Low modulation order with high CQI reports - NSN RU20

[bmaia18]

Guys

I need a help.Im having some cases with cells presenting a very poor Modulation used (mostly using QPSK) while the CQI reported presents samples with very good ranges (which could support higher modulation orders). I made the following analysis...

1 - MAC-HS efficiency is high: Above 90% so its seems not be an issue with BLER and any kind of CQI adjustment based on BLER.

2 - We have just improved by 3 dB the power allowed by HS and also the HS power used is never above 80%. It means that CQI reported is mostly probably not adjusted by the lower power available on Node B. Also the MPO configure is fixed on 6 dB which is quite conservative value for power assumed, thus leading UE to hardly overestimate the power for CQI calculation...

3 - Lots of cells with this behavior are not presenting counters which suggests congestion on Iub side. That could be one of the reasons since in RU10 NSN presents a automatic "spectral efficient link adaptation" which adjusts the power and modulation used to fit the UE data buffer into the TBS calculated by the TFRC algorithm. That feature avoids case of low buffer to use high power and modulation without necessity.

4 - We are not having any Node B MAC-HS buffer full and MAC-D PDU discarted by buffer full, on Node B, is really very low.

5 - 64QAM is enabled too.

6 - MAC-D maximum throughput allowed is compatible with the licenses enabled within the NW, or in other words enough to provide HSDPA+ bitrates

What the hell could limit the Node B to use more often the higher modulation of 16QAM and 64QAM!??? Any hint would be highly appreciated!

Kind Regards
Bruno Maia