Are You in SYNC?

Rapid Evaluation of Telecom Network Synchronization Quality to ITU-T/ETSI Standards

An ANT-20 Application

Rapid Evaluation of Synchronization Quality Using Cost-Effective Analysis Software

As the performance of communications systems increases, so the demands placed upon the underlying network technology also increase. Synchronous networks such as SONET or SDH are designed to cope with the demands of modern communications services as long as their synchronicity can be maintained. Under normal operating conditions, various interference effects lead to deviations from the otherwise rigidly synchronous behavior. Such deviations are apparent as jitter or wander and lead to degradations in transmission quality due to bit errors, slips, loss of data or frequency interference. The synchronization behavior of networks and network elements can be characterized. MTIE and TDEV (see Abbreviations, p. 11) are used for this and both terms are explained in detail in this brochure. The ANT-20 Advanced Network Tester with its Jitter/Wander Option and MTIE/TDEV Analysis Software is the ideal tool for measuring these values. The features of the ANT-20 are listed on the back page of this brochure.

Jochen Hirschinger

Synchronous network architecture

A special synchronization network is set up to ensure that all of the elements in the communications network are synchronous. The network is hierarchically distributed. A primary reference clock source (PRC) controls the secondary clocks for the network nodes (SSU) and network elements (SEC). The inclusion of plesiochronous signals (PDH) within the synchronous system must also be taken into account.

PRC

SSU SEC SEC SEC SSU SEC SEC SEC

PDH

SSU SEC SEC SEC SSU SEC SEC SEC

Example of a synchronization chain (see page 11 for abbreviations)

This type of synchronization signal distribution is also referred to as master/ slave synchronization. The actual synchronization may take place via a separate, exclusive sub-network, or the communications signals themselves may be utilized. Ring structures are also possible. Under ideal conditions, the reference signal from the PRC will be passed on by the next synchronization element in the chain. The clock signal output from the synchronization element is synchronized to

the incoming signal to conform with ITU-T Recommendation G.811 (frequency accurate to 10-11). Clock regeneration in SSUs and SECs is achieved using phase-locked loops (PLL). The control circuit of a PLL basically comprises a phase comparator, a narrow-band filter and a voltage-controlled oscillator (VCO).

PRC
Primary Reference Clock

10-11

Synchronization Supply Unit

Lowpass filter

VCO

Output clock

SEC
SDH Equipment Clock

Accuracy
ETSI ETS 300 462-1 ETS 300 462-3 ETS 300 462-6 ETS 300 462-4

SSU

10-9 (TNC) 10-8 (LNC)

4.6610-6

PDH
Referenceclock PDH Slave Clock

Dj
Phase comparator

1/x
Frequency divider The clock hierarchy

Principle of the phase-locked loop

This circuit is used to pull the output clock to the reference clock. SSUs are designed as clock regenerators, so the filters generally have narrower bandwidths than those in the SECs built in to the NEs. The regenerated clock is therefore of higher quality. Of course, the accuracy of such regenerators is limited. They also introduce degradations due to intrinsic characteristics. As a result, the number of synchronization units in a chain must be limited. ETSI 300 462 specifies that the longest chain originating from a PRC must not exceed 10 SSUs, with not more than 20 SECs between two SSUs. The total number of SECs in a chain should not exceed 60. If all the incoming (higher quality) clock signals fail or are unsuitable for synchronization, the affected unit switches to hold-over mode. In this situation, an attempt is made to hold to the last correctly received signal as precisely as possible. To be able to do this, the frequency correction values for the last 36 hours are stored along with the current oscillator temperature. These data can be used to control the oscillator, achieving a ten- to hundred-fold improvement in stability compared with a free-running oscillator.

The master clock for the entire network is generated by a PRC. All clock signals can be traced back to the PRC, which is realized using a cesium oscillator with LORAN C or GPS support. SSUs are components that are purely for clock recovery purposes. The SECs are normally built into the NEs.

ITU-T Definition Network PRC SSU SEC G.810 G.825 G.811 G.812

ANSI/ Bellcore T1.101/ GR-253 T1.105/ GR-253 T1.101

G.813 ETS (G.81s) 300 462-5

GR-253

Standards

Evaluation methods
Various standardized methods of evaluation are used to determine the quality of a clock. These are described below.

the interval s by the time interval s itself, since: Df/f % MTIE(s)/s Beispiel: MTIE (1s) = 12 ms Df/f = 12610-6 (12 ppm)

TIE (Time Interval Error)

The TIE value represents the time deviation in the clock signal under test relative to a reference source. The measurement is referred to an observation interval, s. It is usual to arbitrarily set the start of the interval to zero, i.e. TIE(0) = 0. TIE measurement is the basis for other calculations (MTIE, TDEV). ITU-T Recommendation G.813 prescribes that at least 30 samples per second should be taken (with low pass filtering with a 10 Hz cutoff frequency).

MTIE (10s) = 15 ms Df/f = 1.5610-6 (1.5 ppm) A more precise result is obtained using the complete algorithm specified in ETSI 300 462. Here, a variable observation interval s travels through the entire measurement time T with the maximum deviation being recorded (MTIE value for the interval s).

TIE TIE MTIE

Time (t) Observation interval s Measurement time T

TIE Time (t) Observation interval s Measurement time T

Determining the MTIE value

Simplified MTIE algorithm:

Determining the TIE value

MTIE (Maximum Time Interval Error)

MTIE is the maximum time interval error (peak-to-peak value) in the clock signal being measured (compared with a reference clock) that occurs within a specified observation time s. In the simplest case measurement (instantaneous value detection), the starting point for the interval is fixed, and the interval is increased by increasing the measurement time. This already allows relative frequency offset Df/f to be detected and determined approximately by dividing the MTIE value for

Observe all 1 s intervals. Determine the maximum time deviation within each observed 1 s interval (MTIE value for 1 s). Enter the highest value against the 1 s mark in the MTIE graph. Observe all 2 s intervals. Determine the maximum time deviation within each observed 2 s interval (MTIE value for 2 s). Enter the highest value against the 2 s mark in the MTIE graph. Repeat for other time intervals (3, 5, 8 s, etc.). The MTIE calculation is suitable for detecting a frequency offset but does not give any information about the spectrum of the error signal.

TDEV (Time Deviation)

The TDEV value is a measure of the phase error variation versus the integration time, i.e. a statistical value. It is calculated from the TIE samples. Put simply, the standard deviation s(si) is calculated for each point si within a measurement time T for an interval s which travels through the entire measurement time (see under MTIE). The calculated values are averaged over T to obtain the TDEV value for the interval s. The interval s is then increased and the procedure repeated for the new value of s. In contrast with the MTIE calculation, TDEV analysis provides information about the spectral content of the phase variation and the results can be interpreted as indicated in the diagram below:

Simplified TDEV algorithm: Observe all 1 s intervals. Determine the standard deviation s within each interval. Average all values of s over the measurement time T (this gives the TDEV value for 1 second). Enter this value against the 1 s mark in the TDEV graph. Observe all 2 s intervals. Determine the standard deviation s within each interval. Average all values of s over T (TDEV for 2 s). Enter this value against the 2 s mark in the TDEV graph. Repeat for other time intervals (3, 5, 8 s, etc.). The TDEV calculation can be considered as a traveling software filter. The TDEV values for the intervals sx are obtained by digital filtering using a bandpass filter with center frequency 0.42/sx followed by calculation of the r.m.s. value.

Measurement time T TIE ti-1

ti sx Observation interval

ti+1 Time (t)

Short interval s1

Medium interval s2

long interval s3

H(f) 0.42 f S1

H(f) 0.42 S1 f

H(f) 0.42 S3 f

RMS value

RMS value

RMS value

TDEV

s1

s2

s3 s Determing the TDEV value

higher lower frequency components

ADEV (Allan Deviation), MADEV (Modified Allan Deviation)

The ADEV and MADEV calculations are comparable with the TDEV calculation (they are also mathematically related). ADEV and MADEV are not used as often as TDEV but are sometimes employed in analysis, as they give more information about the nature of the impairments.

Buffers are used to compensate for the frequency variations caused by pointer actions in digital switches, synchronous cross-connects or add-drop multiplexers. The MTIE value is useful for configuring the buffer, i.e. the buffer is dimensioned according to the specified limit value for MTIE. If this value is not exceeded it can be safely assumed that no buffer overflows will occur and hence frame slips will be absent. The TDEV, ADEV and MADEV traces cannot be used to dimension the buffers
f0 Data in

Buffer

Interpreting the MTIE, TDEV, ADEV and MADEV curves

ADEV, MADEV and TDEV may yield different results depending on the type of interference signal, as the table below taken from ETSI 300 462-1 shows. As well as the obvious effects due to frequency offset and drift, the typical noise processes encountered in oscillators are also listed. As the table shows, the MTIE calculation is the only method described that can detect the important (and frequently occurring) case of frequency offset. On the other hand, the TDEV calculation gives information about frequency drift or oscillator noise. If, for example, the slope of the TDEV curve corresponds to the square root of s (s-1/2), this indicates phase modulation with white noise.

f0

Data out

but they are useful for assessing oscillator performance. ETSI 300 462-3 specifies MTIE and TDEV masks for all synchronization elements (PRC, SEC, SSU, PDH). These indicate the maximum MTIE or TDEV value for each observation interval. To summarize:

MTIE is a measure of the long-term stability of a clock signal, TDEV a measure of the short-term stability.

Process Frequency offset Frequency drift White noise phase modulation (WPM) Flicker phase modulation (FPM) White noise frequency modulation (WFM) Flicker frequency modulation (FFM) Random walk frequency modulation (RWFM)
Interpretation of MTIE, TDEV, ADEV and MADEV curves to ETSI 300 462-1

MTIE s

TDEV s2 s-1/2 s-0 s1/2 s s3/2

ADEV s s-1 s-1 s-1/2 s0 s1/2

MADEV Possible cause Clock not from PRC s Delay variations due to temperature changes s-3/2 Typical parasitic noise processes s-1 in different types of oscillators s-1/2 s0 s1/2

Practical Synchronization Problems

How are synchronization problems manifest?
A poor quality transmission path is usually the first assumption if a connection is faulty. If a bit error rate test does not indicate the cause, the fault may be a synchronization problem. As a bit error test set operates independently from the network, it cannot detect problems within network elements that are due to badly synchronized clock signals. In SDH networks, pointer analysis can give initial information about a poor quality clock signal, since pointer actions are triggered by wander, among other things. Clear assessment of the clock quality is possible by means of wander analysis. On the other hand, wander simulation can be used to test the tolerance of the network elements to phase variations. Typical manifestations of synchronization problems are:
*

Faulty transmission although degradation of the physical characteristics of the line is not apparent. Sporadic and possibly periodic occurrence of interference (e.g. sync. loss defects).

Why do synchronization problems occur?

Ideally, all the network elements in a SDH network should be synchronized to a universal clock (UTC). Under practical conditions, unavoidable interference effects lead to deviations from this rigid synchronicity. Problems due to physical causes are phase noise and drift in the synchronized clock oscillators and temperature-dependent variations in propagation delay through the transmission path. These result in slow, often periodic variations in phase that are generally referred to as wander. Due to the physical problem of distributing the clock signal, a number of local reference sources (PRCs) are used. Most larger network providers run their own reference clock, with smaller network providers often drawing their reference clock from a neighboring larger network. Despite the high frequency accuracy of the PRCs, system-related problems occur in the following cases as a result:
*

Interaction between several synchronous networks, each of which is fed from a different reference source (PRC). Transition between different technologies, e.g. to / from a SDH network to PDH, ISDN, GSM or ATM (in the access sector). Interruption of the synchronization chain: The element (SSU or SEC) immediately downstream of the interruption either receives the reference clock from another source or it switches to holdover mode The local source has reduced frequency accuracy (e.g. 4.6 10-6 for a SEC instead of 10-11).

When should measurements be made?

Correct synchronization of the network is a prerequisite for maximum availability. It is therefore a good idea to monitor the wander characteristics during installation, routinely during operation and especially when changes in network topology are made, and not to wait until a problem occurs.

Where and how are measurements made?

PRC SSU
Schomandl FN-GPS/R (see back page) 10, 5, 2.048, 1.544 MHz alternatives

REF

SEC

SEC

Clock Data

SEC

Standard PDH/SDH optical/electrical data signals

A reference clock is always needed when making a measurement. This can be derived either from an external source or from the next higher-quality clock signal in the synchronization chain (absolute or relative measurement). In many SDH networks, the clock signal is transmitted between the network elements along with the STM-N information signal. Test equipment must therefore be able to recover the clock information from the signals at optical or electrical communications interfaces.

Measurements on network elements in the synchronization chain (relative or absolute measurement)

In the scenario opposite, a smaller network provider draws the clock signal from a larger provider via a data circuit (STM-1, PCM-30). The quality of the clock signal can be checked by absolute measurement against an external reference source.

STM-1, PCM 30

SDH network

SDH network

REF
Schomandl FN-GPS/R

Measuring the clock quality at network limits (absolute measurement)

PRC
2.048 Mbit/s

Switch A
2.048 Mbit/s

Switch B

TIE MTIE

2.048 Mbit/s ATM GSM

SDH
PDH

Measuring a signal transmitted over several networks (relative measurement)

In the following example, a relative measurement is more useful: Two switches (A and B) are synchronized to a PRC. A signal path (e.g. 2 Mbit/s) passes through various transport networks (SDH, PDH, etc.). Impairments caused by propagation time variations, mapping and pointer wander, and oscillator noise can lead to phase variations that are large enough to cause frame slips. A TIE or MTIE measurement can determine whether the phase deviations are within the recommended limit of 18 ms/day (ITU-T G.822 and G.823).

How long should measurements be made?

Wander is, by definition, a long-term phenomenon. To ensure meaningful results when measuring wander, a relatively long measurement time is required. It is not possible to specify generalized values, but measurement times of from several hours to one or more days are often used.

Measurement and qualification

The previous section discusses the how, where and why of synchronization measurements. Some examples now follow. The results were obtained using the ANT-20 Advanced Network Tester. This is not simply a test set, it is a measurement platform that can be used as a basis on which to build test solutions for coming to grips with synchronization problems. The basic version ANT-20 already provides a range of facilities, such as bit error rate tests and pointer analysis, that give initial information about faulty synchronization. The Jitter Option (generator / analyzer) is needed for wander measurement, analysis and qualification. This option also includes useful features such as tolerance to jitter tests and jitter transfer function measurements.

Wander analysis
The Wander Analysis Software (plug-in for the ANT-20 Jitter Option) allows wander measurements on all the affected network elements. TIE and MTIE (instantaneous values) are the quantities measured, with both these measurements being possible at all telecoms system interfaces: PDH, SDH, SONET, ATM, ISDN, GSM Electrical (balanced, unbalanced) and optical interfaces

The analog phase / time function is sampled at a rate of 80 Hz (sample interval 12.5 ms). The TIE is formed by averaging the value over intervals of 1 second. MTIE is determined as the difference between the maximum and minimum TIE values that occurred since the start of the measurement.

The illustration on the right shows a result display. The current TIE and MTIE values are shown numerically. The graph shows TIE versus time.

MTIE/TDEV analysis
The MTIE/TDEV Analysis Software significantly expands the analysis functions of the wander analysis software. MTIE/TDEV analysis based on ETSI Recommendation ETS 300 462 (corresponding to ITU-T G.810, G.813, G.811, G.812) can be performed using the recorded TIE samples. This LabWindows-based software tool can be run on the ANT-20 or on a separate, standard PC. The MTIE/TDEV analysis software calculates the MTIE, TDEV, ADEV and MADEV traces from the TIE samples collected using the ANT-20 wander analysis. The software can evaluate TIE samples from up to 100,000 s re-

MTIE and TDEV analysis. The MTIE and TDEV curves are shown graphically, along with the tolerance masks for an SSU. The observation interval comprises 1000 TIE samples.

cording time. The ETSI 300 462-3 tolerance masks can be used to qualify the synchronization elements PRC, SSU and SEC and for PDH signals. Rapid assessment is provided by a pass/fail software indicator. The built-in software simulator makes it easy to superimpose sine-wave, linear or square-wave signals or even white noise to give a virtual TIE curve. This can also be evaluated using the software and compared with the real measurement traces. Screen shots of the calculated traces can be made and linked in to standard Windows applications such as MS Word or MS Excel to produce a printed record.

Wander simulation and other possibilities

Using WG CATS (CVI Application Test Sequences), user-defined test sequences can be run automatically. The instructions, commands and control criteria are listed in so-called test cases which are processed by the software. The result of the test sequence is a pass/fail assessment. The measurement process is recorded as an ASCII file so that it can be reproduced accurately. A further useful feature is the ability to remote control other components using the software. GPIB (IEEE bus) or RS 232 (V.24) control is possible. One test sequence worthy of mention in this context is wander simulation. This allows controlled offsetting of a clock signal (step width 0.001 ppm, range 500 ppm). Any wander signal shape (e.g. saw-tooth) can be set in this way. The ANT-20 can also be operated remotely if equipped with the Remote Operation software package. This Windows software allows the ANT-20 to be controlled from any standard PC via Ethernet/LAN (TCP/IP) or via a modem link (V.24). The complete Windows user interface of the ANT-20 is reproduced on the remote PC. This means that wander measurements can be controlled and run from a remote location.

A typical application example

The clock source (SEC) of a network element equipped with OC-n or STM-N communications interfaces is synchronized externally via the 1.544 or 2.048 MHz input. The quality of the SEC is to be tested. Wander is measured relative to the reference frequency input. This ensures that only the wander components added by the network element are detected. In this example, access to the clock information is only possible via the OC-n/STM-N interface. The great advantage of the ANT-20 is that it can also measure the wander characteristics at this communications interface.

Built-in jitter and wander analyzer

Clock source (reference frequency) 1.544 or 2.048 MHz

SSU

SEC
Rx Wander measurement on optical or electrical OC-N/STM-N output via power splitter

SONET/SDHnetwork element OCN/STM-N

MTIE/TDEV analysis software

The result graph produced by the MTIE/TDEV analysis software (example on p. 8) shows the evaluation of both MTIE and TDEV together with the appropriate tolerance masks and a pass/fail assessment. This means that tolerance violations can be diagnosed at a glance.

Abbrevations ADEV ANT

Allan Deviation Advanced Network Tester

ATM

Asynchronous Transfer Mode

CATS GPS

CVI Application Test Sequences Global Positioning System Global System for Mobile Communication

GSM

ISDN LNC

Integrated Services Digital Network Local Node Clock

LORAN

Long Range Navigation

MADEV MTIE NE

Modified Allan Deviation Maximum Time Interval Error Network Element

OC

Optical Carrier

PDH PRC SDH SEC SSU

Plesiochronous Digital Hierarchy Primary Reference Clock Synchronous Digital Hierarchy SDH Equipment Clock Synchronization Supply Unit

STM

Synchronous Transport Module

TDEV TIE

Time Deviation Time Interval Error

TNC UTC

Transit Node Clock Universal Time Coordinate

Wandel & Goltermann Worldwide North America Canada Tel. 416 291 7121, Fax 416 291 2638 USA Tel. 919 941 5730, Fax 919 941 5751 Mexico Tel. 5 543 6644, Fax 5 543 8660 Latin America Argentina/Uruguay Tel. 1 784 4700, Fax 1 786 7917 Brazil/Chile/Peru/Bolivia/Paraguay Tel. 11 5505 3266, Fax 11 5505 1598 Colombia/Venezuela/Ecuador Tel. 1 256 4001, Fax 1 616 3267 Guatemala/Central America/Caribbean Islands Tel. 331-8065, Fax 331-8682 Europe Austria Tel. 22 52 85 521-0, Fax 22 52 80 727 Belgium/Luxemburg Tel. 2 725 18 19, Fax 2 725 41 42 France Tel. 01 30 81 50 50, Fax 01 30 55 87 75 Germany Tel. 7121 985610, Fax 7121 985612 Italy Tel. 2 895 12381, Fax 2 895 11780 Netherlands Tel. 40 267 97 00, Fax 40 267 97 11 Poland see Austria Scandinavia Tel. 08-449 48 00, Fax 08-449 48 39 Spain/Portugal Tel. 91-383-9801, Fax 91-383-2263 Switzerland Tel. 31 991 77 81, Fax 31 991 47 07 United Kingdom Tel. 1189-409200, Fax 1189-409210 East and South East Europe see Austria Africa North and West Africa see Germany East and South Africa see Switzerland Asia Middle East see Germany South East Asia see Australia Israel see Germany CIS Countries/Iran/Pakistan see Austria China Tel. 10-6856-1034, Fax 10-6856-1031 Hong Kong Tel. 2528-6283, Fax 2529-5593 India Tel. 11-6878537, Fax 11-6878538 Japan Tel. 45-473-9501, Fax 45-473-9812 Korea Tel. 2-563 2236, Fax 2-563 2239 Singapore Tel. 356-3246, Fax 356-3247 Australia/Pacific Islands/New Zealand Tel. 39-690 6700, Fax 39-690 6750 For all other countries (not listed) please contact: Wandel & Goltermann GmbH & Co. Electronic Measurement Technology Marketing International P.O. Box 12 62 D-72795 Eningen u.A. Tel. +49 (0) 7121-86 16 16 Fax +49 (0) 7121-86 13 33 e-mail: info@wago.de http://www.wg.com

For Your Synchronization Tests:

. . . and a matching option:

The complete solution based on the ANT-20 Advanced Network Tester platform
ANT-20 Mainframe
The portable nerve center for all basic tests: Bit error tests can make sure that the problems are not due to the transmission medium. Pointer analysis allows initial conclusions regarding possible synchronization problems to be drawn. * Tests SDH, SONET, PDH, ATM, TMN (all-in-one concept) * Windows user interface * Portable * Can be remote controlled * Easy to operate

FN-GPS/R Standard Frequency

GPS-synchronized standard frequency with rubidium backup oscillator. Provides the reference clock for wander analysis using the ANT-20. * Standard frequencies: 10 MHz, 2.048 MHz or others * Accuracy class 10-11 * Synchronizes to the GPS cesium reference oscillator * Can be used anywhere in the world * Precision rubidium backup oscillator * Displays geographical position, date and time * Display of control data and alarms * Remote monitoring via modem link A Schomandl company product, marketed by Wandel & Goltermann & Co. for wander applications.

+ Jitter Option
This is required for recording the TIE samples needed for wander analysis and MTIE/TDEV analysis. * Jitter generator * Jitter analyzer * Jitter tolerance tests * Jitter transfer function measurement

+ Wander Analysis Software

Provides precision display of TIE values with the corresponding MTIE values show numerically. * Collection and graphical display of TIE values * Calculation of instantaneous MTIE value

+ MTIE/TDEV Analysis Software

Characterizes the synchronization quality of a network to ETSI 300 462-3 from the MTIE, TDEV, ADEV and MADEV traces. * Calculation over several observation intervals (1 to 100,000 s) * Tolerance masks for PRC, SSU, SEC and PDH signals * Simulator Runs on ANT-20 or on any standard PC

+ WG CATS Software (CVI Application Test Sequences)

Automation, e.g. for monitoring network elements. WG CATS also allows wander simulation. * Automatic running of test sequences * Remote-control of other components

+ Remote Operation Software

Provides remote control of the ANT-20 via Ethernet/LAN (TCP/IP) or dialup modem link (V.24) from any standard PC. * Wander measurements from a remote PC
Subject to change without notice Order no. E5.97/WG1/197/5 Printed in the Federal Republic of Germany