GB2332593A - A method of testing a telecommunications link - Google Patents

Abstract

A method of testing a telecommunications link between a subscriber unit (20) and an exchange (30) is disclosed. The method includes the steps of causing a subscriber unit to emit a predetermined signal, which is then transmitted along the link to the exchange. The received signal is analysed and a determination is made regarding the quality of the telecommunications link based on the received signal.

Description

1 A Method of and Apparatus for Testing a Telecommunications Link 2332593

This invention relates to a method of and apparatus for testing a telecommunications link and in particular to a method of and apparatus for testing a copper pair connecting a telephone exchange to a subscriber unit.

The connection between individual telephone subscribers, be they domestic or business subscribers, and a local telephone exchange has traditionally been provided using copper cables consisting of a number of twisted-pair wires known as copper pairs. When these copper pairs were first deployed in local areas it was assumed that they would be used for transmission of voice signals only. These connections were expected to operate in a frequency range less than 4 kHz. Therefore, the planning rules which were adopted were based on easily controlled and measured parameters such as loop resistance and low frequency attenuation. In the UK the normal planning limits are 1000 Ohms loop resistance and 10 dB attenuation at 1 kHz.

These limits are achieved by a suitable choice of conductor gauge or diameter depending on the route distance between the exchange and the customer. Longer routes clearly require larger conductors in order to meet the resistance and attenuation limits. Conductors tend to be between 0.3 mm and 0.9 mm in diameter, with increasingly larger diameter conductors being used the further the cable extends from the exchange. This allows for bundles of narrow gauge pairs to be grouped together at an exchange thus minimising cable handling problems.

As the number of new subscribers obtaining telephone services from operators utilising optical feeders increases, telephony providers, 2 whose systems are largely constructed of copper pairs, are increasingly looking to the provision of similar wideband and broadband services to their customers over their copper pair links. With the advent of wideband and broadband services such as internet access, video on demand and digital data transmissions as well as increases in the volume of telephony services telephony providers require to test their individual links between exchanges and subscribers in order to ascertain whether or not each link will support the provision of such services. In particular, lines must be tested to see if they will support present ISDN services and as time passes will more frequently need to be tested to ascertain whether or not they will support services requiring technologies such as HDSL, ADSL and VDSL to be used over the copper pairs.

One of the key basic parameters for establishing the suitability of a particular copper pair for carrying such services is its transmission length, as signal attenuation increases with transmission length.

Unfortunately, this is not readily deducible from the records of a particular operator, even if they are accurate. This is because, although the records show duct routes and section lengths, they do not necessarily indicate how a cable is routed through the duct. For example, it is often found that a copper pair in a specific cable will transverse the full length of the duct to a splice point and then return along the same duct as a pair in another, possibly smaller cable.

It is equally misleading to use measurements based on grid references in order to predict lengths, because the necessary scaling factor of actual cable length to direct distance is unknown in any specific instance. For example, in the UK, the average scaling factor is probably somewhere between 1.4 and 1.8. The resulting distances are often enough to render a link unsuitable for the provision of wideband or 3 broadband services and the uncertainty in ascertaining actual cable route lengths makes this method highly inaccurate. In addition, the connectors may include sections of aluminium, which will have different transmission characteristics to the copper sections. Aluminium was used in this way when copper prices made copper less economic than aluminium.

One alternative approach is for operators to dispatch staff to a customer's premises, to undertake one or a series of measurements of the copper cable and its performance. This is a time consuming and consequently expensive solution, especially if the customer decides not to take the service, or takes it only for a short period. A variation on this method is for the operator to take sample measurements of cable lengths and performances from the exchange to the local telephone cabinet, from where individual copper pairs are directed to individual subscribers. As the cables from the exchange to the cabinet are shared this would reduce the cost per customer line, but would only give an indication of a few of the copper pairs from the exchange. Equally this gives no indication of the length and performance of the copper drops from the cabinet into customer's premises. Therefore this method is again rather inaccurate.

Time domain reflectometry (TDR) is a technique primarily used for determining a discontinuity or breakage in a cable. However, this may be usable to test a link from an exchange to a subscriber unit.

Unfortunately, this method is not conducive to copper pairs, as there can be many reflections from imperfections, spliced joints, etc. which may mask the ultimate reflection, if any, from the end of the cable.

Also, although most telephones are quite well matched to the line, they may not produce a reflection at the point where it may be most useful.

Equally, the number of telephones connected at a customer's premises, 4 and hence the differing impedances produced, could cause spurious results. The main difficulty is that TDR measurements require a fast pulse to operate accurately. This is not possible with copper cables beyond a few hundred metres in length. Therefore, again this method is not appropriate.

An object of the present invention is to provide an inexpensive and accurate method of testing a telecommunications link between an exchange and a subscriber.

A further object of the present invention is to provide a method of testing a telecommunications link which does not require an engineer to visit a subscriber.

A further object of the present invention is to provide a method of testing a link in response to a subscriber's telephone enquiry which can be carried out as part of a single telephone call.

According to a first aspect of the present invention there is provided a method of testing a telecommunications link between an exchange and a subscriber unit, comprising the following steps:

a) causing the subscriber unit to emit a predetermined test signal, which is transmitted along the telecommunications link to the exchange; b) analysing the test signal received at the exchange; and c) making a determination regarding the quality of the telecommunications link based on the signal received at the exchange.

Preferably, the predetermined test signal is within the audio frequency range.

More preferably, the subscriber unit is a Dual Tone MultiFrequency (DTMF) telephone having a key pad including a plurality of keys.

Most preferably, the predetermined test signal is a two-tone signal produced by depressing a key on the key pad of said (DTMF) telephone.

The predetermined test signal may be produced by depressing two or more keys.

Preferably, the degree of attenuation and/or group delay of the predetermined test signal due to the link is determined.

The determination may relate to the suitability of the link for propagating audio signals.

Alternatively, a projection may be made from the received audio signal, regarding the suitability of the link to carry broadband telecommunication signals.

Preferably, signature analysis is utilised to project the broadband characteristics of the link.

More preferably, a look up table is formed from empirical data in which known links are tested at audio and broadband frequencies such that when a particular audio frequency signal is received, when a specific link is being tested, a corresponding broadband result can be 6 obtained and a determination made regarding the suitability of the specific link being tested to support broadband services.

Alternatively, an algorithm may be provided to project the expected behaviour of the link with broadband signals from the received audio frequency signal.

Typically, the audio signals have a frequency of less than 4 kHz.

The broadband signals may be in the range 10 kHz to 10 MHz.

Preferably, the broadband signals include ISDN, HDSL, ADSL and VDSL signals.

The telecommunications link may largely comprise a twisted copper pair. Although, the link may include one or more lengths of twisted aluminium pairs.

Different portions of the link may be comprised of twisted copper or aluminium pairs of different diameters in the range 0.3 to 0.9 mm.

Embodiments of the present invention will now be described, by way of example, with reference to the accompanying drawings in which:

FIGURE 1 is a flow chart indicating a method of testing in accordance with the present invention which is instigated due to a subscriber enquiry; FIGURE 2 is a schematic diagram of a telephone system capable of employing the method illustrated in Figure 1; FIGURE 3 illustrates a dual-tone signal generated by a DTIVIF telephone; 7 FIGURE 4a illustrates the attenuation of a typical customer line with distance from an exchange; and FIGURE 4b illustrates the phase shift of a typical customer line with distance from an exchange.

Figure 1 illustrates a method of testing a telecommunications link between a subscriber unit (20) and an exchange (30) (Figure 2). In Figure 1 the method is actuated due to a call from a subscriber as detailed in box 2 of the flow chart. The operator then connects the call to an automatic test system as indicated in box 4 of the flow chart.

The method comprises the step of causing the subscriber unit (20) (Figure 2) to emit a predetermined signal. In this embodiment the signal is produced by the subscriber depressing a key (22) on a keypad (24) on the subscriber unit (20). This step is indicated at box 6 on the flow chart.

The signal thus produced is transmitted along the link (26), in the form of a copper pair, between the subscriber unit and the exchange (30). The predetermined test signal emitted by the subscriber unit (20) is thus connected to a receiver (32) within the test apparatus (18) from where it is transmitted to an analyser (34) for analysis such that a determination regarding the quality of the telecommunications link can be made based on the received signal. The entire process is controlled by a control means (36).

The signal is produced by a dual tone multi-frequency (DTIVIF) telephone in which each of the keys produces a signal which is well specified for example in the UK in terms of frequency and relative amplitude. The frequency of each tone within the dual tone produced by each key is controlled to an absolute accuracy of +/- 1.5%, whilst 8 pre-emphasis of the high tone with respect to the low tone is controlled to be 2.5 dB +/1.5 dB. Each IDTIVIF subscriber unit (20) provides sixteen tones per combination, defined using numbers 0 to 9,, # and letters A to D. These range from the lowest at 697 Hz to the highest at 1, 633 Hz. In practice, most telephones do not include the letters A to D which use the 1,633 Hz high tone, so that the highest available tone generally is 1,477 Hz. It is a primary advantage of the method and apparatus in accordance with the invention that use can be made of these varying tones to determine the performance of the customer's telephone link. An example of a dual-tone signal generated by a IDTIVIF telephone is shown in Figure 3, in the time domain.

The determination made regarding the quality of the link can either be a determination of the link quality in the audio band or it can be a projection of the estimated link quality when supporting broadband or wideband telecommunications signals. As illustrated in box 10 of the flow diagram the method may include signature analysis in either the frequency domain or time domain (or both) which may be utilised to project a broadband characteristics of the link. In such an analysis a look up table is formed from empirical data which is obtained using known links for which both audio and wideband or broadband frequencies are monitored such that when a particular audio frequency signal is received during use a corresponding broadband or wideband result can be obtained from the look up table and a determination made regarding the quality of the specific link being tested, for the provision of broadband or wideband services.

Alternatively as illustrated in box 12 the method may include the running of an algorithm which is provided to project the expected behaviour of the link with broadband or wideband signals from the received audio frequency signal. Wideband and broadband signals 9 include ISDN, HDSL, ADSIL and VDSIL signals. The user is then informed of the result of the test (Box 14) and returned to the operator (Box 16) for further information, or advice on services.

The proposed testing equipment uses digital signal processing (DSP) to analyse the waveform received at the exchange after the connected customer has pressed a key, as instructed by the test apparatus. The composite waveform arriving at the exchange will have been modified by the copper pair between the customer and the exchange in a manner that can be correlated with the propagation characteristics of the line. In the frequency domain there will be a reduction in the relative and absolute amplitudes of the two fundamental tones produced by the subscriber unit, whilst in the time domain there will be a finite group delay between the two tone bursts.

One embodiment of the invention uses algorithms predicting the effect of the characteristics on the relative amplitude and group delay, which can be coded into the DSP unit so that the results of a given test can be translated back into line characteristics.

Alternatively, as discussed above, it would be possible to store representative waveforms from a number of known test lines and use these as a database for comparison with the waveforms received from lines under test. This technique is otherwise known as signature analysis and may be more reliable and less costly to implement than the first mentioned algorithmic approach.

Furthermore, if results of testing with a single key press are inconclusive it should be possible to request the customer to either repeat the first digit or even to press a different digit. The latter will invoke an alternative tone pair which may be analysed so as to bring the test to a satisfactory conclusion.

If results using a single digit are still inconclusive it should be possible to ask the customer to press a predetermined sequence of digits in order to bring the test to a satisfactory conclusion.

Clearly the test apparatus will need to incorporate a standard DTIVIF receiver function in order to confirm that the customer has pressed the correct key(s), before starting analysis of the waveform. It is noted that IDTIVIF transmitter output is permitted to vary with line currents above 40 mA and is unspecified below 25 mA and it is therefore important that the tests are performed with line currents in this range.

At the exchange, the operator who is running the tests with the customer may be sited at a normal operator console or at a test-desk.

In the latter case, it is possible that adjustments may be made to the customer's telephone line current which can then be specially set to a particular current in order to obtain the best performance from the customer's DTIVIF telephone.

An implementation of the Analyser box 34 is to take the outputs from the receiver 32 and process them into a form whereby meaningful data concerning the characteristics of the subscriber's line may be made. This processing can take the received IDTIVIF tones and their relative amplitudes and delays.

This processing may produce either frequency or time domain data series, and can be either linear or non-linear.

11 An example of the linear case is the use of a fast fourier transform (FFT), using DSP, to convert signals received in the time domain into the frequency domain, for assessment of the absolute frequency amplitudes and their relative amplitudes.

A particular example of the non-linear case is to half-wave rectify the received time domain signals which has a special ability to extract lower frequencies from the received signals, as can be seen from the dual tape signal illustrated in Figure 3. Because the frequency is lower than the two DTIVIF tones, its amplitude is less affected by the customer's copper pair and can therefore be used as a reference level for the analysis of the DTIVIF tones. This non-iinear processing also produces higher frequency components which will be quite sensitive to the customer's cable characteristics. This may therefore be used for generating the required signature of the line.

Modifications may be incorporated without departing from the scope of the present invention as claimed.

12

Claims (18)

Claims:

1. A method of testing a telecommunications link between an exchange and a subscriber unit, comprising the following steps:

a) causing the subscriber unit to emit a predetermined test signal, which is transmitted along the telecommunications link to the exchange; b) analysing the test signal received at the exchange; and making a determination regarding the quality of the telecommunications link based on the signal received at the exchange.

2. A method as claimed in Claim 1, wherein the predetermined test signal is within the audio frequency range.

3. A method as claimed in Claim 2, wherein the subscriber unit is a Dual Tone Multi-Frequency (DTIVIF) telephone having a key pad including a plurality of keys.

4. A method as claimed in Claim 3, wherein the predetermined test signal is a two-tone signal produced by depressing a key on the key pad of said (DTMF) telephone.

5. A method as claimed in Claim 4, wherein the predetermined test signal is produced by depressing two or more keys.

13

6. A method as claimed in Claim 1, wherein the degree of attenuation and/or group delay of the predetermined test signal due to the link is determined.

7. A method as claimed in Claim 1, wherein the determination relates to the suitability of the link for propagating audio signals.

8. A method as claimed in Claim 1, wherein, a projection is made from the received audio signal, regarding the suitability of the link to carry broadband telecommunication signals.

9. A method as claimed in Claim 8, wherein signature analysis is utilised to project the broadband characteristics of the link.

is

10. A method as claimed in Claim 9, wherein a look up table is formed from empirical data in which known links are tested at audio and broadband frequencies such that when a particular audio frequency signal is received, when a specific link is being tested, a corresponding broadband result can be obtained and a determination made regarding the suitability of the specific link being tested to support broadband services.

11. A method as claimed in Claim 8, wherein a algorithm is provided to project the expected behaviour of the link with broadband signals from the received audio frequency signal.

12. A method as claimed in Claim 7, wherein the audio signals have a frequency of less than 4 kHz.

13. A method as claimed in Claim 8, wherein the bandwidth of the broadband signals is in the range 10 kHz to 10 MHz.

14 14. A method as claimed in Claim 8, wherein the broadband signals include ISDN, HDSL, ADSL and VDSL signals.

15. A method as claimed in Claim 1, wherein the telecommunications link largely comprises a twisted copper pair.

16. A method as claimed in Claim 15, wherein the link may include one or more lengths of twisted aluminium pairs.

17. A method as claimed in Claim 16, wherein different portions of the link may be comprised of twisted copper or aluminium pairs of different diameters.

18. A method as claimed in Claim 15, wherein the diameter of the twisted copper pair lies in the range 0.3 to 0.9 mm.