F
Member
[Technical note — Fiberaw / F-tone]
Teams often pick "analog RoF" or "digital RoF" from habit. Below is the decision checklist we use before freezing an architecture — not a product pitch.
**When analog RoF usually wins**
1. You need multi-GHz instantaneous bandwidth with one laser + one PD (DAS IF/RF transport, EW, wideband test ranges).
2. Latency must stay near fiber delay only — no DSP framing / FEC round-trip.
3. The RF waveform must stay intact (phase noise, AM-PM, multi-carrier) end-to-end.
4. You can keep the link in the linear region: MZM/PD bias, optical power, and IMD gates are under control.
**When digital RoF (digitize → optics → reconstruct) usually wins**
1. Long haul or many splices where analog CNR / SFDR would die before the far end.
2. You already own a clean ADC/DAC chain and care more about BER than RF fidelity of the optical path.
3. You need wavelength / packet switching, monitoring, or protection switching that digital optics already solve.
4. Regulatory or site rules force regeneration / encryption at digital layers.
**Hybrid red flags (do not mix blindly)**
- Analog RoF into a site that then re-digitizes poorly → you pay optical SFDR twice and still get ADC spurs.
- Digital RoF with a cheap clock tree → you moved the problem from RIN to sampling jitter.
- "Just add an EDFA" on analog without re-checking NF, SBS, and PD overload.
**Minimum sign-off artefacts**
- One-page decision: BW, distance, latency, SFDR/CNR or EVM/BER gate, and which column fails first.
- For analog: RF gain vs optical attenuator sweep (confirm square-law slope ~2).
- For digital: eye/BER vs optical budget and clock-jitter sensitivity.
Discussion: on your last DAS / range / satcom-IF transport job, which constraint forced the choice — latency, SFDR, distance, or ops/monitoring?
— Fiberaw / F-tone 官方
Teams often pick "analog RoF" or "digital RoF" from habit. Below is the decision checklist we use before freezing an architecture — not a product pitch.
**When analog RoF usually wins**
1. You need multi-GHz instantaneous bandwidth with one laser + one PD (DAS IF/RF transport, EW, wideband test ranges).
2. Latency must stay near fiber delay only — no DSP framing / FEC round-trip.
3. The RF waveform must stay intact (phase noise, AM-PM, multi-carrier) end-to-end.
4. You can keep the link in the linear region: MZM/PD bias, optical power, and IMD gates are under control.
**When digital RoF (digitize → optics → reconstruct) usually wins**
1. Long haul or many splices where analog CNR / SFDR would die before the far end.
2. You already own a clean ADC/DAC chain and care more about BER than RF fidelity of the optical path.
3. You need wavelength / packet switching, monitoring, or protection switching that digital optics already solve.
4. Regulatory or site rules force regeneration / encryption at digital layers.
**Hybrid red flags (do not mix blindly)**
- Analog RoF into a site that then re-digitizes poorly → you pay optical SFDR twice and still get ADC spurs.
- Digital RoF with a cheap clock tree → you moved the problem from RIN to sampling jitter.
- "Just add an EDFA" on analog without re-checking NF, SBS, and PD overload.
**Minimum sign-off artefacts**
- One-page decision: BW, distance, latency, SFDR/CNR or EVM/BER gate, and which column fails first.
- For analog: RF gain vs optical attenuator sweep (confirm square-law slope ~2).
- For digital: eye/BER vs optical budget and clock-jitter sensitivity.
Discussion: on your last DAS / range / satcom-IF transport job, which constraint forced the choice — latency, SFDR, distance, or ops/monitoring?
— Fiberaw / F-tone 官方
