Low DL Throughput - LTE 4x4 MIMO 64QAM

4G LTE
, ,
1

Hello Experts.

For low DL Throughput (~225 Mbps, expected 291 Mbps) using 64 QAM configuration with 20 MHz BW 4x4 MIMO on LTE eNB I see very bad PUSCH SINR (-18dB avg) and considerable BLER (8% avg with many spikes during the interval) but other parameters in expected range (CQI=15, numRB=100, RI=4, numCW=2).

Can I suspect the SINR to have caused this low DL Throughput?
If not, what other parameters can be checked for here?
Please note this test in ideal conditions in lab with single UE and only one cell up.

Any pointers would be helpful.

Low DL Throughput - LTE 4x4 MIMO 64QAM
Low DL Throughput - LTE 4x4 MIMO 64QAM991×602 66 KB

2

Hello Experts.

I would like your help in explaining how should do I increase the throughputs in the LTE cells.

I implemented the MIMO 4T4R, but I don’t see an increase in Throughput.

3

CQI
DL QAM
% of good qual samples and TA
RI check on field via tool.

4

Is there any change in RI distribution after 4t4r Implementation?
Do you see any raise of traffic?
Did you check TA pre and post?
These details will help to identify

What is the power in pre (may be 2t2r) and post (4t4r)?

If it’s Ericsson system, then you have counters for RI and ta too you can evaluate… So you don’t need any field support to do the 1st cut analysis…

5

This is Huawei System.

6

I heard in Huawei too there are counters for RI.

7

Dear:

First need to check if you are getting Rank 3 and Rank 4 samples in post benchmark means if it is implemented or not.

Secondly, in many cases it have been witnessed that no major gain in throughout observed after 4T4R implementation for the reason of UE penetration as not all mobile phones supported 4T4R. You can verify it with the help of DT by comparing pre (2T2R) and post (4T4R) results.

Share the trends of MIMO rank 3 and MIMO rank 4 indicator.

8

Is there any way we could optimize as per parameters besides physical optimization?

9

The problem might not be rank3 or rank4 but evm dynamic or static power back off parameter decreases cell dl pwr with offset which helps better 256qam modulation ratio.

For rank4, sweet spot and also receive level between feeders should be aligned, if there is big gap between them UE will not report rank4.

10

4T4R is an alternative to increase Throughput but it is necessary to have goods CQI (that means SNR in the mobile). Since to have 4T4R good DL Quality is required. So, verify or avoid:

  • Overshooting cells
  • Big overlap with neighbor sites

Some other possible actions:

  • Decrease the CRS power and increase pdsch power, by for example setting CRSGain = 0 and pdchtypeBgain =0 (improving UL)
  • Promote a good load balance between frequency cells. This can be achieved by rieviewing LM and load balance strategies.
  • Carrier Aggregation and features related to it. Maybe the faster way to achieve a better Throughput .
  • Activated features as instant 256QAM DL, among others.
11

I guess the root cause may be really due to the bad PUSCH SINR.
If it is open-loop beamforming (FDD), eNB will employing implicit beamforming by estimating the UL SINR, and dNB thought the channel condition is very bad.
If it is closed-loop beamforming(TDD), eNB will employing explicit beamforming via the PUSCH beamforming report, which may be corrupted due to very bad SINR (-18dB).
Thus anyway you can’t get high DL throughput as you expected.