Part7 GSM Interference Analysis and Optimization

GSM Interference Analysis and Optimization NPO Refresher Course July, 1st to 3rd 2010 Vodafone MS – RoB

Jignesh Parmar [email protected] Nokia Siemens Networks National NPO, Ahmedabad, India NSN Internal Document 1 © Nokia Siemens Networks

GSM Interference Analysis/JP/NNPO/ 1st to 3rd July 2010 @ VF RoB

What is interference? • Unwanted signal to the receiver – Internal System Interference Resulting from tight frequency re-use Multipath Hardware Fault Spurious transmission Repeater malfunction – External Interference Radar Other Communication system Jammer

NSN Internal Document 2 © Nokia Siemens Networks


Classification of Interference Sources • Natural Noise – Atmosphere noise – Galaxy noise – Solar noise (quiet period) • Man-made Noise – Interference of ignition systems of vehicles or other engines – Interference of electronic communication system – Interference of power line – Interference of scientific research, medical and household appliances



Main Interference Sources Affecting Mobile Communication • Internal interference – TRX fault: if the performance of TRX is reduced due to manufacture –

– – –

cause or application, which will cause self-excitation of TRX amplification circuit resulting in interference. Spurious emission and inter-modulation: if out-band spurious emission index of BTS TRX or amplifier exceeds the limit, or isolation between TX and RX antennas is not enough, all these will form interference on the receiving channel. Inter-modulation may also occur in passive equipment such as passive antenna and feeder. Improper frequency planning Co-frequency interference Adjacent frequency interference



Main Interference Sources Affecting Mobile Communication • Repeater interference – If the installation of repeater is non standard, causing insufficient isolation between the donor antenna and the serving antenna, so selfexcitation is caused. This affects normal working of BTS that the repeater belongs to. – For the repeater adopting wideband non-linear amplifier, the intermodulation index far exceeds the requirement of the protocol. If the power is comparatively high, the inter-modulation component will be large, and this will cause interference to the BTS around. – If the repeater is malfunctioning due to some fault it will go in oscillation mode results in interference.



Main Interference Sources Affecting Mobile Communication • External interference – Interference of other communication equipment with high power. – Radar station: From 70s~80s of the 20th century, the frequency used by the decimeter wave radar was similar to that of GSM, and its transmitting power was very high, which generally reached tens and hundreds of kilowatts, so the out-band spurious emission is comparatively large. Thus, it easily causes interference to the BTS. – Analog BTS: The frequency band used by the analog mobile BTS overlaps with the GSM frequency band in certain segment. – Communication equipment at same frequency band: As the types of communication equipment are so many, some manufacturers maybe adopt the frequency band but doesn’t comply with the current communication standards. As the GSM frequency band is occupied by the equipments, interference will be caused within the GSM system coverage area.



Impact of Interference • When there is interference in the network, the subscribers usually encounter the following phenomenon:

– The subscriber cannot hear the voice , and the background noise is

too loud. – When fixed telephone subscriber calls MS subscriber, or MS subscriber calls fixed telephone subscriber, call drop occurs after “du, du , du” is heard. – The conversation cannot be carried on smoothly, and call drop often happens. – When interference exists in the network, from the result of traffic statistic, there are some characteristic as following. There are Level 4~Level 5 interference band in TCH measurement function ,

and the measurement value is more than 1. Congestion rate is comparatively high. The call drop rate is higher than other cells. The handover success rate is low. Through Drive Test, it is found that: • It is difficult to handover . • The Rx level is high, but the quality is bad. Through tracing the Abis interface signaling with signaling analyzer, it is found that the bit error rate is higher than other cells.



Carrier to Interference Ratio Interference and reuse/reuse distance

Cellular network

3

• partial overlap of cells

4 2

• only a few frequencies per cell • frequency re-use distance

1

5

7

6

3 4

re-use distance

2 1 7

5 6



Receiver Sensitivity Level impacts Interference & Noise Level [dBm ]

20

0 20

-104 dBm min. sensitivity GSM (BTS)


normal BTS


micro BTS

pico BTS

-40

-60

-80

C/N

-100 C/N


C/N


The necessary CNR increases Bandwidth

Carrier to Interference Ratio Requirements Following GSM 05.05:

Reference interference ratio for all BTS and MS types: • For co-channel interference:

C/Ic =

9 dB

• For (first) adjacent channel interference:

C/Ia1 = - 9 dB

• For (second) adjacent channel interference:

C/Ia2 = - 41 dB

• For (third) adjacent channel interference:

C/Ia3 = -49 dB

At these values, the so called reference interference performance in terms of (maximum) frame erasure rate, bit error rate or residual bit error rate must be met for the different type of channels in different specified propagation conditions (TU3 no FH, TU3 ideal FH, TU50 no FH, TU50 ideal FH, RA250 no FH).



Carrier to Interference Ratio Requirements Co-channel interference: 9 dB Quality (Co Ch related on reuse distance; cs-service)

20

0

V LE Rx

20

Rx LE V

Leve l [dBm ]

-40

+ 9 dB

CIR CoCH

-60

-80

Neighbour cell -100 Distance

Server cell NSN Internal Document 11 © Nokia Siemens Networks

Distance


Adjacent Channel Interference Adjacent channel related quality (CS)

Quality limit of serving cell

-20

ur ed ea s

Rx LE V

f

2

f

m

1

V LE Rx

0

Rx LE V

f

at

2

f1

Level [dBm] 20

- 9 dB CIR AdCH

-40

Adjacent channel Interference BTS and MS

CIR CoCH + 9 dB 60 80 100

Distance

Serving cell Distance



Neighbour cell

Adjacent Channel Interference Adjacent Ch. Interference; cs -service PWR Level [dBm ]

CIR AdCH

10

+ 9 dB

0

21 dB 20

-40

-60

-80

-100

200 kHz F1 NSN Internal Document 13 © Nokia Siemens Networks

200 kHz F2

Frequency range


30 dB

Carrier to Interference Ratio : Exercise • Carrier = - 79 dBm • Interference = - 70 dBm • Carrier to Interference Ratio (C/I) = - 9 dB • Interfering signal is 9 dB higher than Wanted Signal • Carrier = - 80 dBm • Interference = - 104 dBm • Carrier to Interference Ratio (C/I) = 24 dB • Interfering signal is 24 dB lower than Wanted Signal



Adjacent Channel Interference 200 KHz

9 dB

Surface below black curve from fC - 100 KHz to fC + 100 KHz -> serving power

30 dB

Surface below red curve from fC - 100 KHz to fC + 100 KHz -> interfering adjacent power

fC - 100 KHz


fC + 100 KHz


Must be at least 9dB smaller than surface below black curve

Locating Interference



Discovering Interference via OMC Traffic Measurement Discovering potential interference via traffic statistics result • Check the “Average TCH busy time (second)” in TCH measurement function of each cell, the reason is that this index can show the TCH mean occupied time (s”), which is usually called “TCH mean holding time” in the BSC of other manufacturer”, within the measurement period. If it is found that the Average TCH busy time (second) of certain cell is comparatively short (such as less than 10s), then maybe there is strong interference in the cell, causing that handover/call drop happens due to bad quality after TCH channel occupied to MS. • Certainly, if hardware fault occurs in certain TRX (non-BCCH or non-SDCCH carrier) of a cell, the case mentioned above will also appear. NSN Internal Document 17 © Nokia Siemens Networks


Quality Interference analysis DL /UL Rx Quality x Rx Level

Same level – quality distribution for both UL and DL HW Problem: Bad Quality

Good Quality

for all Rx Levels Coverage Problem: Bad quality and Low Rx Level

HW Problem All samples below 100dBm CL10 <-100dBm

High Rv Level Interference Problem: Bad quality and High Rx Level NWD report 204 model



Discovering potential interference via UL interference Stats from OMC BTS area level

BTS level ‘Hit Lists’ over given period/area

200

=============================================================================== = UPLINK INTERFERENCE STATISITICS =============================================================================== The UL interference is measured based on the levels in idle FTCH. The measurement reports the average number of idle FTCH in different bands which are defined by boundaries. Boundaries are set as BTS parameters. -----------------band 1 -----------------band 2 -----------------band 3 -----------------band 4 -----------------band 5 ------------------

boundary B00 (value fixed: -110dBm) boundary B01 (value eg.:

-105dBm, critical)

boundary B02 (value e.g.:

-100dBm)

boundary B03 (value e.g.:

190 UL interference, 24-hour/10-day breakdown BSC2UPS1:SANKUD010:SANKUD2020 boundaries: -110,-105,-100,-95,-90,-47 (dBm)

-95dBm)

boundary B04 (value e.g.: -90dBm) boundary B05 (value fixed: -47dBm)

The boundaries used in the selected BTS area are reported below: Hr -00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23

THRS BOUNDARY1 (BO0) - fixed BTS COUNT ---------------------------------------- ---------110.0 dBm 333 THRS BOUNDARY1 (BO1) BTS COUNT ---------------------------------------- ---------105.0 dBm 327 -95.0 dBm 5 -90.0 dBm 1 THRS BOUNDARY2 (BO2) BTS COUNT ---------------------------------------- ---------100.0 dBm 327 -83.0 dBm 5 -88.0 dBm 1 THRS BOUNDARY3 (BO3) BTS COUNT ---------------------------------------- ---------95.0 dBm 327 -71.0 dBm 5 -86.0 dBm 1 THRS BOUNDARY4 (BO4) BTS COUNT ---------------------------------------- ---------90.0 dBm 327 -59.0 dBm 5 -80.0 dBm 1

Breakdowns for each cell having UL interference

THRS BOUNDARY5 (BO5) - fixed BTS COUNT ---------------------------------------- ---------47.0 dBm 333

Print out, analyze and pay attention to: 1) What is the typical patterns?Sort pages. 2)Colour on the map the cells of similar pattern with same colour. 3) See if the affected cells form an clear are or point the same direction. 4) Use available methods to find out what could be the source (local engineersmay know best the candidates)

The percentual distribution of idle FTCH in the selected area:

Day --sun sat fri thu wed tue mon sun sat

MMDD ---0211 0210 0209 0208 0207 0206 0205 0204 0203

idle FTCH idle FTCH idle FTCH idle FTCH idle FTCH in band 1 in band 2 in band 3 in band 4 in band 5 (%) (%) (%) (%) (%) --------- --------- --------- --------- --------97.4 1.2 0.7 0.3 0.4 97.1 1.4 0.7 0.4 0.5 96.3 1.8 0.9 0.5 0.5 96.3 1.7 1.1 0.3 0.6 96.4 1.7 1.1 0.3 0.5 96.4 1.7 1.0 0.4 0.5 96.5 1.6 1.0 0.4 0.5 97.6 1.1 0.7 0.3 0.4 97.2 1.2 0.8 0.4 0.4

In interference free area all is in band 1. The more weigh is towards band 5 the stronger the interference NSN Internal Document 19 © Nokia Siemens Networks

UL interference, 24-hour/10-day breakdown thu fri sat sun mon tue wed thu fri sat sun BSC2UPS1:SANKUD010:SANKUD2020 27 28 29 30 31 01 02 03 04 05 06 boundaries: -110,-105,-100,-95,-90,-47 (dBm) Hr AUG AUG AUG AUG AUG SEP SEP SEP SEP SEP SEP -- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---thu fri sat sun mon tue wed thu fri sat sun 00 0 0 0 0 0 0 0 0 0 0 0 27 28 29 30 31 01 02 03 04 05 06 01 0 0 0 0 0 0 0 0 0 0 0 AUG AUG AUG AUG AUG SEP SEP SEP SEP SEP SEP 02 0 0 0 0 0 0 0 0 0 0 0 ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---03 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 04 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 05 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 06 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 07 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 08 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 09 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 10 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 12 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 13 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 14 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 16 0 0 0 0 0 0 0 18 0 0 0 0 0 0 0 0 0 0 0 0 0 0 17 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 18 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 19 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 18 0 0 0 20 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 21 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 22 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 23 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

• •

Boundaries can be set as cell parameters Note: Mast Head Amplifier for 1800 and 1900 networks brings constant 12dB gain. MHA for GSM has adjustable gain.


Discovering interference via handover data • When certain cell initiates handover, if the average receiving

quality (uplink) is ≥ 4 (this is true when there is no frequency hopping, if there is, it should be ≥5), and the mean receiving level is ≥25(-85dBm), then it is possibly caused by uplink interference. • When certain cell initiates handover, if the times of receiving quality level above 5 is more than that below 4, then there may be uplink interference as well. • If the times of certain cell attempts to initiate handover (uplink/downlink quality) is more than 10% of total handover attempt times, then there may be interference in the cell. The two indices are all related to quality handover threshold and interference handover threshold within the cell parameters.



Discovering potential interference via call drop index • If the call drop times of certain cell is rather higher the other with same traffic load, and the main cause of call drop is owning to connection fault, then it is possibly caused by interference. • If the average receiving level during call drop is comparatively high (≥25), while the average receiving quality level is ≥6, then the cell should be listed into the interference source.



Discovering potential interference via interference band • BTS will utilize an idle TS in a frame to scan the uplink frequencies of the frequencies used by TRX, and then make measurement to the level 5 interference band. The default setting of interference bands in BSC of Huawei is as follows: 110, 105, 98, 90, 87 and 85 (unit: -dBm) • Compared with other measurement indices, the measurement index of interference band can reflect the cell interference situation more directly, but it can only reflect whether there is interference in the uplink. • If the values of interference band 4 and interference band 5 are comparatively large (≥1), then, there may be co-frequency interference in the cell. If the measurement values mainly distribute in interference band 1 and interference band 2, then the possibility of interference will be small. However, if there is comparatively high value in band 3, then attention should be paid to this. NSN Internal Document 22 © Nokia Siemens Networks


OMC Alarm and Subscriber Complaint • Subscriber complaint is also important clew for finding the potential interference. Information which should be collected from user complaint includes MS number, MS model, called number and fault phenomenon of calling side and fault phenomenon of called side and particular fault location, etc. If the alarm information is more detailed, it will be more easily to find out the network problems. • When there is interference in the network, the direct feelings of the subscriber may be: heavy noise, both parties or either party cannot hear each other clearly, call drops and call cannot be put through, etc. Thus, when many subscribers within the same area complain the same problem, then work should be done to check whether there is interference in the area. NSN Internal Document 23 © Nokia Siemens Networks


Discovering Interference via Drive Test There are two Drive Test methods available: idle mode test and dedicated mode test. • Under the idle mode, the test equipment can measure the signal level of both serving cell and adjacent cell. In addition, the equipment can also perform frequency scanning test to the specified frequency or frequency band. • Under dedicated mode test, the test equipment can measure the signal levels, receiving qualities, power control registrations and time advance, etc. of both the serving cell and adjacent cells. When high level (≥30) and low quality (Rx_Qual≥6) remain in certain section, it can be concluded that interference exists in the section. Further, part test equipment can directly display the frame elimination rate (FER). Generally when the FER ≥25%, subscribers can feel the discontinuous voice, that is to say, interference exists in this section of highway. NSN Internal Document 24 © Nokia Siemens Networks


Finding the Source of Interference



Recommended Procedures for Location and Clearance of Interference • Determine Interference Cell according to Key Performance • • • • •

Index (KPI) Check Alarm of OMC Frequency Planning Check Check Parameter Setting of Cell Drive Test Interference Clearance



Brief introduction to Spectrum Analyzer • Spectrum Analyzer is a broadband signal receiver with high performance which can display • • • • • • •

the spectrum of the receiving signals. The receiving signal resolution bandwidth (RBW): namely the minimum signal bandwidth the Spectrum Analyzer can recognize. The smaller the parameter is, the higher the receiving sensitivity of instrument. Input frequency: the frequency range the Spectrum Analyzer can receive. Sensitivity: generally the minimum receiving level with 1Hz signal bandwidth is defined as receiving sensitivity of the Spectrum Analyzer. The receiving sensitivity of HP85 series Spectrum Analyzer can be below -142 dBm. Video Filter Bandwidth (VBW): it refers to the bandwidth of the intermediate filter after Frequency mixing of the Spectrum Analyzer. The smaller the bandwidth is, the smoother the curve is; Central frequency (F0): it refers to the central frequency of the spectrum that the Spectrum Analyzer can test; Bandwidth (SPAN): it refers to the spectrum span the Spectrum Analyzer can test; Input signal attenuation (ATT): when there is large signal input, it is required to make attenuation on signal properly. The Spectrum Analyzer itself may produce large number of inter-modulation components without attenuation. So it will influence the veracity of the testing result.



Directional Antenna • Directional antenna is used for searching interference sources. The stronger the directionality of antenna is, the higher the antenna gain is. And the ability to search will become better. So the logarithm-period antenna with broad frequency band is the best choice. This kind of antenna has broad frequency band, high antenna gain and strong directionality.



The way to Test Internal Interference •

Set the Spectrum Analyzer to proper state:

– For 900M BTS: f0=902MHz,SPAN=30MHz,ATT=0,RBW=30kHz,VBW=30kHz; – For 1800M BTS: f0=1715MHz,SPAN=10MHz,ATT=0,RBW=30kHz,VBW=30kHz. •

Screw out the connector of output port of CDU divider, then connect the output signal of the divider to the Spectrum Analyzer to carry out a test. If the fractional frequency spectrum level is less than –80dBm, it shows that there is no internal interference; if more than –80dBm, it shows that CDU or TRX inside Base Station are under interference or self-excitation.

•

If internal interference exists, further make sure that it belongs to CDU or TRX. At first confirm TRX carrier board, cut down the cable via which TRX is connected to divider, and use the Spectrum Analyzer to test the main or diversity connector of TRX. If the fractional frequency spectrum level is less than –80dBm, it shows that TRX is normal, otherwise it is required to change carrier board.

•

The three steps above aim at interference measure for the uplink frequency band. If there is suspect that interference exists in the downlink frequency band, please follow steps below.

•

Check interference of transmission band. First, set the Spectrum Analyzer in transmission frequency band of the BTS. Due to the large ouput power of BTS, attenuation should be made on the input signal. Generally ATT is set as 40dB, then the tx_test signal of CDU should be imported to the Spectrum Analyzer to be observed to make sure weather interference signal is generated.



The way to Test External Interference When we are sure that interference is caused by the external cause, first we should confirm the location of interference source and the spectrum distribution state.

• First, set the Spectrum Analyzer to proper state. • Choose output port of divider of cell under interference. • Screw out the selected connector, then use Coaxial Cable to import the output signal of divider to the Spectrum Analyzer;

• View the spectrum distribution state of the Spectrum Analyzer, and find out the abnormal interference signal. The way to calculate the level of interference signal is as follows:

– Antenna port interference level = interference level tested by the Spectrum Analyzer – 15dB Tower Top Amplifier Gain + 3dB cable loss – 7dB divider gain.

– The maximum interference level at antenna port without influence on system = 108dBm sensitivity – 9dB co-channel interference protection= -117dBm



The way to Search External Interference Sources • In the cell under interference, select a test point without building • •

• •

obstruction. Set the Spectrum Analyzer, and connect the directional antenna. If there is rotatable platform, the antenna can be placed on it, and make the wave beam of the antenna point to the front, and the antenna with vertical polarization should be placed vertically; if there is no interferent signal, one can raise the antenna over head with hands. Rotate the antenna slowly, and at the same time view the change of signal of the Spectrum Analyzer. Once there exists abnormal signal, fix the orientation of the antenna immediately and change uptilt of the antenna to make the receiving signal to the strongest. Analyze the signal spectrum distribution carefully, and confirm that it is interference signal, record the signal strength and record the azimuth and downtilt of antenna wave beam. Find new test point along with the direction of antenna wave beam, then return to step 2 to carry out a test till interference source is found.



Antenna Performance Degradation Fault description: There are 5 BTSs for a certain network in a county configured as S4/4/4 and 6/6/6, the BTS type includes BTS20 and BTS30. The interference band 5 in TCH measurement function of some cells is over 15, and there is no alarm information in OMC



Antenna Performance Degradation Fault location process

• Register the statistics task of interference band of 24 hours for the cell

•

• • •

with problems, it is found that interference band 5 mainly occurs in daylight, and in the small hours near middle night, the interference band value is almost 0. After opening the idle BURST of all BTSs and transmitting it in the early morning, it is found that the interference band occurs. It disappears after transmitting is stopped. It can be judged from this phenomenon that the interference comes from internal network and has nothing to do with other equipment. No frequency in the network and data are modified before the interference occurs, accordingly, the interference is irrelevant with the frequency planning. It can be seen from the above second and third points that the problem is relevant with the BTS equipment. Observe the RXM test interface of CDU with the spectrum analyzer in peak hour in the daylight, it can be seen that unstable strong broadband interference and rise of back noise occur.




• First replace all boards (TRX, CDU, FPU, HPA, and power board) of this BTS (BTS20, with Tower Top Amplifier) one by one, at the same time observe the spectrum signal of RXM test interface, it can be seen that interference exists all the time. This indicates that the interference is relevant with the antenna feeder (including divider, combiner, feeder, antenna, lightning arrester, Tower Top Amplifier, jumper and connector) instead of the board. • Since the above BTS under test has the Tower Mount Amplifier, the antenna and feeder check is inconvenient, replace another BTS30(S4/4/4) (dual-CDU, and dual-polarization antenna) with interference and check the antenna feeder. • Since no interference exists in one of the cells while strong interference exists in another two cells in the BTS, interchange the antenna and feeder (changing the jumper at the top of the cabinet) of the cells which are with and without interference in the BTS in the evening. Then send idle BURST, it is found that the interference follows the antenna and feeder. This step helps further locate the fault which should exist in antenna and feeder system. NSN Internal Document 34 © Nokia Siemens Networks



• The situation remains the same even after replacing lightning arrester of antenna feeder and checking all jumper connectors. Then it can be sure that fault exists in the feeder or antenna. • Replace the jumper (i.e., antenna) at the top of the tower, it is found that the interference follows the antenna, so the feeder fault can be excluded while the antenna fault is quite possible. (It should be noted that the external interference at this step cannot be excluded because the actual installation place of the antenna does not change, but the external interference has already been excluded in the above step. ) • Finally check the antenna. The strong interference disappears immediately after the antenna is replaced on the tower by using the dualpolarization antenna. For further verification, replace the antenna of one cell with strong interference in another BTS20 with a new one, and then the interference disappears, thus the problem is solved here.



Antenna Connected Inversely Fault description: the interference bands 4 and 5 often occur in the traffic measurement after certain BTS is on service, the inter-cell handover success rate is very low and the congestion rate is up to 5%. There is no alarm in OMC.



Antenna Connected Inversely Fault location process

• Since the interference bands 4 and 5 occur along with low handover success rate and congestion, it is doubted that the interference causes the above phenomena.

• Check the frequency planning first, no problem is found. The external interference becomes the chief consideration after the frequency planning problem is excluded. Change the original used frequency 9 into the far-away frequency 94 to avoid external interference, but situation si the same. Confirmation made with the operator’s branch office indicates that the BTS is remote and without any high-power radio equipment nearby. It looks as if the frequency planning or external interference should be excluded.

• Since handover failure is involved at the same time, it is found that handover failure occurs between cells 1 and 3 according to the registration of outgoing/incoming cell handover performance measurement.



Page 37

Antenna Connected Inversely Fault location process

• The congestion analysis indicates that TCH assignment failure is usually caused by uplink. After registering the traffic measurement of uplink/downlink balance, it is found that the measurement item of uplink/downlink balance for cell 1 and cell 3 focus on level 1 and 11. This indicates that severe imbalance occurs between uplink and downlink. • The imbalance between uplink and downlink, in combination with much handover failure in cell 1 and 3 turn the doubt to the antenna and feeder which may be connected inversely. • On-site examination indicates that the antennas of cell 1/2/3 become crossed pair which causes the transmitter antennas of cell 1 and 3 to stay in the same cell, while the receiver antennas of them connect to another cell. The interference band and congestion disappear and the handover is all right after it is corrected.



Co-channel Interference Fault phenomenon: The co-channel interference of Huawei early 2.0 BTS (O2) in certain area leads to high call drop rate and poor voice quality. Serious interruption with occasional strong noise (whizz in general) occurs. It is after the BTS’s normal running for a certain time that the call drop occurs. Located in a little town (Du city) on the border of the city, the BTS is surrounded by the BTSs of the other manufacturer.



Page 39

Co-channel Interference Fault location process

• The frequencies assigned for the BTS are 64, and 92 (64 is of BCCH frequency). • In the optimizing test, the receiving quality (quality level is less than 3) is continuously good as the downlink signal level is -95dBm in the direction away from Huanggang and Du City. In the direction from Du City to Huanggang, the receiving quality is also good when the receiving level is more than 70dBm. Then move forward until to the place where TA=5, the receiving quality is sometimes good, and sometimes more than 5 in about 1 minute when the receiving level is about 75dBm. And network-drop occurs frequently when Idle-mode test is made at this place. It is suspected that downlink interference may exist on BCCH frequency. NSN Internal Document 40 © Nokia Siemens Networks



• Carry out continuous conversation test with one test MS and scan test for 64# frequency with another MS. The test carried out again in the section from Du City to Huanggang reveals that the signal strength of 64# frequency is already less than 100dBm near the Huanggang, and call drop already had occured. But the signal strength of 64# frequency rises up to 65dBm and disappears after a duration of 100 seconds when entering the downtown area of Huanggang. So it can be judged that the co-channel interference may be from the TCH frequency of cell nearby.




• Carry out continuous conversation test with one test MS and scan test for 64# frequency with another MS. The test carried out again in the section from Du City to Huanggang reveals that the signal strength of 64# frequency is already less than -100dBm near the Huanggang, and call drop already had occured. But the signal strength of 64# frequency rises up to -65dBm and disappears after a duration of 100 seconds when entering the downtown area of Huanggang. So it can be judged that the co-channel interference may be from the TCH frequency of cell nearby. • Carry out scan test for this frequency after arriving at the hotel, the signal level still remains high, but the conversation is not implemented on this frequency. The next day, carry out designated scan test at the place with the strongest signal of this frequency in the street of Huanggang, and test in Idle mode with another test MS. From the system message, it can be confirmed that 64# frequency is assigned to the TCH in HG08 cell with BCCH frequency 45, and the conversation is actually established once on 64# frequency in multiple times of conversation tests. NSN Internal Document 42 © Nokia Siemens Networks



• The interference disappears after the application is made to the customer to modify the frequency. At the same time, the customer should adjust the adjacent cell relation of peer equipment. • Carry out test again at the place where the original cochannel interference exists after the frequency is modified, the call drop and network drop disappear, and conversation quality Rx_Qual<3. It indicates that the problem of co-channel interference (downlink) is solved.



Adjacent-channel Interference Fault phenomenon: The call drop ratio is universally high and even up to about 15% in busy hour after several BTSs are cut over. And it is difficult for the call to be set up during on-site test. There is no alarm message in OMC system



Adjacent-channel Interference Fault location process

• All BTSs are connected to the same BSC, and call drop occurs after cutover of new BTSs. • The transmission quality is good, and TRX test is carried out for the above BTS with call drop and this indicates that each TRX is all right. No fault is found by checking the data and carrying out the test for 32BIE port corresponding to the BTSs. From the above analysis, the TRX fault, BSC hardware fault, A interface circuit fault and transmission fault can be excluded. • Analysis of traffic measurement result reveals that serious interference occurs in each cell of the above several BTS. Most cells have measurement values in interference band 4 and 5, and the number of channels falling into interference band 5 in several cells is up to 7. So it is sure that the interference in the above several cells is quite serious. NSN Internal Document 45 © Nokia Siemens Networks


Adjacent-channel Interference Fault location process

• It is found that there are many adjacent channels and the frequency planning is irrational after checking the frequency configuration of above BTSs and the adjacent cells. Especially, the area where the above BTSs are located, is newly added, and interference exists among them. And they also have interference in between them and the surrounding running BTSs. • Call drop disappears after adjusting and loading the frequency configuration of this area.



Interference Caused by Over-coverage Fault phenomenon: The hand over success rate of a certain GSM network is low, the call drop rate is high and conversation quality is poor. The hand over success rate is less than 80%, and the call drop rate is more than 2%. It is found that there are many times of downlink/uplink strength hand overs through view and analysis of traffic measurement data. while there are many times of bad downlink quality, and uplink strength among the times of unsuccessful handover. The analysis of cause of call drop indicates that the times of bad downlink quality are more than those of bad uplink quality . There is no alarm message in OMC system



Interference Caused by Over-coverage Fault location process

• From the result of traffic measurement, it can be judged that downlink interference may exist in the system or the coverage is not very good.

• The actual result of Drive Test shows that the strength of outdoor signal can be up to -80dBm above in the downtown area, that is to say, the coverage is all right. But serious over-coverage exists. For example, the service cell used in the building where BTS A is located is cell B with the same BCCH frequency as cell A1, while cell B is 6 kilometers away from BTS A in the suburb. In this way, the problem exists in two aspects:

– 1. The signal of cell B forms co-channel interference which leads to poor downlink link quality in coverage area of cell 1 of BTS A. It is found that ”****” is displayed in the test MS when this cell is locked during the test.

– 2. When cell B is selected as the service cell, its adjacent cell is only geographically adjacent to it, while the cell near the BTS A does not function as its adjacent cell. So when its signal is unavailable, the “effect of isolated island” will occur because the signal of its adjacent cell is poor, too. Then hand over fault and even call drop will easily occur. The on-site survey shows that the antenna of cell B is hung 50m above. The data provided by the customer reveals that the tilt angle is 5°, which is actually far less than 5°.



Interference Caused by Over-coverage Fault location process

• The cause for bad network indices lies in over-coverage, so the basic way is adopted to lower the antenna and adjust downtilt of it to make actual coverage area consistent with planned coverage area. Temporarily this problem can only be solved by adjusting the network parameter. The following operations can be adopted: lower the power level of cell B and add the adjacent cell of cell B, at the same time increase the level threshold of candidate cell from 10 to 15. The network indices exhibit obvious improvement after the above check and modification are performed for all cells in the downtown area, moreover, the hand over success ratio rises up to 85%, while call drop ratio drops to 1.3%.



Repeater Interference Fault phenomenon: subscribers of a certain network complain that they cannot occupy channel in some area for conversation since some day, or noise is heavy after occupancy of channel though the signal of MS is strong at this time. There are two directional BTSs in this area and both of them are BTS30 with the version 05.0529. The azimuth of the first cell directs at north. The BTS in this area works normally and the network indices conform to requirement before the complaint. Both BSC and MSC are devices of Huawei and two BTSs are connected in star mode. The traffic measurement indices show that traffic of both BTSs decreases obviously after occurrence of this problem, especially in the first and third cells. Although the signal of channel is very strong, the quality of voice is poor. Then it can be seen from traffic measurement that the interference band of these four cells is in class three, four, and five, and 95% of channels are under interference. In addition, interference of different classes also exists in other cells. So subscribers complained strongly. And there is no alarm message in OMC system



Repeater Interference Fault location process

• The feedback of subscribers shows the possible causes as follows: 1. Problem occurs in transmission and leads to error code; 2. Problem occurs in antenna feeder; 3. Fault exists in TMU; 4. Internal or external interference may exist. • The traffic measurement console shows the possible reasons as follows: 1. There may exist strong uplink interference signal in the north lean to west in this area. This leads to interference of different levels in the first, second and third cells, especially in the first and third cells; • It is found that it is difficult to put through the call in the first and third cells through on-site dial test. Although the call is put through, the quality of voice is very poor, and the voice is intermittent seriously with strong interference. If MS subscriber calls fixed telephone subscriber in this area, it is hard for fixed telephone subscriber to hear the voice clearly, instead, MS subscriber can hear fixed telephone subscriber clearly. This also proves the above analysis that the interference may be external (It can be judged from this point that interference only exists in uplink.)



Repeater Interference Fault location process

• Carry out test on site with antenna feeder analyzer. No problem of BTS itself is found, and the situation remains the same after TMU is replaced. Therefore, we ask customers whether there are such newly constructed equipment as microwave station, repeater, etc. surrounding the BTS. They told that they didn’t set up them. • It is said that China Mobile sets up a repeater in the area, which is located about two kilometers away in the north lean to the west of both BTSs of Huawei, and when it is activated, the problem will occur in Huawei BTS. Then customers negotiated with China Mobile several times. Finally China Mobile carried out on-site test with Huawei engineers together, and found that only if the repeater is switched off, the interference band and call become normal along with the recovery of Huawei BTS, if the repeater is activated, problem will occur soon in Huawei BTS, i.e. call cannot be put through, or interference is strong even after call is put through. The traffic measurement relevant with interference band in two different cases was handed on to China Mobile, and they approved our viewpoint. • Finally the problem is solved with the coordination of China Mobile. NSN Internal Document 52 © Nokia Siemens Networks


Microwave Interference Fault phenomenon: it is found that call drop rate in the second and third cell of a certain BTS (S2/2/2) in traffic measurement increases abruptly. Call drop rate is up to about 20% at some time.



Microwave Interference Fault location process

• View of BSC traffic measurement shows that idle TCH number in interference band in this BTS begins to increase in interference band 3-5 at about 8:30, in interference band 4, 5 at 10:00, and in interference band 1 at about 22:00. It can be judged from the above phenomenon that interference exists. • Since this BTS operates well before, the problem of frequency planning can be excluded. • Perform power-off restart and replace board for BTS. But interference still exists, so the possibility of TRX self-excitation can be excluded. • TRX management information reveals that interference exists in four boards of the second and third cell in this BTS, and the possibility of damage of the four boards at the same time is little, so the problem of TRX can be excluded. TRX board is replaced for caution, but interference remains NSN Internal Document 54 © Nokia Siemens Networks



• View of all BSC traffic measurement data shows that interference of different levels exists in the cells of all BTSs nearby, which is facing the same direction as the second and third cells in this BTS. And sometimes SDCCH channels in the cells with serious interference are all occupied at the same time, while the occupancy ratio of SDCCH at the same time is very little according to the amount of subscriber. So it can be sure that external interference exists in uplink, but the interference may be relevant with direction instead of frequency. • To further locate, jumpers of the first and third cell are interchanged on the rack top. As a result, it is found that interference occurs in the first cell, but interference disappears in the third cell, so this has proved the above judgement. • Since interference is not relevant with frequency, BTS interference may be caused by high-power signal sent into BTS system.




• Measurement on BTS divider output port with spectrum analyzer shows that high-power signal exists on 904MHz frequency (5M away from the used frequency), and this signal level come up to about –25dbm in BTS with serious interference, while in other BTS it is about –50dbm. So it can be judged that this signal has impact on BTS. • After frequency scanning around BTS with spectrum analyzer, it is found that a microwave antenna outputting high-power signal is at a frequency of 904. • When the microwave equipment is switched off and TRX management information is traced, the interference disappeared. NSN Internal Document 56 © Nokia Siemens Networks


False Interference Fault phenomenon: the reconstruction of some office found that the interference band of an S6/6/6 BTS20 in a county is very high. Theinterference band of two cells in it changed from 5 to 7.



False Interference Fault location process

• During the past interference test ofthis office, some cells with high interference band were located. This is because antenna intermodulation resulted in the interference of normal signal in case of large traffic and it made the interference band very high. So change the antenna, but the interference band did not fall down obviously after that. • During the prior period, the frequency planning has been checked several times, and the ad-frequency interference has been avoided basically. At the same time the frequency that may cause third order intermodulation in the cell has been adjusted. The problem of frequency planning has been basically excluded. It is unknown whether there is something wrong with Tower Top Amplifier, feeder or connector. But no problem was found after checking the antenna feeder system of the BTS several times. NSN Internal Document 58 © Nokia Siemens Networks



• Trace the state of interference band of each channel via Maintenance Console and find that high interference band mainly focuses on the four TRXs of the cell. • First set the frequency with less interference to the TRX with more interference, and find that the interference band does not change. It shows there is no relation to the frequency. Whether there is something wrong with TRX board or not is unknown. However, the situation is unchanged after the TRXs were interchanged. • Then confirm the TRX numbers, and find that the four carriers come from the same divider. From the above process, the problem may be the divider. But it is found that the interference band becomes high when the BTS is expanded from S4/4/4 to S6/6/6. While the combiner/divider used to be normal. Maybe the devices break down after some time. So we decided to change the combiber/divider to have a try.




• The idea that four carriers with high interference band are connected to the same second level divider before the divider is changed. The 7-dB gain of the divider has been considered during interference judgement. However, if the DIP switch is not moved to the right place while cascaded, the interference signal will be amplified by 7dB. Simultaneously, higher interference band will exist because the calculating error of 2.0 station itself is up to 5dB. Check the DIP switch of the divider. It was really not moved to “Off”. After it is corrected, view the traffic measurement of interference band 4 and 5. They fall down to 0. And there is a little value on the interference band three.



Content: Interference Analysis • Statistics – DL C/I and interference matrix – UL receive level of unused time slots

• Drive Tests



DL - C/I and Interference Matrix (BCCH) BTS-11 BTS-5 BTS-7

51 multiframe of BTS-5 51 multiframe of BTS-7 51 multiframe of BTS-11

0

T51,BTS-7

by drive tests

T51,BTS-11

interferers

T51,BTS-5

Search for T51 start time

235400

BTS operating with the same frequencies can be distinguished from each other nevertheless, as the multi frames are not synchronized with each other. So e.g. the strong BCCH signals of nearby BTSs usually are well separated in time. • 51 multi frame -> Measurements of the BCCH • 26 multi frame -> Measurements of the SACCH NSN Internal Document 62 © Nokia Siemens Networks


DL - C/I and Interference Matrix (BCCH) • Detection of interferers: The RX level and RX quality measurements can be performed on per TRX basis. Non hopping network -> with affected TRX single affected frequency recognized Hopping network -> affected MA list recognized only Often there are several nearby cells using the affected frequency or one of the frequencies of the affected MA list • Manual search for interferers: The potentially interfering base stations have to be switched off temporarily - Very time consuming. Especially in hopping networks hardly possible to find out the interferers • Search for interferers on basis of traffic: If the interference level increases with the traffic in a potentially interfering cell, than very probably this is indeed an interfering source



UL Receive Level of Unused Time Slots Measurement of uplink receive level on idle channels = uplink interference

Averaging over interferenceAveragingProcess (AP) = 1..32 SACCH periods Classification into interference bands based on interferenceAveragingProcess (BO1..BO4) = -110..-47 dBm

BSC tries to allocate TCH from best interference band (can be requested by MSC) If not available, BSC tries to take TCH from next band

BO5 –47 (fixed) rxLevUL = -75 dBm BO4 -90 BO3 -95 BO2 -100 BO1 -105 BO0 –110 (fixed) NSN Internal Document 64 © Nokia Siemens Networks

0 1 2 3 4 5 6 7


Interference • Detection of interference: •UL and DL RX level and RX quality statistics for busy TCH: For a RX level guaranteeing a stable connection (for outdoor higher than about -90 dBm) most of the RX quality measurements should indicate either quality 0 or 1 It is an indicator for interference if speech quality is bad and level high. • UL RX level of un-used TCH: The BTS can measure the RX level of unused time slots. Any signal detected there indicates interference directly.



Drive Tests


Example drive test


Drive Tests Example drive test



Drive Test - Exercises Case study • Cell 11 is...

Cell 11

showing high drop rate in NMS statistics. Customer complaints point to a possible problem area around a particular road crossing.

(high call drop rate)

Cell 11 • Cell 11 is ... covering part of a suburban city area including a road crossing. There are other cells in the area (see sketch), but Cell 11 is dominant in and around the road crossing.



(dominant in road crossing)

Drive Test - Exercises Information High traffic ?

• No frequency assignment problems on neighboring cells • Drive tests confirm that the road crossing is causing drops on very many calls • Cell 11 has signal level around –85dBm at the road crossing • Downlink RXQUAL values are showing worse quality than other cells in the area • Uplink RXQUAL is showing same characteristic as other cells in the area (no problem)

- 85 dBm

High interference (fading) ?


Cell 11 (dominant in road crossing)


Call drops

Drive Test - Exercises

Cause of the problem Problem is caused by a competitor site at the building in the upper right side of the road crossing. The competitor cell provides around –35 dBm at the road crossing, which creates signal differences in "our" network of around 50dB between the serving level and the interfering level.

Competitor Cell - 35 dBm

Cell 11 (dominant in road crossing) - 85 dBm




Solution The only solution is to provide a higher signal level at the road crossing, so that the signal difference gets lower than the current 50 dB.

Adding New Site - 40 dBm

- 35 dBm

• The probable solution is to add a new site close to the road crossing. • Plus frequency reallocations

Cell 11 (dominant in road crossing) - 85 dBm


Competitor Cell


Drive Test - Exercises 10

GSM specs requires that a mobile must overcome this kind of interference when the signal difference is around 40dB even if the serving level is close to the sensitivity level of the mobile. There is generally no exact specifications for this kind of interference, but it is commonly known, that this phenomenon is happening. Note: The competitor channel is not adjacent to the serving channel, it might be separated by several MHz!

0

Rx Level [dBm]

20 -35 dBm -40

-60

-80

-100

CIR (competitor server) max. 40 dB allowed

-85 dBm

Cell 11 (dominant in road crossing) 200 kHz F1


Competitor „ channel“


200 kHz Fn

Frequency range


New BTS

A new BTS with a 3+2+3 configuration has been implemented in the network during night-time. The implementation engineer standing next to the BTS has made successful test calls on all timeslots of all TRX's.Using a test mobile a short drive test has been done at approx. 500m from the site and the following verifications have been done :

· · ·

The directions of all 3 sectors by looking at the BCCH frequencies The functionality of the handover relations between the sectors The uplink cabling of each TRX by making one outdoor test call on each TRX

The implementation engineer parks his car to make some phone calls to arrange the rest of his work. The first call is successful. Another call is also successful. The 3rd call he has to make fails during the setup. Again he makes some phone calls to check the new site and get's the following results : 1. OK 2. OK 3. NOK 4. OK 5. OK 6. NOK 7. OK All the time he is standing in the area between sector 1 and 2 and making phone calls on sector 1. NSN Internal Document 73 © Nokia Siemens Networks


Drive Test - Exercises Description of the actions • The BSS Technician went on site for checking the hardware, and the operator's OMC controller checked the alarms. Nothing was found. The frequency plan was checked, it was clear that too many frequencies were involved here and that interference can't easily be avoided. Different parameter fine tuning trials were implemented, but they didn't improve anything. Base band hopping was deactivated for some hours to have statistics per TRX. But nothing came out of this test. • Some frequencies were swapped and changed, but no real improvement was seen. The problem should come from an external source or from a problem with the hardware itself (intermodulation problem was proposed as it was seen on other cells already). • The operator went back on site to measure the UL interference, and concluded that UL interference is coming from MS of other cells. It was reported that the MS were coming from the back.



Drive Test - Exercises Cause of problem It turned out that the side and back lobes from the antenna were picking-up UL signals from MSs on cells behind

Solution Change the antenna

• •

Replace the antenna with one which has less back and side lobes Move the antenna on the roof, so that the building is a screen for the MS on the cells behind and the back/side lobes. The MSs operating on the cells behind (co-channels) are not seen by the antenna's back/side lobes anymore

Solution 1

Move the sector Solution 2 NSN Internal Document 75 © Nokia Siemens Networks


Thank You



Part7 GSM Interference Analysis and Optimization

Recommend Documents