Freq hopping & freq planning

CHAPTER 7

FREQUENCY HOPPING & FREQUENCY PLANNING

Section Objectives

Upon completion of this section, the student will be able to understand thefollowing:

The principles and appreciate the advantages of frequency hopping.

The two types of implementation of frequency hopping .

The Hardware Requirements for frequency hopping.

The comparison between BBH and SFH

Economical advantages of Hopping

Frequency Planning for Non Hopping and Hopping Scenarios

FREQUENCY HOPPING TECHNIQUE

Frequency hopping is an old technique introduced firstly in military transmissionsystems to ensure the secrecy of the communications and combat jamming. Thephilosophy is as simple as changing the frequency used in transmission at regularintervals. It has been included in GSM specifications mainly in order to deal with twospecific problems which affect transmission quality:

* Multipath Fading: The immunity to fading increases by exploiting itsfrequency selectivity, because using different frequencies the probability of beingcontinuously affected by fading is reduced, so the transmission link quality is improved.This characteristic is normally referred as Frequency Diversity. This improvement ismuch more noticeable for slow moving mobiles.

* Interference: The situation of permanent interference coming from neighbourcells transmitting the same or adjacent frequencies is avoided using Frequency Hoppingbecause the calls will spend the time moving through different frequencies not equallyaffected by interfering signals. This effect is called Interference Averaging.

As used in GSM, the frequency is changed on a per burst basis, which meansthat all the bits in a burst are transmitted in the same frequency (known as slow frequencyhopping).

FREQUENCY DIVERSITYIt is well known that considering a cellular urban environment in most cases

multipath propagation will be present and, as a consequence of that, important short termvariations in the received level are frequent . This is called Rayleigh fading which resultsin quality degradation because some of the information will be corrupted.

For a fast moving mobile, the fading situation can be avoided from one burst toanother because it also depends on the position of the mobile so the problem is not soserious, but for a stationary (or quasi-stationary) one the reception may be permanentlyaffected resulting in a very bad quality, even a drop call.

Once the information is received by the mobile or the base station, the only‘weapon’ to cope with the disturbance produced by the fading (errors in the informationbits) are the decoding and deinterleaving processes, with an effectiveness limited by thenumber of errors they have to deal with.

Frequency hopping is able to take advantage of the frequency selective nature ofthe fading to decrease the number of errors at the same time that they are temporallyspread. If a mobile is stationary (or quasi stationary) at a point where a deep fade existson the frequency in question and the system uses frequency hopping, the call will spendtime on frequencies which are not faded at that point. As a result, the decoding anddeinterleaving processes can more effectively remove the bit errors caused by burstsreceived whilst on the faded frequencies (errors will be randomly distributed instead ofhaving long bursts of errors). This increase in effectiveness leads to a transmissionquality improvement of the same proportion.

INTERFERENCE AVERAGINGThe second effect of frequency hopping is referred as averaging the interference

experienced by the calls. Considering a non hopping system, the set of calls on theinterferer cells which can interfere with the wanted call is fixed for the duration of thosecalls and some calls will be found with very good quality (no interference problems)whereas some others with very bad quality (permanent interference problems). Withhopping, that set of interfering calls will be continually changing and the effect is thatcalls tend to experience an average quality rather than extreme situations of either goodor bad quality (all the calls will suffer from a controlled interference but only for shortand distant periods of time, not for all the duration of the call).

This interference averaging means again spreading the raw bit errors (BERcaused by the interference) in order to have a random distribution of them instead ofbursts of errors, and therefore enhance the effectiveness of decoding and deinterleavingprocesses to cope with the BER and lead to a better value of FER.

NON HOPPING SYSTEM

Wanted Call (f1 fixed)

Interfering Call (f1 fixed)

Corrupted Bursts because of Interference

Figure 3: Cochannel Interference for a Conventional Fixed System

f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1

f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1 f1

HOPPING SYSTEM

Wanted Call (Pseudorandom Hopping over f1, f2, f3)

Interfering Call (Pseudorandom Hopping over f1, f2, f3)

Corrupted Bursts because of Interference

Figure 4: Cochannel Interference for Hopping System

It is easy to see from figures 3 and 4 how this procedure works to improve the quality ofthe system. Frequency Hopping can be used to improve quality but also, advantages of FH allowto add more carriers while keeping the same level of quality so, there is a trade-off Quality vs.Capacity whit Frequency Hopping. Considering a fixed system with a certain quality and capacityas a reference, just only enabling the frequency hopping feature, a system with better quality andthe same capacity will have been achieved. Provided that the reference quality was good enough,more new carriers (meaning an increase of capacity) could be added to the hopping system untilthe system quality goes back to the previous value in the fixed system. The result will be ahopping system with the same quality as the non hopping one but with much more capacity. (SeeFigure 5).

Figure 5: Quality - Capacity Trade-off

f1 f3 f1 f2 f3 f1 f2 f1 f3 f1 f2 f3 f2 f1 f2 f1 f3 f2 f3 f1 f2 f1 f3 f1 f3 f2 f1

f3 f2 f1 f3 f1 f3 f2 f3 f1 f2 f1 f3 f1 f2 f3 f1 f2 f1 f3 f2 f1 f3 f1 f2 f1 f2 f3

HOPPING SYSTEMHOPPING SYSTEM HOPPING SYSTEM

C

A

P

Q

U

A

NON HOPPINGSYSTEM

C

A

P

Q

U

A

C

A

P

Q

U

A

OROR Q

U

A

C

A

P

The main idea is very simple: “For the same capacity FH improves the quality,and for a given average quality FH makes possible increase the capacity”.

FREQUENCY HOPPING OPERATIONSo far, only the aspects associated with the effect of using different frequencies

have been treated and nothing about the way in which the system hops over the set offrequencies has been commented.

From the infrastructure point of view, there are two ways of implementingFrequency Hopping in a Base Station System (BSS), one referred as Base Band FH andanother as Synthesiser FH. Their operation philosophies only differ in the way theyestablish the Base to Mobile Station link (downlink) in the Base Station part (consideredfrom the Mobile Station, there is no difference at all between both types of frequencyhopping) but it is worth to discuss the impacts in the operability.

BASE BAND FREQUENCY HOPPING (BBH)Its main characteristic is that the transmitting units (DRCUs) are always

transmitting a fixed frequency and frequency hopping is performed by moving theinformation for every call among the available DRCUs on a per burst basis. A call willstart in a particular timeslot of one DRCU and will move to the same timeslot of the otherDRCUs spending the time associated to a burst (about 577 microsec.) in each DRCU(and hence in each different frequency). Changing the frequency implies changing theDRCU (the call hops between DRCUs). It must be noticed that although data aretransmitted by different DRCUs, all the processing (coding, interleaving, etc.) is done bythe digital part associated to the DRCU the call was initially assigned to, and only afterthat, the information is routed to the corresponding transmitting unit.

Looking at the uplink, MS to BS direction, the call is always received by theDRCU the call was initially assigned to.

The following table in Figure 6 will explain the performance: Assuming a cellwith 4 DRCUs and 4 frequencies (f1 to f4), Base Band Hopping in a cyclic way and acall assigned to DRCU 3 timeslot 5 the call process will be described.

f4 f3 f2 f1

TRANSMISSION RECEPTION

BURST # Timeslot Frequency DRCU # Timeslot Frequency DRCU

1 5 f3 3 5 f3 3

2 5 f4 4 5 f4 3

3 5 f1 1 5 f1 3

4 5 f2 2 5 f2 3

5 5 f3 3 5 f3 3

6 5 f4 4 5 f4 3

7 5 f1 1 5 f1 3

... ... ... ... ... ... ...

Figure 6: Call Progress on a Base Band Hopping System

The first consequence is that as many DRCUs as frequencies in the hoppingsequence need to be physically equipped in the cell, which means that the restriction inthe number of frequencies to hop over will come from the traffic requirements in the cell(number of DRCUs equipped in the cell). At the same time, because the DRCUs do notneed to retune each burst, this type of FH can be used in cells where the combination of

TCH

DRCU

3

TCH

DRCU

4

TCH

DRCU

2

BCCH

DRCU

1

transmission signals to the antenna is done through Remote Tune Combiners (highcapacity cells usually equip that kind of combiners).

As required by the GSM system, the BCCH frequency must be always on the airtransmitting the maximum power (Power control do not apply for BCCH carriers), andparticularly its timeslot 0 can not hop (if additional Common Control Channels -CCCH-are allocated in other timeslots -1 to 7-, they will not be allowed to hop as well). From theworking philosophy described for BBH there is no problem at all for timeslots 1 to 7 ofthe BCCH carrier to hop (provided that they are not used to allocate CCCHs) since thepermanent presence of this frequency on the air is guaranteed, but the busts using thisfrequency will be transmitted at maximum power. This is another important point inBBH: The BCCH frequency can be included in the hopping sequence and also the BCCHcarrier can carry hopping calls in timeslots 1 through 7. If Downlink Power Control isenabled in the Base Station it will only take effect for the bursts transmitted in the non-BCCH frequency.

SYNTHESISER FREQUENCY HOPPING (SFH)In this type of hopping the DRCU changes the transmitting frequency each burst

and the call always stays in the same DRCU where it started. The DRCU is able to retuneto a different frequency for transmission every 577 microsecs., and because such fastfrequency changes, Remote Tune Combiners (RTC) must not be equipped if synthesiserFH is to be used. So, Synthesiser Frequency Hopping requires the use of widebandcombiner devices such as hybrid combiners.

The main advantage of SFH is that there is no restriction on the number ofcarriers equipped in the cell. The number of DRCUs will be determined by the trafficneeded to be handled, but they can hop up to over 64 different frequencies (limitationcoming from GSM specifications) if they are available according to the planning.

Since with SFH the number of frequencies can be greater than the number ofcarriers, if the BCCH frequency is included in the hopping sequence, its presence on theair would not be guaranteed unless the BCCH carrier transmits it when no other carrierdoes. This has two implications:

* The BCCH frequency can be included in the hopping sequence (SFH throughBCCH) but timeslots 1 to 7 from BCCH carrier can not be used to carry traffic becausethey must be reserved to put the BCCH frequency on the air when necessary (dummyBursts -DB-). At the same time, for the bursts transmitted in the BCCH frequency theDRCUs will do it at the same power used by the BCCH carrier (BCCH power).

* The BCCH carrier will never hop. It will either carry traffic in timeslots 1 to 7on the BCCH frequency (if not included in the hopping sequence) or transmit dummybursts.

Because of this, BCCH frequency is not included in the hopping sequence, soBCCH timeslots do not hop and non-BCCH timeslots do. Following is a more detailedexplanation of this, considering an example.

Assuming a cell with 2 DRCUs and 5 frequencies (fb for the BCCH and f1, f2,f3 and f4 for hopping -fb being the lowest one-), doing Synthesiser Hopping in a cyclicway on DRCU 2 and a call assigned to DRCU 2 timeslot 5 the call process is describednext in the table of Figure 7.

BCCH Frequency: fb

Hopping Frequencies: f1, f2, f3, f4

Transmission and reception are always routed through the same timeslot in thesame carrier (it does not happen for transmission in Base Band Hopping). In this case, fortimeslot 5, depending on the inclusion of BCCH frequency in the hopping sequence ornot, the evolution of the call will be different (see table of Figure 7):

TCH

DRCU

2

BCCH

DRCU

1

SFH through BCCH SFH

(BCCH frequency not included in thehopping sequence)

BURST # BCCHDRCU

Frequency

TCH DRCU

Frequency

BCCH DRCU

Frequency

TCH DRCU

Frequency

1 - fb(BCCHPow)

fb f1

2 fb (DB) f1 fb f2

3 fb (DB) f2 fb f3

4 fb (DB) f3 fb f4

5 fb (DB) f4 fb f1

6 - fb(BCCHPow)

fb f2

7 fb (DB) f1 fb f3

... ... ... ... ...

Figure 7: Call Progress on a Synthesiser Frequency Hopping System: Effect of Inclusionof BCCH Frequency in the Hopping Sequence.

For frequency hopping operability, GSM defines the following set ofparameters:

* Mobile Allocation (MA): Set of frequencies the mobile is allowed to hopover. MA is a subset of all the frequencies allocated by the system operator to the cell(cell allocation) although it can be the same.

* Hopping Sequence Number (HSN): Determines the hopping order used inthe cell. 64 different HSNs can be assigned, where HSN = 0 provides a cyclic hoppingsequence and HSN = 1 to 63 provide various pseudorandom hopping sequences.

* Mobile Allocation Index Offset (MAIO): Determines inside the hoppingsequence which frequency the mobile starts to transmit on.

* Frequency Hopping Indicator (FHI): Defines a hopping system made up byan associated set of frequencies (MA) to hop over and a hopping sequence (HSN).

Motorola equipment allows the system operator to define the hopping system ona per timeslot basis. So, allowing different hopping configurations for different timeslots.

This is very useful for the purpose of interference averaging and to randomise thedistribution of the errors.

HARDWARE REQUIREMENTS FOR FREQUENCY HOPPINGAs it was previously mentioned, Frequency hopping has been a feature included

in Motorola equipment since the first stages of GSM deployment. As a consequence ofthat, all the equipment, even the oldest ones, has the possibility of working withfrequency hopping without any modification at all in the hardware.

The only thing that must be considered is the combining issue: Cavity combinersoblige to use Base Band Hopping, so if Synthesiser Frequency Hopping is to beequipped, hybrids will be required for combining purposes.

The main consequence of using hybrids for combination is that it must be donein an intelligent way in order not to be affected by the initial limitations of hybridscompared with cavity combiners. The situation can be summarised as follows: Up to sixcarriers can be combined through a Cavity Combine Block (CCB) giving the maximumspecified transmit power at the top of the rack (20 W), whereas only two carriers will bepossible through hybrid combiners if those 20 W of Power are required.

This does not mean any limitation at all, since the same situation obtained usingcavity combiners can be achieved with hybrids by using the different antennae thatsystem operators install for spatial diversity purposes. Considering a cell where sixcarriers are going to be equipped, and three antennae available (typically one transmittingand two receiving with diversity), the two possibilities are:

Cavity Combining Blocks: The six carriers are combined and transmitted bythe same antenna all together (Figure 8). The maximum power can beachieved at the output of the cabinet (20W).

Figure 8: Cavity Combination of Carriers used with BBH

DRCU

DRCU

DRCU

DRCU

DRCU

DRCU

CAVITY COMBINER

TX RX2RX1

Max. Power (20W)

* Air Combining: The six carriers are combined by pairs through hybridcombiners, and the three outputs fed each one to the three antennae (one transmitting andtwo receiving, for diversity) through a duplexer (only necessary in the two antennae usedsimultaneously for transmission and reception). The DRCUs are calibrated to transmitmaximum Power (20W) at the output of the duplexer (top of the rack). The six carriersare combined on the air, after being radiated by the antennae, and the duplexers permit touse the same antennae for both, transmission and reception (Figure 9). With thisconfiguration, two antennae transmit and receive, and the other one only transmits. So,for both configurations, cavity combining blocks and air combining, the maximum powerat the top of the rack (20 W) can be transmitted.

Figure 9: Air Combination used with SFH

An additional advantage of this configuration is that fewer frequencies arecombined in the same antenna (2 in each antenna, compared to 6 in one antenna withcavity combiners) so, less intermodulation products are generated.

Even more than 6 carriers can be air combined maintaining the maximum powerat the output without increasing the number of antennae, by using dual polarisationantennae. In fact, one of these antennae can be considered as being equivalent to twoconventional antennae, so up to four carriers can be combined in one antenna usinghybrids and transmitting the maximum power.

DRCU

DRCU

DRCU

DRCU

DRCU

DRCU

HYBRID

TX3 TX2/`TX1/RX1

Max. Power (20W)

HYBRID HYBRID

DUPLEXER DUPLEXER

GENERAL COMPARISON OF BBH AND SFHThe main interest of this chapter is to compare BBH with SFH and point out theiradvantages. The most important aspects will be considered for comparison purposes.

* Quality: The improvement in quality is the consequence of the frequencydiversity, so both BBH and SFH offer the same possibilities to enhance thequality of a network, provided that the same conditions are considered. In thatway, for low traffic cells, (few carriers) SFH allows higher improvement becauseof the higher number of frequencies that can be used.

* Capacity: The increase in capacity that can be achieved with frequency hoppingis related to the reuse of the frequencies. The tighter this reuse, the higher thecapacity increase. As it has been described, the commonly used patterns for BBHand SFH lead to tighter reuse patterns in the last case (SFH) and thus the capacitycan be increased more if Synthesiser Frequency Hopping is implemented.

* Frequency Planning: Frequency Hopping leads to a simplification in planningissue, and a reduction of time and work. Base Band Hopping implementation stillrequires a frequency plan, because frequencies continue being associated todedicated DRCUs (each DRCU transmits continuously the same frequency).Synthesiser Frequency Hopping, on the contrary, needs only the planning ofBCCHs (non hopping) with no planning at all required for the hopping carriers.The frequencies are regularly distributed in the cells of a site, and reused in thatway for all the sites (1x3 reuse pattern). SFH reduces the planning issue only toBCCH planning, always using a number of frequencies high enough to achieve agood BCCH layer, and hence being easy to plan. The best solution to save timeand money in planning is to select SFH implementing a 1x3 reuse scheme.

* Optimisation: Simplicity in optimisation tasks is achieved using Frequencyhopping. Base Band Hopping reduces the optimisation periods, since it is done ona per cell basis, instead of doing it on a per carrier basis. According to SynthesiserHopping, it is done in a different way, because of the different philosophy ofworking, where the correspondence between cells and frequencies is not so direct.Dealing with a controlled interference environment, the optimisation must beapproached by modifying the hopping sequences looking for orthogonalitybetween them in order to reduce the probabilities of cochannel interference. Thisissue is quite easy and fast, pointing at SFH as the most advisable type of hoppingin order to take advantage of the reduction in optimisation.

* Hardware impacts: Both types of hopping can be implemented in all Motorolaequipment without any change, so there is no hardware impact at all. The onlyrestriction Synthesiser Hopping introduces is that Cavity Combiners can not beused together with SFH. Air Combining, as described in previous chapters, allowsmaximum power of transmission even if high number of carriers are equipped inthe cells so there is no limitation at all to use hybrid combiners, having thepossibility of implementing both, BBH or SFH.

* Hopping on the BCCH: Implementing Base Band Hopping, all the channelsdedicated to carry traffic (TCH carriers and non signalling timeslots on the BCCH

carrier) can take advantage of it, since the BCCH is allowed to hop as well on nonsignalling timeslots. On the contrary, using Synthesiser Hopping, it is no worth toconfigure the hopping through the BCCH, since its non signalling timeslots wouldnot be available for carrying traffic, losing capacity in the cell.

* Flexibility: The main advantage of Synthesiser Hopping appears at the time ofintegrating a new site, because of the simplicity of this task: Only a cleanfrequency is needed for BCCH purposes, and the same frequencies used in all thesites will be reused in the new one, following the conventional 1x3 pattern. Thegrown of the network is easier and faster if SFH is implemented on it. There isanother point to be considered and it is the interdependence between all thecarriers in a cell existing in a Base Band Hopping system: As all the calls movearound all the carriers, a faulty DRCU will affect all the calls. On the other hand,using SFH, all the timeslots work independently from the point of view ofhopping, so there is no influence at all from problems appearing in any of them.

* Economic advantages: From the economic point of view, Synthesiser Hoppinggives more advantages, because the planning and optimisation tasks are almostinexistent, so the period of time necessary to deploy a network is much shorter. Atthe same time, as more capacity per site can be achieved, additional savings canbe obtained because fewer sites will be required for the same capacity goal.

The table in Figure 10 summarises all the previous results.

ASPECTS OF COMPARISON BETTER OPTION TOIMPLEMENT

QUALITY IMPROVEMENT BBH & SFH

CAPACITY INCREASE SFH

SIMPLICITY IN FREQUENCYPLANNING

SFH

SIMPLICITY IN OPTIMISATION SFH

HARDWARE IMPACTS BBH & SFH

HOPPING ON THE BCCH BBH

FLEXIBILITY SFH

ECONOMIC ADVANTAGES SFH

Figure 10: Comparative Analysis BBH vs. SFH

ECONOMIC ADVANTAGES OF FREQUENCY HOPPING

There are some points involved in the deployment of a system where FrequencyHopping gives the possibility of saving a significant amount of money. A system operatorthat has decided to implement FH in his system can save money taking advantage of:

* Reduction in the number of sites needed for the same capacity andquality, respect to the fixed system case.

* Reduction in time required for planning and optimisation tasks, respectto the fixed system case.

The first aspect, fewer sites required, can be understood with the followingexample: “A new system operator has to deploy a system from scratch to give service to200000 subscribers. The spectrum assigned to this operator is 9.6 MHz, meaning 48frequencies are available”. Three possibilities can be considered:

* Conventional Fixed Frequency System

* Base Band Hopping System

* Synthesiser Frequency Hopping System

Considering standard values used for planning issues:

* 2% Blocking

* 25 mErlangs by subscriber

* Traffic offered given by the table in Figure 14.

The differences can be summarised in the next table (Figure 11):

COMPARATIVE ASPECTS FIXED SYSTEM BBH SYSTEM SFH SYSTEM

BCCH Reuse Pattern: 4x3 4x3 4x3

Frequencies Required for BCCHs: 12 12 12

TCH Reuse Pattern: 4x3 3x3 1x3 (n/2n)

Frequencies Assigned for TCHs: 36 36 36

Carriers per Cell: 4 5 7

Site Configuration: 4/4/4 5/5/5 7/7/7 (*)

Sites Required (3 cells/Sector): 77 57 39

Relative Investment: 100 75 50

(*) Air Combining will be used.

Figure 11: Comparison between BBH, SFH and Fixed systems

The main conclusions are (See Figures 11 and 12):

* The same system capacity can be achieved and only 75% of theinvestment necessary to implement a Fixed System is required in a BaseBand Hopping System.

* The same system capacity can be achieved and only 50% of theinvestment necessary to implement a Fixed System is required in aSynthesiser Frequency Hopping System.

Figure 12: Reduction in Costs Achieved with Base Band/Synthesiser Frequency Hopping

The second point in which costs can be reduced is the frequency planning andoptimisation activities:

* In a BBH system, all the calls move around all the frequencies (carriers) andbecause of that, the resultant performance observed is, in a certain way, anaverage of the behaviour experienced by all the frequencies separately. The mainresult is that planning and optimisation on a per carrier basis (each frequencyseparately) is not necessary and it is enough by doing it on a per cell basis (groupof frequencies).

* In an SFH system, as it was discussed, the reuse patterns are much more tighterthan the ones used in a fixed system, so the planning is more simple. For thecommonly implemented 1x3 pattern, all the frequencies are reused in a regularway in all the sites. The available frequencies are splitted into three groups andassigned regularly as the hopping sequence to each sector in all the sites. In thisway, the planning work is reduced to plan the BCCHs (so, no TCH planningrequired), which will work without hopping, and built up the hopping sequences.The optimisation task is reduced to take care of the specific areas where collisionscould appear because of irregularities in the sectorisation of the network. Theoptimisation is as simple as removing common frequencies from the sequences in

INVESTMENT REQUIRED

10050% Saving25% Saving

75

50

FIXED SYSTEM BBH SYSTEM SFH SYSTEM

the cells where this interference can be present, which is referred as “look fororthogonality in the hopping sequences”.

Although this approach has been done from the point of view of developing anew system, it must be considered as well that moving a system from fixed frequencyoperation to Base Band Hopping does not imply any additional cost at all. The increase innumber of subscribers can be achieved without adding new sites, but more carriers needto be equipped in the existing sites. Similarly, for Synthesiser Frequency Hopping, theonly requirement is that hybrids be used for combining the carriers into the antennae.

As seen above, the technical advantages of frequency hopping described, resultin significant savings for system operators, not only because of the capacity increase butalso for less time/resources required for frequency planning and network optimisation.

SUMMARYFrequency Hopping operation can be summarised as follows:

* Two effects:

- Frequency Diversity: Protection against frequency selective deepfading, affecting stationary or quasi-stationary mobiles.

- Interference Diversity: Randomise the interference situation and spreadthe errors to enhance decoding and deinterleaving effectiveness.

* Two ways of working:

- Base Band Hopping (BBH): The calls hop between DRCUs. Diversitygain depends on the traffic in the cell (number of carriers).

- Synthesiser Frequency Hopping (SFH): The DRCUs hop (change theirtransmitting frequency). Diversity gain depends on the spectrum allocation(number of frequencies to hop over).

* Possibilities offered:

- Improve the system quality.

- Enhance spectral efficiency leading to significant capacity increases.

- Reduce the number of sites, planning and optimisation required todeploy the network, saving money to the system operator.

FREQUENCY PLANNING & REUSE PATTERNSFrequency Planning Aspects:

The primary objective of frequency planning is to ensure that, given the limited RFspectrum, we achieve the required capacity (traffic channels), keeping the interferencewithin specified limits.

There are 2 types of Frequency Planning:

- Frequency Plans based on Reuse patterns ( manual )- Frequency Plans based on heuristic algorithms ( automatic).

Manual planning is done by dividing the available frequencies in to a number offrequency groups ( as per a selected reuse pattern ) and assigning frequencies to varioussectors/cells.

Suppose we have ‘n’ frequencies. For a 3 cell repeat pattern with 3 sectors, we have 9frequency groups, each group having n/9 frequencies.

The sectors are labelled A1,A2,A3, B1,B2 and so on.

Assuming that an operator has 32 frequencies, for example, from ARFCN 63 to 94, thefrequencies could be grouped as shown in the table below.

A1 A2 A3 B1 B2 B3 C1 C2 C363 64 65 66 67 68 69 70 7172 73 74 75 76 77 78 79 8081 82 83 84 85 86 87 88 8990 91 92 93 94

OR,

A1 B1 C1 A2 B2 C2 A3 B3 C363 64 65 66 67 68 69 70 7172 73 74 75 76 77 78 79 8081 82 83 84 85 86 87 88 8990 91 92 93 94

m X n reuseThe basic principle, cellular systems are based on, is the reuse of the frequencies

in order to obtain the highest capacity with the minimum spectrum. The more thefrequencies are used inside a certain coverage area, the more amount of traffic (capacity)can be carried. The possibilities of reusing the frequencies are limited by interferenceproblems arising when the same frequency is used in two cells which are too close eachother (cochannel interference).

It can be considered that a cellular network is made up by a basic unit, in whichall the frequencies are used, repeated all along the area which is intended to be covered.This basic unit (set of cells) is usually known as cluster, and it is the pattern used todeploy the network. The size of the cluster is directly related to the capacity that can beachieved: The smaller the size, the more times will be needed to be reused in thecoverage area, so the higher the reuse ratio and hence, the capacity.

Figure 13: Frequency Reuse Pattern: Cluster

A1

A2A3

B1

B2B3

C2C3

C1

A1

A3A2

A1

B1

C1

A3

B3C3

B2

C2

A2

A1

C1B1

The usual way to refer to a reuse pattern is by giving the number of cellsincluded in the cluster as well as its configuration. In that way, a cluster made up by msites with n cells per site, giving a total of p = m*n cells, will be referred as “mxn reusepattern”. Any frequency will be used once and only once inside the cluster.

As an example, a 12 cell cluster made up by 4 three-cell sites, known as 4x3reuse pattern, is represented in Figure 13, meaning that one frequency will be reused onceeach 12 cells or, equivalently, that 12 frequencies (one per carrier) will be needed toconfigure this cluster (a 4x3 reuse pattern with, for instance, 3 carriers per cell wouldrequire up to 36 different frequencies).

Figure 14: 4x3 Frequency Reuse Pattern

Considering a conventional fixed frequency system in GSM, it has been agreedthat a 4x3 reuse pattern is the best compromise solution taking into account the cochannelinterference and the reachable capacity.

Higher capacity goals, without allocating more spectrum, lead to differenttechniques able to control the interference and allow the system operator to use smallerclusters (tighter frequency reuse patterns). Frequency hopping is the most efficient one,considering the very small cluster size that can be achieved. As it will be described laterin the document, a 3x3 reuse pattern (see figure 15) can be successfully implemented in asystem working with Base Band Hopping.

Figure 15: 3x3 Frequency Reuse Pattern implemented with BBH

More aggressive reuse patterns such as 1x3 (all the sites reuse the same set offrequencies) are possible in a system working with Synthesiser Frequency Hopping(Figure 16), although it must be noticed that with SFH more than one frequency can beassigned to each carrier. This reduction in the size of the cluster, respect to the 4x3 one,can be used to increase capacity. The results achieved in the systems alreadyimplemented using that configuration prove its effectiveness to allow a very highcapacity increase.

Figure 16: 1x3 Frequency Reuse Pattern implemented with SFH

It is important to notice that, either for a fixed system or a Base Band Hoppingsystem, the ratio Number of Frequencies/Number of carriers is always 1, whereas for a

Synthesiser Frequency Hopping system, it can be higher than one. In fact, as it will beshown later in this document, for these systems it is recommended to have at least doublenumber of frequencies than carriers.

SFHMost of the SFH networks employ 2 different re-use plan for the BCCH and TCH layers.Since the BCCH will not be hopping, conventional fixed frequency re-use plans such as4X3 or 5X3 will be used. It is always a design goal to have a best BCCH layer, within theresource of the network. As for the TCH layer, the common methodology would be 1X3(1 site 3-sector) re-use pattern. This is a much more efficient spectrum utilization, whichis not possible in a fixed frequency system as the resultant C/I would be degraded badlybeyond ½ of the cell radius. An even more aggressive re-use plan 1X1 (1 site 1 sector) isfeasible in networks where the operating environment permits it. 1X1 is by far the mostefficient and yet practical aggressive re-use plan tested and proposed by Motorola.Nevertheless, careful planning has to be practiced to achieve good results. The guidelines are outlined in the next section.

Loading Factor (or sometimes termed as Fractional load factor) is an importantparameter in SFH systems. It is calculated as:

Since the number of frequency channels is always higher than the transceiver count in acell, some channels will be idle at one time. Thus, loading factor is equivalent to themaximum channel-occupancy to total-channel ratio in a cell at any given instant. Thelower the value the lower is the channel loading, which indicates fewer collisions offrequencies and hence better quality.A theoretical maximum of 50% is permitted in 1X3 SFH. Any value higher than 50%practically results unacceptable quality. Some commonly used loading factor are 40%,33%, 25% etc. In 1X1 SFH, a practical tested loading factor is 1/6 or 16.7%. For a roughcomparison, this is about equivalent to a 33% loading in 1X3 SFH or a well-planned4X3Xn fixed re-use network, as far as average quality is concerned. In terms of spectrumutilization or capacity, 1X1 at 16.6% loading is equivalent to 1X3 at 50% loading.

BBHDifferent re-use patterns are employed in BBH systems. Since the number of hoppingfrequencies must equal or less than the number of transceivers in the cell, the quality gainof BBH is higher in the cells with higher transceiver count. As a result, a progressive re-use pattern is usually used. This is analogy to a layered cake with a loose BCCH plan atthe base and progressively tighter plan for each subsequent transceiver added to the cell.

loading factor = (highest non BCCH transceiver count in a cell)(Number of hopping channels)

For example:BCCH – 4X3 plan1st TCH – 3X3 plan2nd TCH – 2X3 plan … and so on

Alternatively, a homogeneous re-use plan that is tighter than conventional 4X3 can beused. The widely used pattern would be homogeneous 3X3 re-use plan, which yieldscomparative results as in progressive re-use mentioned above.

Planning GuideThe ultimate goal of frequency planning in a GSM network is attaining and maintaining ahighest possible C/I ratio every where within the network coverage area. A generalrequirement is at least 12dB C/I, allowing tolerance in signal fading above the 9dBspecification of GSM.The actual plan of a real network is a function of its operating environment (geography,RF etc) and there is no universal textbook plan that suits every network. Nevertheless,some practical guide lines gathered from experience can help to reduce the planningcycle time.

Rules for SFHAs the BCCH carrier is not hopping, it is strongly recommended to separate bands forBCCH and TCH. This has the benefits of: Making planning simpler, Better control of interference.

BCCH (4X3)

1st TCH (3X3)

2nd TCH (2X3)

Progressive re-use

1st TCH (3X3)

BCCH (3X3)

2nd TCH (3X3)

Homogeneous re-use

n channels m channels

BCCH TCH

Guard band

If micro cells are included in the frequency plan, the below band usage is suggested.

Practical rules for 1X3

BCCH re-use plan: 4X3 or 5X3, depending on the bandwidth available and operatingenvironment.

Divide the dedicated band for TCH into 3 groups with equal number of frequencies(N). These frequencies will be the ARFCN equipped in the MA list of a Hoppingsystem (FHI).

Use equal number of frequencies in all cells within the hopping area. The allocationof frequencies to each sector is recommended to be in a regular or continuoussequence. (see planning example)

Number of frequencies (N) in each group is determined by the design loading factor(or carrier-to-frequency ratio). A theoretical maximum of 50% is permitted in 1X3SFH. Any value higher than 50% would practically result unacceptable quality.Loading factor (sometimes termed as fractional load factor) represents the Somecommonly used loading factor are 40%, 33%, 25% etc. As a general guide-line,

For example: mixture of 4-4-4 and 5-5-5 site configurations and loading factor of33%. Then N = 5/(0.33) = 15 frequencies in the MA list. As loading factor has directeffect on the overall network quality and its setting is highly dependent on the RFenvironment, a smaller scale trial is recommended to obtain the necessary data andexperience before larger scale deployment. As a general rule, SFH with 33% loadingis equivalent to a well-planned 4X3 fixed frequency system.

Use same HSN for sectors within the same site. Use different HSN for different sites.This will help to randomize the co channel interference level between the sites.

Use different MAIO to control adjacent channel interference between the sectorswithin a site.

The following example illustrates the above planning guide.

N = (highest non BCCH transceiver count in acell)

(loading factor)

Macro BCCHMicro TCH

Macro TCH(SFH)

MicroBCCH

Bandwidth : 10 MhzSite configuration : mixture of 2-2-2, 3-3-3 & 4-4-4.Loading factor : 33%Multi layer environment (micro & macro co-exist)

The spectrum is split as shown:

A total of 49 channels are available and the 1st and last one are reserved as guard band.Thus, there are 47 usable channels. 12 channels are used in the BCCH layer with 4X3 re-use pattern.Based on 33% loading and 4-4-4 configuration, N is calculated as N = 3 / 0.33 = 9hopping frequencies per cell. Thus, a total of 27 channels are required for the hoppingTCH layer. The remaining 8 channels are used in the micro layer as BCCH.One of the possible frequency plan and parameter settings are outlined in the below table:

ARFCN HSN MAIOSector A 21,24,27,30,33,36,39,42,45 Any from {1,2,…63} 0, 2, 4Sector B 22,25,28,31,34,37,40,43,46 Same as above 1, 3, 5Sector C 23,26,29,32,35,38,41,44,47 Same as above 0, 2, 4

The above MAIO setting will avoid all possible adjacent channel interference amongsectors within the same site. The interference (co or adjacent channel) between sites willstill exist but they are reduced by the randomization effect of the different HSN.

Practical rules for 1X1

1X1 is usually practical in rural area of low traffic density, where the averageoccupancy of the hopping frequencies is low. With careful planning, it can be used inhigh traffic area as well.

BCCH re-use plan: 4X3 or 5X3, depending on the bandwidth available and operatingenvironment.

The allocation of TCH frequencies to each sector is recommended to be in a regularor continuos sequence.

Use different HSN to reduce interference (co and adjacent channel) between the sites. Use same HSN for all carriers within a site and use MAIO to avoid adjacent and co-

channel interference between the carriers. Repeated or adjacent MAIO are not to beused within the same site to avoid co-channel and adjacent channel interferencerespectively.

Macro BCCHMicro TCH

Macro TCH(SFH)

MicroBCCH

12 channels 27 channels

8 channels

Maximum loading factor of 1/6 or 16.7% is inherent in a continuous sequence offrequency allocation. Since adjacent MAIO is restricted, the maximum number ofMAIO permitted would be:

In a 3-cell site configuration, the logical maximum loading factor would be 1/6 or16.7%.

Max MAIO = ½ x (Total allocated channel)

Rules for BBH

All the rules outlined for SFH are generally applicable in BBH. As the BCCH is in thehopping frequency list, a dedicated band separated from TCH may not be essential. Anexample of spectrum allocation is shown below:

BBH channels & micro TCH

Micro BCCH

HSN = 1

HSN = 1

Different MAIO toavoid co-channel

Non adjacentMAIO to avoid

adjacent-channel

HSN = 1