R
Member
An intensity-modulated, direct-detection RoF link is a cascade of an electrical-to-optical conversion, an optical path, and an optical-to-electrical conversion. The RF gain depends on modulator slope, laser power, photodiode responsivity, and optical loss, and it often comes out negative.
A useful first-order chain:
Modulator: V_pi or slope efficiency sets RF-to-optical conversion
Optical loss L_opt (dB) attenuates optical power; in the RF domain the detected RF power changes by roughly 2x L_opt (dB) because photocurrent is proportional to optical power and RF power to current squared
Photodiode: i = R x P_opt, with R ~ 0.8-0.9 A/W near 1550 nm for a typical PIN
Load: RF power = i^2 x R_load / 2
So doubling fibre span, or adding a 3 dB optical splitter, costs about 6 dB of RF gain. This is the single most common surprise when moving from a coax run to a RoF run.
Noise terms to separate:
Laser RIN
Shot noise, proportional to DC photocurrent
Receiver thermal noise and any post-amplifier noise figure
Optical amplifier ASE if an EDFA is present
Measurement sequence:
1) Measure RF gain and S21 with a VNA across the band, with the optical power recorded at the receiver.
2) Measure noise figure using the Y-factor or cold-noise method with the optical link in place; a standard NF meter assumes a purely electrical DUT and can mislead.
3) Check SFDR / two-tone intermodulation, since modulator nonlinearity sets the upper end.
4) Sweep optical received power and plot NF and gain together; there is usually an optimum.
Takeaway: an RoF link is specified by RF gain, noise figure, and dynamic range as a set. Quoting only optical loss or only bandwidth hides the dominant trade-off.
A useful first-order chain:
Modulator: V_pi or slope efficiency sets RF-to-optical conversion
Optical loss L_opt (dB) attenuates optical power; in the RF domain the detected RF power changes by roughly 2x L_opt (dB) because photocurrent is proportional to optical power and RF power to current squared
Photodiode: i = R x P_opt, with R ~ 0.8-0.9 A/W near 1550 nm for a typical PIN
Load: RF power = i^2 x R_load / 2
So doubling fibre span, or adding a 3 dB optical splitter, costs about 6 dB of RF gain. This is the single most common surprise when moving from a coax run to a RoF run.
Noise terms to separate:
Laser RIN
Shot noise, proportional to DC photocurrent
Receiver thermal noise and any post-amplifier noise figure
Optical amplifier ASE if an EDFA is present
Measurement sequence:
1) Measure RF gain and S21 with a VNA across the band, with the optical power recorded at the receiver.
2) Measure noise figure using the Y-factor or cold-noise method with the optical link in place; a standard NF meter assumes a purely electrical DUT and can mislead.
3) Check SFDR / two-tone intermodulation, since modulator nonlinearity sets the upper end.
4) Sweep optical received power and plot NF and gain together; there is usually an optimum.
Takeaway: an RoF link is specified by RF gain, noise figure, and dynamic range as a set. Quoting only optical loss or only bandwidth hides the dominant trade-off.
