CRS Port Number 2 vs 4 -- LTE -- Huawei

[fahmi]

Dears,

With a 4T4R capable cell (Huawei) and a 2T2R capale UE, when setting CRS Port number to 4, there no obvious positif gain on throughput.

When analizing Logs, we notices that, with a 4T4R cell and RS PWR set to 15.2dBm

- if CRS port Number = 4 , the reference signal pwr is 15.2

But If CRS Port Number = 2 , the reference signal pwr from logs is 18.2 and not 15.2 !! (We did not change the RS power setting, always 15.2)

any explanation for this behavior ?

Many thanks

[auto_art]

Dears,

With a 4T4R capable cell (Huawei) and a 2T2R capale UE, when setting CRS Port number to 4, there no obvious positif gain on throughput.

When analizing Logs, we notices that, with a 4T4R cell and RS PWR set to 15.2dBm

- if CRS port Number = 4 , the reference signal pwr is 15.2

But If CRS Port Number = 2 , the reference signal pwr from logs is 18.2 and not 15.2 !! (We did not change the RS power setting, always 15.2)

any explanation for this behavior ?

Many thanks

this is because with 2T you have total RRU power being pumped of two ports.
with 4T, you have 4 TX ports pumping so RRU power is divided in four ports hence per port powe is reduced.

[ls2398]

If you can't increase PDSCHCfg.ReferenceSignalPower then you can always adjust mobility thresholds to keep the same coverage. Which model of RRU are you using and what bandwidth?

[subway]

Dears,With a 4T4R capable cell (Huawei) and a 2T2R capale UE, when setting CRS Port number to 4, there no obvious positif gain on throughput. When analizing Logs, we notices that, with a 4T4R cell and RS PWR set to 15.2dBm- if CRS port Number = 4 , the reference signal pwr is 15.2But If CRS Port Number = 2 , the reference signal pwr from logs is 18.2 and not 15.2 !! (We did not change the RS power setting, always 15.2)any explanation for this behavior ?Many thanksThis is completely normal, the beamforming documentation actually mentions this as well.And the overhead is not just the loss of power: you will also see an increased overhead of you set CRS=4. The signalling overhead will go from around 11-12% (CRS2) to about 22-24% with CRS4. Even if you would have 4x4MIMO capable CPEs, think twice before you switch to CRS4 (and 4x4 MIMO), as if you are not in an ideal environment (large number of 4x4 capable CPEs, good SINR and good refelctive environment), you will just see the loss of power and the CRS overhead, and nothing from the gain. So its not just that you will not see any gains, but you can see reasonable amount of degradation.We have 4x4 capable Huawei CPEs, and in a rural environment we only seen the downside of CRS=4, very little amount of close-by CPEs actually switched to 4x4. So we switched it off.In rural environment I would recommend DL 2x2 MIMO with TM9, and multi user beamforming. If you have old CPEs, then Dual-stream beamforming (TM8).Do you have 4T4R or 8T8R antennas?

[fahmi]

This is completely normal, the beamforming documentation actually mentions this as well.And the overhead is not just the loss of power: you will also see an increased overhead of you set CRS=4. The signalling overhead will go from around 11-12% (CRS2) to about 22-24% with CRS4. Even if you would have 4x4MIMO capable CPEs, think twice before you switch to CRS4 (and 4x4 MIMO), as if you are not in an ideal environment (large number of 4x4 capable CPEs, good SINR and good refelctive environment), you will just see the loss of power and the CRS overhead, and nothing from the gain. So its not just that you will not see any gains, but you can see reasonable amount of degradation.We have 4x4 capable Huawei CPEs, and in a rural environment we only seen the downside of CRS=4, very little amount of close-by CPEs actually switched to 4x4. So we switched it off.In rural environment I would recommend DL 2x2 MIMO with TM9, and multi user beamforming. If you have old CPEs, then Dual-stream beamforming (TM8).Do you have 4T4R or 8T8R antennas?Dear, for my case it's not CPE but ordinary 2T2R smartphones. Could you please share the document you are refering to ? My problem is with the value of RS Power reported in SIB 2. I set it to 15.2, but when CRS 2 it's reported as 18.2 and not 15.2 in SIB 2 msg and when CRS port 4 , it's reported as 15.2 !

[slava121]

Hi, fahmi !What radio unit RRU 4T4R do you use in your case?
Dear, for my case it's not CPE but ordinary 2T2R smartphones. Could you please share the document you are refering to ? My problem is with the value of RS Power reported in SIB 2. I set it to 15.2, but when CRS 2 it's reported as 18.2 and not 15.2 in SIB 2 msg and when CRS port 4 , it's reported as 15.2 !

[fahmi]

Hi, fahmi !What radio unit RRU 4T4R do you use in your case?Dear,We are using 4T4R RRU 4*40W RRU3971

[slava121]

Do you set the power to 40 watts? How many resource blocks do you use?

[fahmi]

Do you set the power to 40 watts? How many resource blocks do you use?Yes, total power used is 40W. I don't set PWR in Huawei, I just set RS, PA and PB and total power is calculated based on these parameter and it's 40W per each port. (in total 4*40W) and BW is 20M (100 RB)

[slava121]

Do you have a description of DL 4x4 MIMO HUAWEI?

[slava121]

Do you have Skype?

[subway]

Dear, for my case it's not CPE but ordinary 2T2R smartphones. Could you please share the document you are refering to ? My problem is with the value of RS Power reported in SIB 2. I set it to 15.2, but when CRS 2 it's reported as 18.2 and not 15.2 in SIB 2 msg and when CRS port 4 , it's reported as 15.2 !

"In TDD 4T or 8T scenarios, the RSRP decreases by 3 dB after the number of CRS ports changes from 2 to 4. If the transmit power is not increased"

This is from the Huawei docs.

I think your confusion is why CRS4 eats more power? Well the reason is you have physical and logical antenna ports.

Example: if you configure your cells to 4T4R and CRS2, you will get all the 40watts on each antenna port. (this is your recommended setting by the way).

If you configure CRS4, you will add logical antenna ports, and the extra RS needs extra power (on each port).

For you usecase, I recommend 4T4R cells with CRS2. You can introduce TM7 if you have capable UEs. I would highly recommend "4-way received diversity", and IRC (both for PUCCH and PUSCH), these are all require enodeb licenses.

42096