You are on page 1of 32

..

2419 4 8,000 64 1983 AMPS (Analog Advance Mobile Phone Service) 824-894 FDMA - Frequency Division Multiple Access 1990 GSMGlobal System for Mobile Communication TDMA-Time Division Multiple Access 890-960 1991 IS 54 - Interim Standard - 54 TDMA 824-894 1993 IS-95 CDMA 824-894 1,850-1,980 AMPS 1G 1G 2G 9 TDMA IP CDMA - Code Division Multiple Access 2G 3G Universal Mobile Telecommunication Systems (UMTS)

CDMA (wideband) WCDMA 3G CDMA CDMA2000 IS95 (HDR-High Data Rate) 1G 2G GPRS-General Packet Radio Service 9.06, 13.4, 15.6 21.4 GPRS 3G EDGE-Enhanced Data Rate for GSM Evolution 3G 384 2 Licensed Spectrum GSM 890-960 10 700

Adaptive signal - AMPS GSM IS54 IS95 DCS1900 1983 1990 1991 1993 1994 FDMA TDMA/FDMA TDMA/FDMA CDMA TDMA/FDMA WCDMA 824-894 890-960 824-894 824-894 1850-1980 1840-1990 1885-2025 2100-2200 , , , , , , , , ,

WCMA (CDMA200 2000 0) IMT2000

1. 2. 3. 4. 3G WAP - Wireless Application Protocol WAP 3G

GSM SYSTEM IN GOUND


The Global System for Mobile Communications (GSM) is the most successful cellular system in operation today. It is gaining worldwide acceptance and is being adopted as a standard by many countries. As GSM continues to grow, an ever-increasing number of manufacturers, network operators and repair organizations will become involved with the measurement and characterization of GSM equipment. Gaining a good understanding of GSM measurement is essential for all those involved in the testing of GSM equipment. Understanding the factors which drive the need to test, and the way in which specialized GSM measurements are carried out, will help those involved appreciate why testing is required and how GSM measurements have been designed to ensure high levels of system performance. This presentation looks at GSM measurements from the perspective of mobile test. In the first section, we look at factors, which create the need to measure GSM equipment. We investigate the balance created by the need to maintain reliable radio links for individual users, while maintaining high levels of overall system performance and low levels of interference for non-GSM users. In the second section, we look in detail at the measurements needed to verify a mobile is capable of supporting a reliable level of service. In the final section, measurements that characterize the amount of interference generated by a mobile are described

Agenda

T h e B a s is F o r G S M M e a s u re m e n ts
M a in ta in in g th e R F L in k I n t e r fe r e n c e G e n e r a t io n

1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

1. GSM Cells

G S M C e lls

1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

GSM is a family of digital cellular systems. GSM900 operates at 900MHz, DCS1800 at 1800MHz and the proposed PCS1900 at 1900MHz. Each of these GSM systems share modulation, channel coding, system 9

10

architecture and signaling. In fact, the only noticeable differences between the systems are their operating frequencies, channel numbering and transmitter power levels. The comments in this presentation apply equally to all GSM systems. The nature of the GSM system determines the types of measurements, which need to be made on mobiles. The GSM radio spectrum is divided in three ways: spatially into cells, by frequency into channels (FDMA) and by time into timeslots (TDMA). Blocks of spectrum are allocated to each cell. Neighboring cells use different channels, while distant cells can re-use channels. Within each cell, users are assigned an FDMA channel to share with seven other users by TDMA. Speech and control information is digitally encoded onto the radio frequency (RF) carriers using digital modulation to form a link between each mobile and a base station.

2. GSM Measurement Printcipals

G SM M easurem ent P rincipals


M aintain H igh Q uality R adio Link U nder All Expected C onditions
e Me Me
M

MeM

D ont Interfere w ith O ther G SM Links or O ther U sers of the R F Spectrum


Sh e a re ar S h S hare Share

100 01 50 5.p pt 07-9 8

10

11

GSM is probably the most complex cellular system in operation. Despite its complexity, there are two guiding principles, which lead to most GSM specifications and measurements. As a cellular telephone system, GSM is designed to provide a communication link for its users as they travel around the coverage area. In many situations this area will span entire continents. The overriding concern for an individual subscriber is the quality of their phone link. Are they experiencing dropped calls, poor voice quality, problems in some countries or areas? Or, does GSM work perfectly whenever they use it? Purely selfish considerations drive a whole series of GSM tests, many of which are described in this presentation. However, these measurements alone would not guarantee a working system. Other equally important considerations balance the design of the system and lead to more demands on mobile performance. Mobiles should operate effectively without interfering with other GSM users, other GSM networks or other users of the radio spectrum.

3. GSM MEASUREMENT BASIC

M a in ta in in g th e L in k Just E nough P ow er
oss a th L P

T a lk T im e

E n o u g h P o w e r to O v e r c o m e P a th L o s s T o o M u c h P o w e r D r a in s B a tte r y
5 1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

11

12

It is easy to see why lots of output power is an advantage in any radio system. Power overcomes path loss and makes fewer demands on receiver sensitivity. Power causes problems as well. In portable equipment power drains batteries, and in cellular systems, power stops cells from being small, and creates interference in adjacent areas. A balance is required, and just enough output power is the answer. GSM has tight specifications for a phone's output power. Outputs are variable too. Transmitters can be commanded to produce more or less power, as conditions require 4. Maintaining the link, Just Enough Power 4. Maintaining the link, Just Enough Power

M a in t a in in g t h e L in k Just E nough P ow er
P a th Los s

T a lk T im e

E n o u g h P o w e r to O v e rc o m e P a th L o s s T o o M u c h P o w e r D r a in s B a tte ry
5 1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

It is easy to see why lots of output power is an advantage in any radio system. Power overcomes path loss and makes fewer demands on receiver sensitivity. Power causes problems as well. In portable equipment power drains batteries, and in cellular systems, power stops cells from being small, and creates interference in adjacent areas. A balance is required, and just enough output power is the answer. GSM has tight specifications for a phone's output power. Outputs are variable too. Transmitters can be commanded to produce more or less power, as conditions require. 5. Modulation Quality

12

13

M a in ta in in g th e L in k M o d u la tio n Q u a lity
Q

I Q I

R e c e iv e d
Q I

T ra n s m itte d
6

In te rfe re n c e
1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

Producing enough output power to overcome radio path loss is just one of the ingredients required for a reliable link. The quality of the signal being produced is equally important. The objective of establishing and maintaining the radio link is to transfer information from one end to the other. GSM uses 0.3GMSK (Gaussian Minimum Shift Keying) to modulate digital information onto the RF carrier. The quality of the transmitted modulation directly affects the quality of the radio link. This is particularly evident when considering the effect of the interfering signal, illustrated on the slide. In this example, the interference is coming from a mobile on a call with a distant base station. Cellular systems use spatial frequency re-use, so that while neighboring cells use different sets of RF channels, distant cells can re-use the same channels. Assigned to the same RF channel and timeslot, the signal from the interfering mobile is able to combine with that coming from the nearby mobile, making reception at the base station more difficult. For the base stations receiver, this is co-channel interference To understand the effect of the interfering signal, we need to understand 0.3GMSK modulation. 0.3GMSK represents one's and zeros by 90-degree changes in carrier phase, relative to the carrier center frequency. Absolute phase states are unimportant. Information is conveyed only by phase changes. A continuous stream of one's would be represented on the I/Q plane by a steadily rotating vector, moving 90 degrees for each bit period. A stream of zeros would be similarly represented by the vector moving in the opposite direction. Any binary pattern can be represented by the vector making 90 degree phase changes in either direction. The Gaussian filter applied to the data prior to modulation determines the exact phase trajectory for any data sequence. If, instead of generating a perfect phase trajectory, the transmitter introduces a phase error, the phase at any particular instant might be represented by the solid arrow. The ideal phase is represented by the dashed arrow. The dashed circle indicates the area of uncertainty introduced by the error. At the receiver, the 13

14

transmitted modulation combines with the interferer. The vectors sum, producing a much larger area of uncertainty. With 90 degree phase changes representing one's or zeros, it is easy to see how the combination shown could result in misinterpretation by the receiver causing bit-errors. The sketch to the right is shown for comparison. If the transmitted modulation had been perfect, the effect of the interferer would have resulted in a much smaller area of uncertainty, and as a consequence, a much lower likelihood of bit-errors. Poor quality modulation reduces the ability of the receiver to cope with noise and interference. 6.Receiver Performance

M u ltip a th & D o p p le r

M a in ta in in g th e L in k R e c e iv e r P e rfo rm a n c e
U sa b le R ange S e n sitiv ity A d ja c e n t T im e slo t B lo c k in g & S p u r io u s A d ja c e n t & C o -C h a n n e l

1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

GSM transmitters could be perfect, enough power, ideal 0.3GMSK phase trajectory, but still the characteristics of the radio environment would conspire to test receivers to their limits. As we saw in the last slide, receivers need to perform well in the presence of co-channel interference. There is also the possibility of interference from adjacent channels. Other users of the radio spectrum such as radar and TV transmitters can generate large sources of interference, placing demands on receiver. The most common challenges facing the receiver result from radio propagation. The variation in path loss as the mobile moves very close to, or a long way from, the base station tests the receiver's dynamic range, sensitivity and linearity. Complex multipath fading and Doppler need to be simulated in order to verify receiver performance.

7. Interfere With Other Users

14

15

D o n t In te rfe re W ith O th e r U s e rs T im e
A m p litu d e
3 1 0 1
2 2

7 5 4 6

T im e slo t
S p u rio u s

F requency
O u tp u t R F S p e c tru m

ARFCN
P o w e r V e rs u s T im e

1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

GSM specifications need to be balanced. As we have seen, the system needs to be designed and tested so that a high quality radio link is maintained, but not at the expense of interference generation. The TDMA and FDMA outputs from GSM transmitters must stay within their allocated portion of spectrum. Energy should not leak outside the designated timeslot or into adjacent frequency channels (ARFCN) where it could affect other GSM users. Any spurious outputs might interfere with other users of the radio spectrum. 8.The Basis for GSM Measurements

The Basis for G SM M easurem ents


R F M easurem ents Signaling and Protocol Tests Extrem e C onditions Test M ethods Type Approval Lim its

G SM 11.10

ETSI
9 10001505.ppt 07-98

15

16

The GSM system has been designed by the Special Mobile Group (SMG) working under the European Telecommunications Standards Institute (ETSI). The GSM system is described by a large number of 'core' specifications. GSM 05.05, for example, describes the design objectives for the RF portions of mobiles and base stations. Specifications like 05.05 have been carefully designed to appropriately make the trade-offs described in this presentation. There are other trade-offs and considerations too. Specifying and controlling mobile performance very tightly makes the job of designing networks easier, but makes mobiles more expensive to manufacture and test. Poorly specified mobiles may be inexpensive, but make networks unreliable and difficult to design. GSM 05.05 and the many other core specifications have been used to build a type-approval specification for mobiles: GSM 11.10. This document lists test methods, limits and procedures for mobiles. While RF measurements play an important part in GSM 11.10, much of the document is concerned with protocol and signaling test. Hewlett-Packard has been involved for many years, working as members of the ETSI committees formulating the test methods necessary to qualify GSM equipment. Many Hewlett-Packard products have been developed in parallel with the drafting of GSM specifications, and are designed specifically to meet the needs of GSM measurement. While GSM 11.10 can be used as a basis for most GSM measurements, performing all the tests in the specification can take many weeks, with equipment which may cost millions of dollars. This level of testing is necessary to type-approve a mobile, or test its design. In manufacturing, incoming inspection and repair, testing focuses not on the mobile's design, but on characteristics which vary from phone to phone, or as a result of failure. In these situations mobile test can be performed in seconds, often with single dedicated GSM testers such as the HP 8922F/H GSM MS Test Sets. Even though test times are dramatically reduced, test methods can be traced to the procedures in GSM 11.10. 9.Speak the Same Language

16

17

S p e a k th e S a m e L a n g u a g e

L in k A c c e s s P ro to c o l
L ayer 3 L ayer 2 L ayer 1
11 1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

L ayer 3 L ayer 2

GSM is becoming a truly global cellular system, being used in all world regions. Equipment is designed and manufactured in many countries. Interoperability is a key feature of GSM. Base stations from Sweden, mobiles from the UK, being used in Australia, whatever the country, equipment will work together. A single set of agreed GSM standards and type-approval for mobiles helps to guarantee the principles of GSM are maintained. In the following slides, we look in detail at some of the RF specifications of GSM, but these are just part of the overall picture. Much of the hidden complexity of GSM is in the form of protocol; a signaling standard which allows equipment from different manufacturers to operate effectively together. GSM mobiles all speak the same language. The interchange of information between mobile and network is enormous, and includes location update, authentication, placing calls, channel allocation and Tx power control. To guarantee conformity, signaling test forms a major part of type-approving a mobile and verifying its firmware design. Unlike RF hardware, GSM signaling firmware can be reproduced without error. Once a mobile has been type-approved, with one unit having been shown to operate satisfactorily, each copy will have good firmware, removing the need for further test. 10. Mobile Station Measurement Reports

17

18

M o b ile S ta tio n M e a s u r e m e n t R e p o r ts
U p lin k S A C C H

BCH BCH

TCH
T C H : R X Q ual 0 < 0 .2 % B E R 1 0 .2 t o 0 .4 % 2 0 .4 t o 0 .8 % 3 0 .8 t o 1 .6 % E t c .
12

BCH

T C H , H o m e & A d ja c e n t B C H R X L evel . 2 7 -8 4 to -8 3 d B m 2 6 -8 5 to -8 4 d B m 2 5 -8 6 to -8 5 d B m E tc .
1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

Uplink Slow Associated Control Channel (SACCH) measurement reports provide a "mobile's eye view" of the GSM system. Each mobile measures the power of the serving cell and adjacent cells Broadcast Channel (BCH). Mobiles also measure the signal strength and quality being received on the Traffic Channel (TCH). Measurement of Received Level (RXLev) and Received Quality (RXQual) are reported back to the serving cell base station over the SACCH. The RXLev numbers can be mapped directly onto signal levels in dBm. The RXQual reports can be mapped directly onto bit-error ratio. Measurement reports are used by the network to drive a number of processes. If the TCH RXLev indicates a low power level is reaching the mobile, the base station output power can be increased to compensate. If the RXLev from an adjacent BCH is higher than the serving cell, a hand-over to the stronger cell would improve link reliability. If the RXQual looks poor, but RXLev is high enough, an interfering signal could be disturbing the link. The base station could assign the mobile to another ARFCN or to a hopping call. If RXLev and RXQual reports are inaccurate, the network will respond incorrectly to the mobile's situation. Low, or channel-dependent RXLev reports might cause unnecessary hand-overs, high reports might lead to late hand-overs and dropped calls. Verifying the measurement reports coming from a mobile is an essential part of checking its overall performance. Measurement reports can be displayed and checked easily by GSM Test Sets. 11. Tx Levels

18

19

T X L e v e ls
P a th L o s s

TCH

D o w n lin k S A C C H
T x L evel 5 6 7 * 14 15
13

P ow er dB m 33 31 29 * 16 13
1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

Mobile transmitter output power is controlled in 2dB steps or Tx Levels. The mobile's Tx Level is controlled by the base station which sends commands on the downlink SACCH. The reliability of the link dep ends on the mobile producing the correct power for each Tx Level 12. Tx Power
T X Pow er
8 A v e r a g e P o w e r D u r in g U s e fu l P a r t = X d B m + /3dB 8 M a x im u m O u t p u t f o r P o w e r C la s s = X d B m + /- 2 dB Pow er B 8 I n s t a n t a n e o +u 1 . 0Pdo w e r + / - 1 d B R e l a t i v e t o A v e r a g e X s
- 1 .0 d B 1 4 7 U s e fu l B its 5 4 2 .8 s

T x L evel 5 6 7 * 14 15

P ow er dB m 33 31 29 * 16 13

T im e

14

1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

Measuring the output power from a GSM transmitter is complicated by the TDMA multiplexing scheme. Each frequency channel is shared among eight users by TDMA. The mobile transmitter turns-on only during its active timeslot. During the central or useful part of the TDMA burst, when data is being transmitted, the mobile has to control its transmitter output to within +/-1dB relative to the mean value. The absolute output power is defined as the average during the useful part of the burst. The power measurement is designed to ignore the effects of overshoot and undershoot at the ends of the burst, and concentrate on the portion where data is being transmitted. The average power must be within +/-3dB of the 19

20

value specified for the selected Tx Level. The maximum output power available from a mobile depends on its power class. At the maximum output power for the class, the specification is tightened to +/-2dB. The GSM power measurement can be made conveniently with tuned or wide-band power meters, provided they are capable of averaging only during the useful portion of the burst. Conventional peak power meters, not specifically designed for GSM signals, will have their readings affected by the overshoot or undershoot of the burst. Thermal power meters, or other devices with long term averaging properties, can sometimes be used with TDMA systems by taking into account the 1:8 duty cycle of the signal being measured. This technique is generally not recommended for GSM signals. The relatively slow rising and falling edges of the burst, and the variation from phone to phone in pulse rise shape and burst length can cause large changes in actual duty cycle. This, combined with the effects of overshoot and undershoot can lead to poor measurement results. Measuring the output power from a GSM transmitter is complicated by the TDMA multiplexing scheme. Each frequency channel is shared among eight users by TDMA. The mobile transmitter turns-on only during its active timeslot. During the central or useful part of the TDMA burst, when data is being transmitted, the mobile has to control its transmitter output to within +/-1dB relative to the mean value. The absolute output power is defined as the average during the useful part of the burst. The power measurement is designed to ignore the effects of overshoot and undershoot at the ends of the burst, and concentrate on the portion where data is being transmitted. The average power must be within +/-3dB of the value specified for the selected Tx Level. The maximum output power available from a mobile depends on its power class. At the maximum output power for the class, the specification is tightened to +/-2dB. The GSM power measurement can be made conveniently with tuned or wide-band power meters, provided they are capable of averaging only during the useful portion of the burst. Conventional peak power meters, not specifically designed for GSM signals, will have their readings affected by the overshoot or undershoot of the burst. Thermal power meters, or other devices with long term averaging properties, can sometimes be used with TDMA systems by taking into account the 1:8 duty cycle of the signal being measured. This technique is generally not recommended for GSM signals. The relatively slow rising and falling edges of the burst, and the variation from phone to phone in pulse rise shape and burst length can cause large changes in actual duty cycle. This, combined with the effects of overshoot and undershoot can lead to poor measurement results. 13.Modulation Quality

20

21

M aintaining the Link M odulation Q uality


Q

I Q I

R eceived
Q I

Transm itted
15

Interference
100 01 50 5.p pt 07-9 8

As we discovered in the first part of the presentation, transmitter modulation quality must be good enough to maintain a low-error digital radio link in the presence of a variety of interference sources. Nearperfect modulation would be ideal, but require very complex and expensive transmitter design. The balance between cost-effective design and the desire for high quality modulation has been used to develop the GSM specification. The phase trajectory across a GSM burst is allowed to have a peak error of 20 degrees and an rms error of 5 degrees. In addition, the whole I/Q pattern is allowed to rotate slightly as a result of center frequency error. Frequency error should be less than 90Hz for GSM900 mobiles and 180Hz for DCS1800 mobiles. Measuring 0.3GMSK phase and frequency error has required new test methods to be developed. Hewlett-Packard was able to contribute to this pioneering work by developing the first and most accurate commercial equipment and algorithms for measuring 0.3GMSK phase and frequency error 14. phase and frequency error

21

22

M e a s u r in g P h a s e a n d F r e q u e n c y E rro r S a m p le A c tu a l P h a s e T r a je c to r y
Q I +180 +90 0 -9 0 -1 8 0

D e m o d u la t e

Q I +180 +90 0 -9 0 -1 8 0

1010011000111001

C o m p u te P e r fe c t T r a je c to r y

16

1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

Before the process of calculating phase and frequency error can begin, a sampled record of the transmitter's phase trajectory during one TDMA burst is captured. There are a number of techniques for obtaining this phase trajectory. Hewlett-Packard's patented method uses high-speed sampling and digital signal processing to ensure high accuracy and repeatability. The incoming RF burst is down-converted and digitized directly. The sampled data is processed to extract the phase trajectory. Obtaining the phase trajectory using digital processing avoids accuracy and repeatability problems often associated with techniques using analogue I/Q demodulators prior to digitizing. To aid understanding, the phase trajectory must be thought of as being relative to the phase of the carrier center frequency. Streams of ones will cause a phase decrease of 90 degrees each, while zeros cause 90 degree phase increases. The rounding caused by Gaussian filtering stops the phase trajectory from meeting its 90 degree target points. The sampled phase trajectory is processed to produce a demodulated data pattern. The data pattern is used by the digital signal processor to synthesize a perfect phase trajectory. 15.phase and frequency error
M e a s u r in g P h a s e a n d F r e q u e n c y E r r o r
+ 1 8 0 + 9 0 0 -9 0 -1 8 0 S u b tr a c t P er fe c t T r a je cto r y F r o m + 2 0 + 1 0 0 -1 0 -2 0
1 7 1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

Q I

A c tu a l

22

23

Overlaying the sampled trajectory with the perfect trajectory highlights the imperfections in the measured modulation. Subtracting the two wave-forms produces a plot of phase error at each point across the TDMA burst. Overlaying the sampled trajectory with the perfect trajectory highlights the imperfections in the measured modulation. Subtracting the two wave-forms produces a plot of phase error at each point across the TDMA burst.

16. phase and frequency error

M e a s u r in g P h a s e a n d F r e q u e n c y E r r o r
+20 +10 0 -1 0 -2 0 F r e q u e n c y E r r o r (m e a n
/ )

P h a se E rror

G SM 900 < 90 H z D C S1800 < 180 H z

P eak < 20 deg. R M S < 5 deg.


18 1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

The phase error trace has two ingredients: slope and roughness. A best-fit straight-line is used to calculate the slope. The slow change of phase across the burst, shown by the straight-line, is removed from the phase error calculation and expressed separately as frequency error. The remaining phase error trace is summarized by calculating its peak error and rms error. The entire process of sampling a GSM burst, calculating its phase trajectory, demodulating, producing a perfect trajectory and calculating frequency error, peak and rms phase error, can be carried out using highspeed digital signal processors in one second or less. 17. Receiver Performance

23

24

M u ltip a th & D o p p le r

M a in ta in in g th e L in k R e c e iv e r P e rfo rm a n c e
U s a b le R ange S e n s itiv ity A d ja c e n t T im e slo t B lo c k in g & S p u r io u s A d ja c e n t & C o -C h a n n e l

19

1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

GSM receivers have to operate effectively in a variety of complex radio environments. The receiver has to meet specification when subjected to multipath and Doppler fading, low signal levels, high signal levels, interference from other GSM transmitters and interference from other users of the radio spectrum. For the mobile to work effectively, and provide a good quality of service, it must be able to demodulate the incoming 0.3GMSK signal with the minimum of bit-errors. GSM speech channels carry voice information encoded as type Ia and Ib bits with error correction and type II bits with no error correction. Each bit-error introduced by the receiver misinterpreting the 0.3GMSK modulation will affect the quality of the decoded speech to some degree. Errors occurring in type II bits will directly affect speech quality. Small numbers of error in the type Ia and b bits will be corrected. Larger numbers will overcome the effectiveness of the error correction algorithm, leading to very noticeable speech distortion or drop-outs. 18. Bit Error Ratio

24

25

B it E r r o r R a t io M o b ile
PO W ER

Test Set
DeM od S peech Coder M od

PRBS

L o o p -B a ck
M od DeM od C o m p a re

Z e r o B it E r r o r s F o r N o r m a l S ig n a ls

B it E r r o r R a tio

20

1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

Bit-error-ratio is used to assess the effectiveness of the receiver in the presence of a variety of simulated inputs. In analogue radios it is usually possible to measure the performance of the receiver by listening to, or measuring, the demodulated audio signal. In a digital system, such as GSM, speech coding and error correction makes it almost impossible to accurately assess the performance of the receiver from the quality of the decoded audio signal. It is always necessary to look at the demodulated data before it goes to the speech decoder. This presents some practical measurement problems. The demodulated data bits are generally not available outside the phone to be measured. They may only exist within an integrated circuit and be impossible to probe. To solve this problem, and allow these data bits to be measured, the GSM specifications require all mobiles to have a special 'loop-back' mode. The mobile can be commanded to loop-back demodulated data bits by sending a command on the downlink SACCH from a GSM Test Set. Loop-back mode can only be activated when the mobile has a GSM test SIM installed. Once the mobile is looping back demodulated data, bit error ratio measurements can be made. To give confidence that bit errors are due to the phone's receiver, a high quality 0.3GMSK signal is produced by the GSM Test Set, emulating a GSM base station. The phone camps, a call is set up and the mobile is commanded into loop-back mode. The speech data, normally encoded on the TCH, is replaced within the Test Set by a Pseudo Random Binary Sequence (PRBS). The PRBS is modulated onto the downlink TCH with the correct midamble and frame / multiframe structure. The mobile receives and demodulates the PRBS, loops it back, re-modulates it onto the TCH uplink. Within the Test Set, the TCH is demodulated to reveal a delayed version of the PRBS. Correlating and comparing with the original PRBS allows any bit-errors to be determined and counted. Error bits can be expressed as a percentage of the total number of bits being sent. The demodulator in the Test Set is designed to produce zero bit-errors with mobiles, which meet the GSM transmitter modulation specifications for phase and frequency error. Under normal signal conditions, the

25

26

mobile and Test Set produce no bit errors. As interference or other stresses are imposed on the mobile's receiver, bit errors are counted, indicating the performance 19. Receiver Static Sensitivity

R e c e iv e r S ta tic S e n s itiv ity M o b ile


PO W ER

Test Set
T y p e II T y p e Ib
B it E r r o r R a tio

D o w n lin k P o w e r

< 2 % T y p e II B E R a t -1 0 2 d B m
21

< 0 .1 % T y p e II B E R a t -1 5 d B m
1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

As the mobile moves around the cell, the received downlink signal level varies due to changes in the propagation path loss. Base stations use the RXLev reports from the mobile as an indication of the power being received. When the reports show received power is low, the base station can increase its output power in a particular timeslot. As the mobile moves closer, the base station can reduce its output power in the timeslot being used. This process is called dynamic power control, and is available over a 30dB range. Even with dynamic power control, the variation in signal level experienced by the mobile is huge. GSM specifications for a hand-held mobile require it to operate with less than 2% bit-error-ratio at signal levels of -102 dBm, and less than 0.1% at signal levels as high as -15 dBm. The low signal level of -102 dBm tests the mobile's performance at the edge of a cell, some distance from the base station. The high signal level case at -15 dBm simulates situations where the mobile is close to a base station. The effects of error correction present only on type Ia and Ib bits can be seen in the graph. Type II bits, with no error correction, show a gradual increase in bit-error ratio, as signal conditions become more demanding. The error correction algorithms of the Ib bits correct a certain number of errors. When the signal quality gets too poor, the correction algorithms break-down more quickly, giving a more sudden increase in bit-error ratio. Type II bits give the best indication of the performance of the receiver since the effect of any bit-errors are not masked by error correction algorithms. 20. Receiver Sensitivity under Fading Conditions

26

27

R e c e iv e r S e n s it iv ity U n d e r F a d in g C o n d it io n s R F C h a n n e l S im u la to r
D o p p le r 6 o r 1 2 p a th s
PO W ER

A tte n

R a y le ig h

D e la y

T est Set

R a y le ig h 3 -D S ta n d in g W a v e s M o b ile

O n e P a th
22

D o p p le r

1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

In urban or hilly areas, the radio path between base station and mobile is usually not direct, or line-ofsite. The signal received by the mobile is reflected many times from buildings or rock-formations. Just like ripples or standing waves, which form in a water container, the radio signals form three dimensional standing wave patterns in space. This effect is called Rayleigh fading. Signal variations of up to 40 dB are possible in very short distances as the mobile moves from areas of constructive interference to destructive interference or nulls. The rate at which the mobile moves through the standing wave pattern depends on its speed of motion. The mobile's speed also results in Doppler frequency shifts of the received signal. These Rayleigh and Doppler effects are modeled by a single path of an RF channel simulator such as the HP 11759C. In a typical radio environment, several major paths from a base station, with different path delays and attenuation, will combine in an obscured area to form separate overlaid Rayleigh patterns, adding together in mean power, but each with a different information timing or delay. Six or twelve Rayleigh / Doppler paths are combined in the RF channel simulator to emulate these conditions. Each path has a separately programmed delay and attenuation, described by the GSM standard fading models. The GSM specifications list a variety of propagation models simulating rural, urban and hilly conditions, with a range of mobile speeds. For each condition, the mobile must meet a specified bit-error-ratio performance.

21. Receiver Performance with Interference

27

28

R e c e iv e r P e r f o r m a n c e w it h I n t e r f e r e n c e

POW ER

C o -C h a n n e l

A d ja c e n t C h a n n e l

B lo c k in g

A d ja c e n t T im e s lo t W a n te d
23

I n t e r m o d u la t io n S p u r io u s R e s p o n s e
In te r fe r e
1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

GSM mobiles need to be able to operate effectively in the presence of interference from other GSM channels and other non-GSM users of the radio spectrum. Co-channel interference, arriving at the mobile's receiver, could come from distant base stations using the same frequency channel and timeslot as the downlink. Adjacent channel interference could come from nearby base stations, using adjacent ARFCN, or from other downlink channels in the same cell. Adjacent channel interfering signals are often at a higher power level than the wanted signal, and test the selectivity of the mobile's receiver. Blocking tests the ability of the phone's receiver to operate in the presence of large signals within the GSM band, but not in the channel being used. Blocking signals might come from other GSM transmitters very close to the mobile, for example those being used by another network operator, or from non-GSM signals in the same band, such as the analogue NMT or TACS cellular systems. Intermodulation tests the phone's response to two interfering signals in the GSM band. The signals are chosen so that when their frequencies combine in any non-linear receiver components, they mix to form signals coinciding with the channel in use. Spurious response investigates the effect of large signals outside the GSM band, for example from radar, private mobile radio or TV transmitters. Non-GSM signals are injected from 100kHz to 12.75GHz. The adjacent timeslot test looks at the effects of the base station's downlink dynamic power control on the mobile's receiver. The base station will increase the power in certain timeslots by up to 30dB to reach distant mobiles. A mobile positioned close to the base station may be expected to receive on a timeslot, which is 30dB lower than the two adjacent slots. The phone must be able to demodulate the information in the wanted timeslot without being affected by the higher signal levels in the adjacent slots. In each of these tests, interference is injected into the downlink, and the phone's bit-error ratio monitored. 28

29

D o n t I n t e r f e r e W it h O th e r U s e rs T im e
A m p litu d e
3 1 0 1
2 2

7 5 4 6

T im e s lo t
S p u rio u s

F requency
O u tp u t R F S p e c tr u m

ARFCN
P o w e r V e r s u s T im e

25

1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

In this section, we look in more detail at some of the measurements designed to ensure that mobiles don't interfere with other GSM or non-GSM, users of the radio spectrum. Interference can be generated in the form of spurious emissions, or by the mobile's transmitted TDMA burst not fitting in the designated timeslot. Output RF spectrum checks for interference spilling into adjacent ARFCN. Power versus time checks for interference with other timeslots. 22. The GSM Burst

T h e G S M B u rs t

T im e

A m p litu d e

F req u e n c y
26 1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

Before examining the output RF spectrum and power versus time tests it is useful to understand some of the characteristics of the GSM TDMA burst. The diagram shows a three dimensional, enlarged representation of one of the bursts shown on the previous slide. Moving along the time axis, the burst rises in amplitude as the 29

30

transmitter turns on at the start of the burst, maintains a constant amplitude during the useful part, and then falls. The useful part of the burst is the central portion, when 0.3GMSK modulated data is being transmitted. As we saw in earlier sections, the amplitude of the burst is controlled to within +/- 1dB during this portion. The shape in the frequency / amplitude plane is mainly caused by the Gaussian filtered 0.3GMSK modulation. During the useful part of the burst, its arched profile changes little with time. The most noticeable changes to the arched shape are near the ends of the burst. Raised amplitude lobes can clearly be seen on either side of the center frequency, at each end of the burst. These lumps are caused by transient energy contained in the amplitude rising and falling edges spilling out into a wider frequency range. At these points, the spectral energy in the rising and falling edges rises above that caused by the 0.3GMSK modulation. 23.Power versus Time
P o w e r V e rs u s
P o w e r
+ 4 .0 d B + - 6 .0 d B 1 .0 - 1 .0 d B d B - 6 .0 d B

T im e

- 3 0 .0 X d B

d B

X d B O n / O ff R a tio t y p ic a lly 5 9 d B , b u t d e p e n d s o n P o w e r a n d R a d io ty p e

- 3 0 .0 X

d B d B

1 0 S

8 S

1 0 S

1 0 S

8 S

1 0 S

T im e
2 7 1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

Slicing the TDMA burst along the center frequency, in the time / amplitude plane reveals its power versus time profile. GSM specifications demand that the mobile's amplitude profile fall within the mask shown. The uppermost portion of the burst, from -30 dB to +4 dB is specified by a standard mask for GSM mobiles of all types, and is the same for all transmitted power levels. The lower corners of the rising and falling profiles, referred to as the on / off ratio, are controlled by a more complex specification. The on / off ratio specification varies with phone type and transmitted power level. Power versus time can be measured conveniently with a time-gated spectrum analyzer, set to zero frequency span and tuned to the channel center frequency. The resolution bandwidth is chosen to be narrow enough to give a signal-to-noise performance necessary to display the burst's full dynamic range. The resolution bandwidth also needs to be wide enough not to distort the profile by slowing down transitions or displaying ripple induced by modulation during the useful part of the burst. GSM Test Sets frequently use digital sampling and signal processing to make high quality measurements of the upper 30 to 40dB of the burst. Some Test Sets employ built-in spectrum analyzers to measure the high dynamic range on / off ratio.

30

31

24. Output RF Spectrum

O u tp u t R F S p e c tr u m

D u e t o M od u la t ion

D u e t o R a m p in g
28 1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

Cutting the burst in the frequency / amplitude plane reveals different cross-sections at various points along the time axis. During the useful part of the burst, the cross-section stays constant and reveals the spectrum due to modulation. At the ends of the burst, the cross-section is modified by the energy contained in the rising and falling amplitude ramps. The cross-section here reveals the spectrum due to ramping. GSM specification limits are recorded as power levels at tabulated offsets from the channel center frequency. Interpreting these specifications as a frequency mask gives the shapes shown in the slide. 25. Measuring Output RF Spectrum\

31

32

M e a s u r in g O u tp u t R F S p e c tr u m

~ ~ ~

Log Am p

D e te c to r

R e s o lu t io n B a n d w id t h F ilt e r

V id e o G a te

C e n te r F re q + /- O f f s e t

29

1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

Output RF Spectrum can be one of the most difficult GSM measurements to visualize or understand. Matters are further confused by the fact that most pieces of measurement equipment display Output RF Spectrum traces as amplitude versus time at a particular frequency offset, not as most expect, amplitude versus frequency. The measurement is made using a time-gated spectrum analyzer, set to zero frequency span and tuned to the channel center frequency +/- an offset. The offset frequencies allow the analyzer to take amplitude versus time slices from the measured bursts at the GSM specified frequency offsets. A reference measurement begins the sequence by establishing the amplitude at the center frequency (zero offset). The reference measurement is used to convert the results at each offset to relative or dBc values. Within the analyzer, the resolution bandwidth filter defines the width of each time domain slice taken from the burst. The log amp improves display dynamic range; the detector converts the down-converted and filtered input signal to a video wave-form suitable for display. The video-gate provides synchronization by selecting the correct portion of each TDMA frame for display and post-processing. The ripple displayed during the center of the burst is an expected by-product of the measurement technique. The instantaneous input frequency will be varying approximately +/-67kHz due to the 0.3GMSK modulation and data pattern. Since the resolution bandwidth filter is tuned to select a narrow slice of the burst, the instantaneous input signal will move backwards and forwards across the selected frequency many times during the burst. The energy gathered by the filter during each crossing depends on the exact data pattern being transmitted. This produces the random ripple pattern shown. 26. Output RF Spectrum Due to Modulation

32

33

O u tp u t R F S p e c tr u m D u e to p M o d u la tioS ne c ific a tio n d B c


T u n e d A v e ra g e P o w e r A c ro s s U s e fu l P a rt o f B u rs t

O ffs e t k H z C h a n n e l C e n te r

Pow er T im e

30

1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

The Output RF Spectrum due to modulation is measured by averaging the trace at a particular offset frequency across the useful part of the burst. The average is compared to the value obtained at the channel center frequency to provide a dBc result. Results at individual offsets are compared with the GSM specification limits. Stepping away from the channel center in very small offsets would reveal the smooth arch shaped modulation spectrum shown.

27. Output RF Spectrum Due to Ramping

O u t p u t R F S p e c t r u m D u e t o R a m p in g
S p e c if ic a t io n d B c

Tuned Peak Power

O ffs e t k H z C h a n n e l C e n te r

P ow er F req.
31

T im e

1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

The spectrum due to ramping is measured using a similar process. The resolution bandwidth and video post processing are modified to reveal the humped characteristics at the end of the burst. Instead of

33

34

averaging across the trace, as in the modulation case, the highest peak value is used as the result. The results at each frequency offsets can be compared with the GSM specifications. Typically, phones with faster amplitude ramps produce poorer spectrum due to ramping performance. 28. Spurious Emissions

S p u r io u s E m is s io n s
P r o t e c t O t h e r U s e r s o f R a d io S p e c t r u m 9 k H z to 1 2 .7 5 G H z I d le M o d e a n d O n a C a ll

PO W ER

32

1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

GSM specifications call for a variety of spurious emissions tests over a wide frequency range. These tests are designed to protect other users of the radio spectrum from unwanted emissions from GSM mobiles. Specifications vary for mobiles on a call, or in idle mode. Depending on the design of the mobile, conducted and radiated spurious emissions will need to be checked. Testing spurious output over a variety of extreme power supply and temperature conditions can sometimes be revealing. When a mobile's battery voltage droops, the circuitry should switch-off cleanly, rather than getting stuck in an unpredictable mode with unwanted RF outputs. Spurious emissions tests can be made conveniently using spectrum analyzers. Searching over the entire RF spectrum from 9kHz to 12.75GHz can be very time consuming, and require screened RF measurement rooms to keep out other radio energy, which could interfere with the measurement. In some cases, detailed knowledge of the phone's block diagram allows abbreviated measurement procedures, which look only at likely problem areas. 29. GSM Measurement Principles

34

35

G S M M e a s u r e m e n t P r in c ip a ls
M a in ta in H ig h Q u a lit y R a d io L in k U n d e r A ll E x p e c te d C o n d it io n s
e M e Me
M

M eM

D o n t In te r fe r e w ith O th e r G S M L in k s o r O th e r U s e rs o f th e R F S p e c tru m
Sh a re re ha S h a r e S S h a re

33

1 0 0 0 1 5 0 5 .p p t 0 7 -9 8

During this presentation, we have looked at the balance of GSM principles shown on the slide. We have also looked in detail at how some of the important measurements, indicated by these principles, are carried out. It is important for all involved in the manufacture, design, incoming inspection or repair of GSM mobiles to have an appreciation for the theory behind the measurements required by GSM, and how these tests are performed. With this knowledge, it is possible to develop an understanding of how to prioritize and choose important measurements, interpret failures and build effective test and repair strategies.

35

You might also like