L
Admin
[Technical article / FAQ, rfforum Editorial Team]
An analog RoF or microwave-photonic receiver that "goes quiet" after a brief optical overload is often not a RF chain failure. The photodiode and TIA are still recovering from a large optical step; the RF looks dead while the front-end bias and photocurrent settle. Blaming the laser or the RF amp first invents the wrong ECO.
1. What overload actually does
- A large optical pulse or accidental connector flash drives the photodiode into high photocurrent. The TIA rail, bias network, and AC-coupling time constants then dominate recovery, not the RF matching.
- Recovery can last from microseconds to tens of milliseconds depending on the TIA design, AC-coupling, and whether a protection clamp engaged. During that window, RF gain and noise figure look catastrophic.
- A link that fails only after connector mating, cleaning, or a hot-plug optical path is a recovery story until proven otherwise.
2. Freeze list before you redesign the RF chain
- Optical power into the PD before and after the event (or a calibrated attenuator A/B)
- Scope capture of PD bias / TIA output during recovery, not only RF spectrum after the fact
- Whether a fixed optical attenuator or a slower optical ramp removes the outage
- Cable and connector: did the event coincide with a mate, a dirty ferrule flash, or a EDFA transient?
3. Habit
Never open a "RoF RF dead after connector work" ticket without an optical-power and recovery-time note. Archive max optical power, recovery time to within 1 dB of steady RF, and whether a clamp fired. If a fixed pad or soft optical ramp kills the outage, the RF amp was never the patient.
4. Practical gate
- Cap optical power during mating; use attenuators when probing
- Measure recovery time to RF within 1 dB; write it next to the link budget
- If recovery > your protocol gap, fix optical handling or PD/TIA protection before RF redesign
Discussion: on your last RoF "dead after mate" event, how long did RF take to return, and do you require max optical power + recovery time on every acceptance log?
— rfforum 编辑部 / Editorial Team
An analog RoF or microwave-photonic receiver that "goes quiet" after a brief optical overload is often not a RF chain failure. The photodiode and TIA are still recovering from a large optical step; the RF looks dead while the front-end bias and photocurrent settle. Blaming the laser or the RF amp first invents the wrong ECO.
1. What overload actually does
- A large optical pulse or accidental connector flash drives the photodiode into high photocurrent. The TIA rail, bias network, and AC-coupling time constants then dominate recovery, not the RF matching.
- Recovery can last from microseconds to tens of milliseconds depending on the TIA design, AC-coupling, and whether a protection clamp engaged. During that window, RF gain and noise figure look catastrophic.
- A link that fails only after connector mating, cleaning, or a hot-plug optical path is a recovery story until proven otherwise.
2. Freeze list before you redesign the RF chain
- Optical power into the PD before and after the event (or a calibrated attenuator A/B)
- Scope capture of PD bias / TIA output during recovery, not only RF spectrum after the fact
- Whether a fixed optical attenuator or a slower optical ramp removes the outage
- Cable and connector: did the event coincide with a mate, a dirty ferrule flash, or a EDFA transient?
3. Habit
Never open a "RoF RF dead after connector work" ticket without an optical-power and recovery-time note. Archive max optical power, recovery time to within 1 dB of steady RF, and whether a clamp fired. If a fixed pad or soft optical ramp kills the outage, the RF amp was never the patient.
4. Practical gate
- Cap optical power during mating; use attenuators when probing
- Measure recovery time to RF within 1 dB; write it next to the link budget
- If recovery > your protocol gap, fix optical handling or PD/TIA protection before RF redesign
Discussion: on your last RoF "dead after mate" event, how long did RF take to return, and do you require max optical power + recovery time on every acceptance log?
— rfforum 编辑部 / Editorial Team
