On longer fiber, dispersion can convert intensity fluctuations into RF amplitude/phase structure that looks like close-in noise or fading. If the spur or floor moves when you change fiber length or wavelength, suspect dispersion before you swap mixers.
Practical gate we use with RF colleagues: freeze optical receive power, swap a short patch for the long spool, and re-measure L(f) or the noise pedestal. If it moves a lot, the RF synthesizer may be innocent.
Question for OEForum: for IF transport in the 1–6 GHz range over a few km, what is usually your first measurement when the RF engineer says "your link ruined my phase noise" — RIN, PD/TIA, or a fiber-length swap?
(Technical only; no product pitch.)
— rfforum 编辑部
When an analog RF-over-fiber link looks "noisy" on a spectrum analyzer, teams often blame the laser RIN first. On the bench the floor is frequently a stack of three contributors, and the order of attack matters.
1. Laser RIN, transferred by the link gain.
Directly modulated links often show the laser RIN peak in the same offset region as close-in RF phase noise. Measure RIN (or a well-known optical attenuator substitution) before you redesign the RF LO.
2. Photodiode + TIA / post-amp noise.
After the optical hit on the PD, shot noise and the TIA's input-referred current noise set a floor that scales differently with received optical power than RIN does. A quick Popt sweep separates "RIN-like" (improves then flattens) from "TIA thermal" (keeps improving with more optical power until PD compression).
3. Fiber dispersion turning intensity noise into phase / amplitude structure.
On longer fiber, dispersion can convert intensity fluctuations into RF amplitude/phase structure that looks like close-in noise or fading. If the spur or floor moves when you change fiber length or wavelength, suspect dispersion before you swap mixers.
Practical gate we use with RF colleagues: freeze optical receive power, swap a short patch for the long spool, and re-measure L(f) or the noise pedestal. If it moves a lot, the RF synthesizer may be innocent.
Question for OEForum: for IF transport in the 1–6 GHz range over a few km, what is usually your first measurement when the RF engineer says "your link ruined my phase noise" — RIN, PD/TIA, or a fiber-length swap?
(Technical only; no product pitch.)
— rfforum 编辑部
