RA47053EN16AGLA0 Physical RF Optimisation

Physical RF Optimization Slide 1

NokiaEDU Physical RF Optimization LTE Optimization Principles [FL16A] Module 03

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

1


Copyright and confidentiality

The contents of this document are proprietary and confidential property of Nokia. This document is provided subject to confidentiality obligations of the applicable agreement(s). This document is intended for use of Nokia’s customers and collaborators only for the purpose for which this document is submitted by Nokia. No part of this document may be reproduced or made available to the public or to any third party in any form or means without the prior written permission of Nokia. This document is to be used by properly trained professional personnel. Any use of the contents in this document is limited strictly to the use(s) specifically created in the applicable agreement(s) under which the document is submitted. The user of this document may voluntarily provide suggestions, comments or other feedback to Nokia in respect of the contents of this document ("Feedback").

2

Such Feedback may be used in Nokia products and related specifications or other documentation. Accordingly, if the user of this document gives Nokia Feedback on the contents of this document, Nokia may freely use, disclose, reproduce, license, distribute and otherwise commercialize the feedback in any Nokia product, technology, service, specification or other documentation. Nokia operates a policy of ongoing development. Nokia reserves the right to make changes and improvements to any of the products and/or services described in this document or withdraw this document at any time without prior notice. The contents of this document are provided "as is". Except as required by applicable law, no warranties of any kind, either express or implied, including, but not limited to, the implied warranties of merchantability and fitness for a particular purpose,

RA47053EN16AGLA0

are made in relation to the accuracy, reliability or contents of this document. NOKIA SHALL NOT BE RESPONSIBLE IN ANY EVENT FOR ERRORS IN THIS DOCUMENT or for any loss of data or income or any special, incidental, consequential, indirect or direct damages howsoever caused, that might arise from the use of this document or any contents of this document. This document and the product(s) it describes are protected by copyright according to the applicable laws. Nokia is a registered trademark of Nokia Corporation. Other product and company names mentioned herein may be trademarks or trade names of their respective owners.

© Nokia 2016

RA47053EN16AGLA0

2


Module Objectives • After completing this module, you will be able to:

• Describe how to detect interference by means of field measurements • Give an overview on interference and coverage issues via performance measurement counters

• Explain the impact of interference on peak throughput • Describe the relation of load and interference • Discuss the importance of interference analysis for the overall network performance

4

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

3


Index - Detecting interference using field measurements - Detecting interference and bad coverage from counters - DL Interference reduction mechanisms • eICIC • DL Interference shaping - Detecting overshooting cells • Timing advance counters - Impact of interference on peak throughput • idle/loaded other cell interference • PCI collision impact in TD-LTE - PCI Confusion - Impact of interference on LTE network performance – importance of physical RF optimization • Impact of network load • MIMO X-feeders • A trial example

5

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

4


Detecting Interference – Indicators • Three quantities • SINR • RSRQ

• RSRP - Which one should be used for drive test analysis?

6

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

5


Detecting Interference – SINR • SINR measurements can indicate interference areas, but it doesn’t necessarily see all interference sources: • Impacted by network load. Traffic in the neighboring cells will reduce serving cell SINR. • Depends on the measurement method (RS or SCH) and tool

• Depends on PCI planning (RS SINR) • Results can be misleading!

7

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

6


Detecting Interference – SINR • Example: SSS-CINR + RS CINR versus top-N RSRP

8

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

7


Detecting Interference – RSRQ • RSRQ depends on network load, including own cell load

9

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

8


Detecting Interference – RSRQ • RSRQ depends on serving and neighbor cell load • Fluctuates quickly • Hence difficult to interpret results

• Similar to Ec/N0 in 3G

10

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

9


Detecting Interference – RSRP • RSRP measurement with scanner is the most reliable way to detect areas with possible interference problems and bad dominance • Not impacted by network load • RSRP measurement appears to be consistent between UEs/scanners

• The number of PCIs in e.g. 5 dB power window is a useful indicator - A scanner with good dynamic range and PCI tracking capability needed

Bad dominance

11

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

10


SINR (worst case estimate) calculated from RSRP • Measured with PCTel MX scanner in TD-LTE network – RS-SINR, SCH-SINR, RSRP • Calculated SINR is worst case estimate for SINR (i.e.100% neighbor cell load). In TD-LTE it should be equal to SCH-SINR.

60

Calculated SINR goes very high in the locations where no neighbors are detected

40

20

Calculated SINR follows SCHSINR nicely in the most places 11:30:37 11:30:47 11:30:57 11:31:07 11:31:17 11:31:27 11:31:37 11:31:47 11:31:57 11:32:07 11:32:17 11:32:27 11:32:37 11:32:47 11:32:57 11:33:07 11:33:17 11:33:27 11:33:37 11:33:47 11:33:57 11:34:07 11:34:17 11:34:27 11:34:37 11:34:47 11:34:57 11:35:07 11:35:17 11:35:27 11:35:37 11:35:47 11:35:57 11:36:07 11:36:17 11:36:27 11:36:37 11:36:47 11:36:57 11:37:07 11:37:17 11:37:27 11:37:37 11:37:47 11:37:57 11:38:07 11:38:17 11:38:27 11:38:37 11:38:47 11:38:57 11:39:07 11:39:17 11:39:27 11:39:37 11:39:47 11:39:57 11:40:07 11:40:17 11:40:27 11:40:37 11:40:47 11:40:57 8:57:28

0

Average of LTE_Scan_RS_CINR_SortedBy_RSRP_0

-20

Average of LTE_Scan_SCH_CINR_SortedBy_RSRP_0 Average of Calc. SINR dB

Average of LTE_Scan_RSRP_SortedBy_RSRP_0 -40

Calculated SINR (worst case) = RSRP_serving/ (∑RSRP_others + Noise)

-60

-80

[Noise figure 9dB] -100

-120

12

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

11


Detecting Interference – Pilot pollution Counting PCIs less than XdB RSRP difference • From drive test with test terminal. • Serving PCI vs. Top N PCIs • Less than 5dB difference to the serving PCI can be considered a potential interferer.

• A common rule for antenna tilt optimization consideration: 3 or more PCIs inside 5dB window.

13

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

12


Detecting Interference – Summary •

Absolute SINR measurement values can’t be used as a reliable performance indicator. • Do not to blindly believe measured SINR values. • Relative SINR changes can be used as performance indicator, if the same measurement tool is used all the time.

• SINR measured from S-SCH and RS behaves differently depending on the interference situation (intra/inter eNodeB).

• Detailed SINR measurement methods of the terminals and scanners are not known. • The most robust and reliable measurement quantity seems to be RSRP

14

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

13


Index - Detecting interference using field measurements - Detecting interference and bad coverage from counters - DL Interference reduction mechanisms • eICIC • DL Interference shaping - Detecting overshooting cells • Timing advance counters - Impact of interference on peak throughput • idle/loaded other cell interference • PCI collision impact in TD-LTE - PCI Confusion - Impact of interference on LTE network performance – importance of physical RF optimization • Impact of network load • MIMO X-feeders • A trial example

15

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

14


Use Case  A BTS or a group of BTSs is having bad KPIs • Q: Is this because of bad coverage, UL/DL interference or both?

• How to analyze this from counters?

16

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

15


Bad Downlink vs Good Downlink • Example from Live Network 70000000

CQI = 14 60000000 Data Sum of M8010C036 UE Reported CQI Level 00 Sum of M8010C037 UE Reported CQI Level 01 Sum of M8010C038 UE Reported CQI Level 02 Sum of M8010C039 UE Reported CQI Level 03 Sum of M8010C040 UE Reported CQI Level 04 Sum of M8010C041 UE Reported CQI Level 05 Sum of M8010C042 UE Reported CQI Level 06 Sum of M8010C043 UE Reported CQI Level 07 Sum of M8010C044 UE Reported CQI Level 08 Sum of M8010C045 UE Reported CQI Level 09 Sum of M8010C046 UE Reported CQI Level 10 Sum of M8010C047 UE Reported CQI Level 11 Sum of M8010C048 UE Reported CQI Level 12 Sum of M8010C049 UE Reported CQI Level 13 Sum of M8010C050 UE Reported CQI Level 14 Sum of M8010C051 UE Reported CQI Level 15

50000000

Good DL coverage 40000000

30000000

Fairly bad DL coverage (or DL interference)

20000000

10000000

Check CQI offset from LTE_5432b E-UTRAN Average CQI Offset

0 100589

100953 BTS

17

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

16


Bad uplink vs good uplink 500000

450000

UE Power Headroom: -1dB <= PHR < +1dB.

400000

350000

300000

250000

200000

Fairly good UL coverage

150000

100000

50000

0 100589 BTS

18

Data Sum of M8005C054 UE Power Headroom for PUSCH Level 1 Sum of M8005C055 UE Power Headroom for PUSCH Level 2 Sum of M8005C056 UE Power Headroom for PUSCH Level 3 Open-loop UL Sum of M8005C057 UE Power Headroom for PC used PUSCH Level 4 Sum of M8005C058 UE Power Headroom for PUSCH Level 5 Sum of M8005C059 UE Power Headroom for PUSCH Level 6 Sum of M8005C060 UE Power Headroom for PUSCH Level 7 Sum of M8005C061 UE Power Headroom for PUSCH Level 8 Sum of M8005C062 UE Power Headroom for PUSCH Level 9 Fairly bad UL Sum of M8005C063 UE Power Headroom for coverage PUSCH Level 10 Sum of M8005C064 UE Power Headroom for PUSCH Level 11 Sum of M8005C065 UE Power Headroom for PUSCH Level 12 Sum of M8005C066 UE Power Headroom for PUSCH Level 13 Sum of M8005C067 UE Power Headroom for PUSCH Level 14 Sum of M8005C068 UE Power Headroom for PUSCH Level 15 100953 Sum of M8005C069 UE Power Headroom for PUSCH Level 16 Sum of M8005C070 UE Power Headroom for PUSCH Level 17 Sum of M8005C071 UE Power Headroom for © Nokia 2016 PUSCH Level 18

UE Power Headroom: -15dB <= PHR < -13dB.

RA47053EN16AGLA0

Open-loop UL power control used in this example. NOTE: UL PC settings will impact reported power headroom values

RA47053EN16AGLA0

17


PUSCH SINR, PUSCH RSSI Measurement • PUSCH RSSI and PUSCH SINR measurement can be used to detect UL coverage and UL interference problems • Interpretation of counter values depends on UL PC settings • Measurements are not correlated

SINR

UL CL PC upper SINR thrshld Ideally all samples are in this box

bad UL coverage

UL CL PC lower SINR thrshld UL interference RSSI

19

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

18


PUSCH SINR, PUSCH RSSI Measurement • Noise rise impacts SINR versus RSSI

Interference drives counter samples to this region

20

RA47053EN16AGLA0

© Nokia 2016

BTS noise figure assumed 2dB

RA47053EN16AGLA0

19


PUSCH SINR, PUSCH RSSI measurement • Impact of power control settings on PUSCH SINR UL tx pwr too high, generates interference

UL tx pwr too high, generates interference

21

RA47053EN16AGLA0

© Nokia 2016


RA47053EN16AGLA0

20


UL interference and noise rise •PUSCH RSSI and SINR are measured only when there is UL traffic

Methods for UL noise rise detection: Calculating UL noise rise from PUSCH RSSI and SINR counters.

- Definition, NR = (I+N)/N, N is thermal noise, I is interference - Assumption: PUSCH signal power = PUSCH RSSI - Calculation in linear: PUSCH_RSSI/PUSCH_SINR/N = PUSCH_RSSI/[PUSCH_RSSI / (I + N) ]/N = (I+N)/N = NR

22

RA47053EN16AGLA0

© Nokia 2016


RA47053EN16AGLA0

21


• This page intenionally left blank

23

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

22


Bad coverage analysis from HO counters M8015 - Neighbor cell HO measurements • •

Visualize HO cell pairs with poor HO performance In the case of poor performance to all neighbors in one directions  Coverage problems or unsuitable HO parameters (A3 offset)

Example: Cell pairs with HO SR<80% & HO Att>10 shown

24

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

23


Content • Detecting interference using field measurements - Detecting interference and bad coverage from counters - DL Interference reduction mechanisms • eICIC • DL Interference shaping - Detecting overshooting cells • Timing advance counters - Impact of interference on peak throughput • idle/loaded other cell interference • PCI collision impact in TD-LTE - PCI Confusion - Impact of interference on LTE network performance – importance of physical RF optimization • Impact of network load • MIMO X-feeders • A trial example

25

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

24


LTE1113 /LTE1496 eICIC Muted subframes known as ABS (Almost Blank)

During ABS-frames small cells can serve cell edge mobiles

SMALL cell transmission subframes MACRO cell transmission subframes

During all subframes good radio condition mobiles served

Ma

cro

-int e

rfe re

n ce si

g

l na

 with eICIC larger Range Extension values can be applied + better conditions for small cell edge camped mobiles

Range extension

pico-eNB macro-eNB Requires timing synchronization between the cells either,

UE

• •

X2-link with eICIC coordination

26

RA47053EN16AGLA0

LTE80 GPS synchronization, LTE891 Timing over Packet with phase synchronization

© Nokia 2016

The eICIC functionality will establish eICIC partnerships between a macro eNB and its subtending pico eNBs •Operator configures one macro cell partner on each pico. •A pico cell can only have one macro cell partner. •Following X2 setup, if a pico is configured with a macro cell partner the pico will initiate the eICIC partnership establishment by sending an X2: LOAD INFORMATION message to the macro. Macro Cell (LTE1113) does not schedule UEs during ABS based on Cell Load information from Micro maximizes performance by adjusting:

– cell individual offsets (Range Expansion) – ABS Micro Cell (LTE1496) schedules Cell Range Extension UEs during ABS sends Cell Load information to Macro via X2 applies information from Macro to maximize performance

RA47053EN16AGLA0

25


LTE1113 /LTE1496 eICIC eICIC benefit in a nutshell •eICIC principle is sacrifice of macro cell resources for higher small cell user offload and better performance Significant increase in the # configured ABSs does not improve throughput too much

-Therefore main benefit of the feature can be observed in the small cells, especially at the edge

Macro throughput

End user throughput

-Please keep in mind that network level gain is a trade-off between macro and small cell capacity

Small cell throughput

ABS, CRE

With eICIC #ABS Without eICIC

27

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

26


LTE1113 /LTE1496 eICIC

eICIC area capacity improvement

Cell-edge user throughput improvement

Up to

20% 60%

Small cell users increase in the CRE region

Up to

28

27%

70% RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

27


LTE1113 /LTE1496 eICIC •Dynamic eICIC increases the user throughput especially on cell edge and allows a larger offloading of users from macro cell by cell range expansion of small cells

Dynamic ABS (RL70)

Dynamic ABS + CIO

Area cell capacity

31%

23%

User throughput

59%

47%

Cell-edge user throughput

62%

21%

SINR in small cell

+2.5 dB

+1.3 dB

Interference in small cell

-3.3 dB

-3.7 dB

29

RA47053EN16AGLA0

© Nokia 2016

The dynamic adaptation algorithms for ABS and CIO reduce the maximum gains of the feature (especially on cell edge). Reasons is the speed of the adaptation algorithms and that the outcome can only be optimized to one considered KPI or be a compromise for all

RA47053EN16AGLA0

28



30

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

29


LTE1800: Downlink interference shaping Introduction: Fractional load and CQIs frequency

Traffic pattern in the cell – the load is fractional, but now allocations are placed only within preferred area

This profile of traffic creates spatially localized interference to the neighboring cell

Preferred allocation area

Squeezing the traffic to a section of the bandwidth we create distinct “busy” and “clean” areas

Area excluded from allocations

time

31

RA47053EN16AGLA0

Now the subband CQIs will be able to well depict the interference pattern generated by neighboring cell

© Nokia 2016

RA47053EN16AGLA0

30


LTE1800: Downlink interference shapingactDlIntShaping Activate downlink interference shaping LNBTS; False, True

Fractional loaded Cell to Highly loaded Cell The simplest scenario where this feature is expected to provide a gain is a fractionally loaded cell with a neighboring highly/fully loaded cell.

Heavy load

CQI

Fractional load

f

Cell1 Cell2

PRB utilization

f

32

RA47053EN16AGLA0

Utilization of PRB resources in cell 1 becomes visible in the frequency selective channel quality reporting of a cell edge UE in cell 2.

© Nokia 2016

actDlIntShaping activates the feature within the eNB actdLIsh activates feature on Cell level

RA47053EN16AGLA0

31


Benefits and gains Benefits and gains  The feature provides significant cell edge throughput and fairness gains in networks with heterogeneous cell load.  No gain for full load or balanced load scenarios  In comparison with other proposals for multi-cell coordination and centralized scheduling it has the following benefits:  Compatible with 3GPP release 8 onwards

 No proprietary X2 message extensions needed  No strict latency requirements over X2  No additional network element for centralized scheduling

33

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

32



34

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

33


Overshooting cells • Cells with too large or largely distributed dominance area. • Will cause increased interference to other cells • Can collect excessive amount of traffic. • How to detect overshooting cells? • Drive tests

• Analyzing HO performance and neighbor cell measurements from drive test logs. • Counters • Cell pair HO analysis with inter site distance information. Indicating HOs to cells with long inter site distance. • Timing Advance trace • Cell Timing Advance trace analyzes to find long distance users.

35

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

34


Overshooting cells analysis from HO counters Optimizer - Visualize adjacencies and HO KPIs •

36

To find long distance cells pairs with HO attemps -> overshooting cell detection.

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

35


Overshooting cells analysis from HO counters Other tools to analyze Neighbor cell HO counters •

M8015 - Neighbor cell HO measurements

•

Visualize long distance HO cell pairs (e.g. >10km)

37

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

36



38

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

37


LTE1697 Timing Advance Histogram PM-counter Histogram configuration • Based on an operator-configured parameter (LNCEL:expectedCellSize ) eNB is applying a predefined set of PM-counter bins (TA-histogram set) for mapping the measured instantaneous TA-value into a corresponding bin -

Each of the histogram PM-counters represents a bin for a certain distance range #

The value of the parameter is always stored in additional counter TIMING_ADV_SET_INDEX (M8029C0) Counter bins

• Configuration of the 'expected cell size' can be performed Manually by the operator or

-

Automatically during the system upgrade when doing the database conversion  Thereby a simple configuration rule is applied, which picks a proper expectedCellSize depending on the configured PRACH cyclic shift ( prachCS) and high-speed flag (hsFlag).

18000

2.1 km range

16000

Cell range

5 km range

14000 Range [m]

-

10 km range

12000

15 km range (def)

10000

8000 6000 4000

2000 0

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 Counter bin

39

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

38


Index • Detecting interference using field measurements - Detecting interference and bad coverage from counters - DL Interference reduction mechanisms • eICIC • DL Interference shaping - Detecting overshooting cells • Timing advance counters - Impact of interference on peak throughput • idle/loaded other cell interference • PCI collision impact in TD-LTE - PCI Confusion - Impact of interference on LTE network performance – importance of physical RF optimization • Impact of network load • MIMO X-feeders • A trial example

40

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

39


RF Peak Throughput under Neighbor Cell Interference • Measuring peak MIMO dual-stream throughput in the field can be tricky because of interference

• An idle cell produces common channel + RS interference to impact peak throughput  need to find good interference-free measurement spot. Inter-site cell border, non-frame synchronized cells

Intra-site cell border, framesynchronized cells

41

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

40


Impact on Peak Tput from Idle Neighbor Cell Interference

• Measurement example #1, Samsung terminal, 20MHz. Inter-site and intra-site neighbor are unloaded (no PDSCH traffic) PHY tput, CINR, RSRP

All neighbor cells attenuated 50dB

120

-50

-60

Inter-site interference, adjacent site cell about 5 dB weaker RSRP than serving cell

Intra-site interference, adjacent cell about 5 dB weaker RSRP than serving cell

-70

80 -80

60

-90

RSRP dBm

PHY tput Megabits/sec, CINR dB

100

Data Average of Phy DL TP(Mbps) Average of SCell-CINR Average of SCell-RSRP

-100 40 -110 20 -120

06/11/2010 09:53:02.801

0 06/11/2010 09:54:45.317

-130 06/11/2010 09:56:29.840

Intra-site neighbor frame-synced, no RS interference

06/11/2010 09:58:14.347

Time

All neighbor cells attenuated 50dB` 42

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

41


Impact on Peak Tput from 100% Loaded Neighbor cell • Measurement example #2, Samsung terminal, 20MHz. Unloaded and 100% loaded PHY tput, CINR, RSRP

inter-site neighbor

Neighbor site cell attenuated 50dB -50

120 UDP download 100Mbps

neighboring site cell in idle mode

80

-60

Neighbor site cell about 6 dB weaker

-70

-80 60 -90 40

RSRP dBm

PHY tput Megabits/sec, CINR dB

100

Data Average of Phy DL TP(Mbps) Average of SCell-CINR Average of SCell-RSRP

-100 20 -110

06/11/2010 10:10:27.444

06/11/2010 10:09:36.436

06/11/2010 10:08:45.430

06/11/2010 10:07:54.924

06/11/2010 10:07:04.918

Neighbor site cell about 1 dB weaker

06/11/2010 10:06:14.411

06/11/2010 10:05:23.905

06/11/2010 10:04:33.398

-20

06/11/2010 10:03:42.892

06/11/2010 10:02:52.385

06/11/2010 10:02:00.821

0

-120

Typical SINR= 15-17 dB at inter-site cell border, unloaded neighbor.

-130

Time

43

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

42


TD-LTE: Impact of Idle Mode Interference on Tput • UE FTP downloading in the middle of two sectors of the same site, RSRP from both cells ~ -70dBm • First the second cell is off (rebooting), then comes on-air but no traffic carried (only common channels

34Mbps vs 15Mbps

Neighbor cell switched on

SINR

tput

Serv RSRP

transmitted)

44

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

43



45

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

44


Impact of PCImod3 Collision on Tput, TD-LTE •

Case: UE at the border of two cells who have the same PCImod3, RSRP from both cells = 67dBm in both measurement cases (only PCI changed)

•

Nokia 7210 TD dongle, 2.6GHz, 10MHz bandwidth 16 14

tput, Mbps

12 10

no PCImod3 collision

8

PCImod3 collision

6 4 2 0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53

seconds

46

RA47053EN16AGLA0

© Nokia 2016

PCIs in the two measurement cases were 9 /10 and 10 / 13.

RA47053EN16AGLA0

45


PCImod3 Collision Impact, 2.3GHz@20MHz, Qualcomm TD-LTE Dongle Example PCI= 88/90 RSRP = -97dBm SINR = 12dB

PCI= 87/90 (mod3 collision) RSRP=-101dBm SINR=2dB

Tput = ~15Mbps

Tput = ~21Mbps

47

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

46


Tools for Parameter Planning - NetAct Optimizer • PCI planning • PRACH planning

• UL DM RS sequence planning is a future feature candidate - Atoll • Automatic PCI planning supported - Asset 7

• PCI planning - Alpha (Nokia-internal tool) • PCI planning

• UL DM RS planning - MUSA (Nokia internal) - post processing - Daisy (Nokia-internal tool)

• PCI planning

48

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

47


Index • Detecting interference using field measurements - Detecting interference and bad coverage from counters - DL Interference reduction mechanisms • eICIC • DL Interference shaping - Detecting overshooting cells - Timing advance counters - Impact of interference on peak throughput • idle/loaded other cell interference • PCI collision impact in TD-LTE - PCI Confusion - Impact of interference on LTE network performance – importance of physical RF optimization • Impact of network load • MIMO X-feeders • A trial example

49

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

48


LTE1685 neighbor Relation Robustness LTE1685-C: Warning for PCI Confusion • Starting RL50/RL35TD: eNB reports a Warning with fault ‘neighbor cell ambiguity detected’ towards NetAct and BTS Site Manager. PCI collision is solved by manual PCI re-planning.

eNB

Cell A PCI-1, Freq1

neighbor cell ambiguity detected!

PCI Collision! • CGI of the cell which detected the collision • Cell A • PCI and frequency which are not unique • PCI1, Freq1 • List of CGIs for which the collision was detected • CGIs of CellA and NbCell1

ReportCGI: ECGI-1 PCI-1, Freq1

ECGI1 PCI-1, Freq1

NbeNB1

50

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

49


LTE1685 neighbor Relation Robustness LTE1685-C: Warning for PCI Confusion • Starting RL50/RL35TD - neighbor Relation Status (nrStatus) can either be ‘available’ or ‘unavailable’, which determines if the CGI exists or not in eNB (LNADJL) • NR status is evaluated in these events: - Creation of LNADJL object (via X2, reportCGI or plan file) - Modification of LNADJL parameters phyCellId or fDlEarfcn (via X2, reportCGI or plan file) - Deletion of LNADJL (X2, plan file) - Creation of LNREL object (measurement report, plan file) - Creation and deletion of LNCEL object (plan file) - eNB restart (as long as LNCEL creation causes BTS restart)

51

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

50


LTE1685 neighbor Relation Robustness LTE1685-C: Warning for PCI Confusion • Starting RL70/RL55TD: PCI confusion is detected via multiple CGI resolutions and warning will be raised in NetAct and BTSSM. Re-assignment of PCI has to be corrected manually, but BTS notification makes it easier to detect and resolve the conflict.

ENB neighbor List: • LNADJL-1: ECGI-1  LNREL-1 (invalid) (available)

eNB


• LNADJL-2: ECGI-2  LNREL-2 (available) Cell A


ECGI1 PCI-1, Freq1


ECGI2 PCI-1, Freq1

NbeNB2

PCI Confusion! • CGI of the cell which detected the PCI confusion • Cell A • PCI and frequency which are not unique in the eNB • PCI1, Freq1 • List of CGIs of neighbor cells for which the confusion was detected. • CGIs of NbCell1 and NbCell2

NbeNB1

52

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

51


LTE1685 neighbor Relation Robustness Alarm Fault name

PCI confusion detected

Fault ID and name

6279: EFaultId_PciConfusionDetectedAl

Reported alarms

7655 CELL NOTIFICATION

Unit status

Working

LED display

Blinking yellow

Meaning

Based on CGI measurements it has been detected that two or more cells operating on the same carrier frequency use the same PCI Alarm contains: 1. CGI of the cell which detected the PCI confusion. 2. The CGI of the NR which is switched from available to invalid 3. The CGI of the NR which is switched from invalid to available

Unit actions after fault detect Detection method start/transient

None The neighbor relation states 'invalid' and 'available' are exchanged between two LNREL instances of the same cell (pointing to the same PCI and frequency) after reception of a CGI measurement report.

Detection method cancel None Effect

Possibly increased handover failure rate for the affected cell.

Instructions

Check for E-UTRAN cells that operate on the same frequency layer and use the same PCI and that are in close vicinity. Re-assign PCIs for the discovered cells.

53

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

52



54

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

53


RF Optimization • Basic physical RF optimization is very important (of course..) • Clear cell dominance areas, minimize cell overlapping • Avoid sites shooting over large areas with other cells

• “Can’t fix bad RF by tuning parameters” • Antenna tilting and antenna placement has big impact on other cell interference!! • What is the impact on network performance?

55

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

54


Impact of DL load, 0% vs. 70% DL load •The same drive test route driven twice, with the same UE setup - LTE819: DL Inter-cell Interference Generation to generate load

- 0% load versus 70% DL load - Compare distribution of throughput and SINR, the same drive test route twice with and without load - 20MHz OL-MIMO, FTP download, 1UE inside the car, Samsung BT-3710, UE-internal antennas

- average throughput is 58% better without interference - Selection of drive test route strongly affects result, here only results for one drive test route Empirical CDF

Empirical CDF

1

1

0.9

0.8

0.7

0.7

0.6

0.6

0.5

Mean = 36Mbps

0.4

CDF

CDF

70% OCNG 0% OCNG

0.9 70% OCNG 0% OCNG

0.8

0.5 0.4

0.3

0.3

Mean = 57Mbps 0.2

0.2

0.1

0.1

0

56

0

10

20

30

40 50 60 throughput [Mbps]

70

80

RA47053EN16AGLA0

90

100

0 -5

0

5

10 15 SINR [dB]

20

25

30

© Nokia 2016

RA47053EN16AGLA0

55



57

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

56


MIMO X-Feeder Assuming:

•

ANTL-1 and ANTL-7 are defined active for sector 1

•

ANTL-3 and ANTL-9 are defined active for sector 2

•

Then the configuration in the upper picture is correct

•

The configuration in the lower picture is incorrect and results in sectors overlapping with each other  bad throughput due to interference

58

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

57


MIMO x-feeder, Example 1 Scanner Measurement

Sectors 241 and 242 equally strong in area where 242 should dominate

242

241

59

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

58


MIMO X-Feeder Example 2 Scanner Measurement

22 21 Site (PCIs=21,22)

PCIs

60

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

59


MIMO X-Feeder Example 2, Scanner measurement Corrected Feeders

21

Site (PCIs=21,22)

PCIs

61

22

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

60



62

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

61


Antenna tilt tuning example (1/3) A reference cluster • Drive test measurements

• SINR before and after tilt tuning.

63

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

62



• CQI before and after tilt tuning.

64

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

63



• HO attempts before and after tilt tuning.

65

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

64


Summary • Building good dominance is essential for network performance – also in LTE !!! • “Can’t fix bad RF with parameters…” • …except by fixing missing neighbors

66

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

65


NokiaEDU

RA47053EN16AGLA0

© Nokia 2016

RA47053EN16AGLA0

66

RA47053EN16AGLA0 Physical RF Optimisation

Recommend Documents