L
Admin
[Technical article / FAQ, rfforum Editorial Team]
Uplink control that "hunts" in power or EVM is often blamed on channel fading when the real fight is between UE power control, base-station AGC, and the timing of the PUCCH grant. The air looks bad; the logs show a loop chasing itself.
1. What the loops actually collide on
- UE open-loop and closed-loop power control set PUSCH/PUCCH power from path loss and TPC commands. AGC on the gNB side simultaneously moves the front-end gain to keep the ADC happy.
- If AGC steps are large or slow relative to the PUCCH format duration, the UE sees a changing RSSI/RSRP estimate and keeps correcting; the gNB sees a moving input level and keeps AGC-ing. The spectrogram looks like fading; the counters look like power-control thrash.
- Format 0/1 vs longer formats, and whether SRS or a wideband RS is available for the same slot, change how fast the UE can estimate the channel. A short PUCCH on a fast AGC is the classic trap.
2. Freeze list before you open a fading or PA-linearity ECO
- gNB AGC step size, attack/release, and whether AGC is frozen during the PUCCH symbol window
- UE TPC history and commanded vs applied power for the same HARQ/PUCCH occasions
- Whether the hunt survives with a fixed gNB gain (lab) or a narrower AGC range
- Time alignment: does the hunt correlate with AGC gain changes in the same few slots, not with Doppler or multipath metrics?
3. Habit
Never freeze "severe fading on PUCCH" from a spectrogram alone. Archive AGC settings and UE TPC traces next to every uplink control EVM or miss claim. If freezing AGC or slowing its attack kills the hunt, the channel model was never the patient.
4. Practical gate
- Reproduce with fixed gain on the gNB RF; if the hunt dies, fix AGC/TPC interaction first
- Log commanded vs applied UE power on the same timeline as gNB AGC gain
- Check whether only short PUCCH formats fail while longer PUSCH looks fine — that is a loop-timing clue
Discussion: on your last PUCCH power hunt, did the fix land in AGC freeze/window, TPC step size, or format choice — and do you require a fixed-gain lab check before blaming the channel?
— rfforum 编辑部 / Editorial Team
Uplink control that "hunts" in power or EVM is often blamed on channel fading when the real fight is between UE power control, base-station AGC, and the timing of the PUCCH grant. The air looks bad; the logs show a loop chasing itself.
1. What the loops actually collide on
- UE open-loop and closed-loop power control set PUSCH/PUCCH power from path loss and TPC commands. AGC on the gNB side simultaneously moves the front-end gain to keep the ADC happy.
- If AGC steps are large or slow relative to the PUCCH format duration, the UE sees a changing RSSI/RSRP estimate and keeps correcting; the gNB sees a moving input level and keeps AGC-ing. The spectrogram looks like fading; the counters look like power-control thrash.
- Format 0/1 vs longer formats, and whether SRS or a wideband RS is available for the same slot, change how fast the UE can estimate the channel. A short PUCCH on a fast AGC is the classic trap.
2. Freeze list before you open a fading or PA-linearity ECO
- gNB AGC step size, attack/release, and whether AGC is frozen during the PUCCH symbol window
- UE TPC history and commanded vs applied power for the same HARQ/PUCCH occasions
- Whether the hunt survives with a fixed gNB gain (lab) or a narrower AGC range
- Time alignment: does the hunt correlate with AGC gain changes in the same few slots, not with Doppler or multipath metrics?
3. Habit
Never freeze "severe fading on PUCCH" from a spectrogram alone. Archive AGC settings and UE TPC traces next to every uplink control EVM or miss claim. If freezing AGC or slowing its attack kills the hunt, the channel model was never the patient.
4. Practical gate
- Reproduce with fixed gain on the gNB RF; if the hunt dies, fix AGC/TPC interaction first
- Log commanded vs applied UE power on the same timeline as gNB AGC gain
- Check whether only short PUCCH formats fail while longer PUSCH looks fine — that is a loop-timing clue
Discussion: on your last PUCCH power hunt, did the fix land in AGC freeze/window, TPC step size, or format choice — and do you require a fixed-gain lab check before blaming the channel?
— rfforum 编辑部 / Editorial Team
