Will CFI be equal 3 for 1.4, 3 and 5 MHz system bandwidths in all cases?

Only in 5G VoNR QCI1 1 can be sent in RLC AM mode.

To avoid drops RLC AM has been used.

QCI1 used RLC UM.

Yes, you’re absolutely right.
But changes have done for that too.
How? I don’t know…

Either way, it is still a best effort.
You will not see a VoLTE released when GBR is not fulfilled.

Point is: if we Optimize PDCCH Spare Power for WEAK COVERAGE USER will it increase CFI 3 and degrade DL user throughput as Large CFI value means less space for PDSCH?

But it does not use all symbols it uses as per requirement of UE scheduling.
If CCE are more and one symbol cannot handle then this dynamic CFI is useful.
It doesn’t use other two symbols all resource element or else it is a waste.
One CFI will be used rest two CFI re will be used depending on CCE require.
So they have PDCCH max code rate.
In Huawei.
To compensate both Coverage and Capacity.

CFI mode auto is better for < 5 Mhz due to SR and CQI per sector admission would be better controlled (under 900/1800 layer).

For weak coverage CCE used can be shifted from 8 to 4 cce - I don’t remember the name.
PDCCH agg and CCE utility KPI can let us know whether we have spare available for PDSCH resource, which can be used for Throughput.
Oher, if no scheduling no Throughput for that UE.

No of PHICH group depends on Ng & is given by Ng * Ceiling (No of RBs/8).
A PHICH group occupy 12 symbols and can carry ACK/NACK for 8 UEs.
For eg, in 20 MHz, for Ng=1, no of PHICH group= 13, Total no of PHICH Symbols= 13*12=156 Symbols.
If Ng=1/2, the no PHICH Symbols will become half and the no of CCEs will increase.

Yes.
This is called PDCCH power boosting.
All the operators are doing this as I know to overcome the capacity constraint.
Higher is the aggregation level lower is the capacity.