WO1993015563A1

WO1993015563A1 - Test signal transmission between stations for group delay measurement

Info

Publication number: WO1993015563A1
Application number: PCT/GB1993/000178
Authority: WO
Inventors: John David Brownlie
Original assignee: British Telecommunications Public Limited Company
Priority date: 1992-01-28
Filing date: 1993-01-28
Publication date: 1993-08-05
Also published as: GB9201737D0

Abstract

Modems 1, 2 undertake a start-up sequence in which modem 1 transmits a first multi-frequency test signal; modem 2 measures the SNR of the first test signal; modem 2 transmits a time reference to modem 1; modem 1 responds by switching from the first test signal to a second such signal having a different power from the first; modem 2 measures the SNR of the second test signal and also recognises the timing of the power transition to determine the duration of the delay between the time reference and the transition.

Description

TEST SIGNAL TRANSMISSION BETWEEN STATIONS FOR GROUP DEL-AY MEASUREMENT

The present invention is concerned with modems, that is to say, modulation/demodulation apparatus by means of 5 which digital signals to be transmitted control one or more of the properties (phase, frequency, amplitude) of a carrier, and digital signals to be received are recovered from an incoming modulated carrier. The precise form of modulation to be used is not however critical to the present invention,

10 which is more concerned with so-called "start-up" procedures. A transmission link comprising two simple modems connected by a transmission path requires little preparation before data can be sent; however where-as is common - adaptive echo cancellers are used to improve the quality of transmission it

15 is usual for each modem in turn to transmit a period of unmodulated carrier so that the other modem' s echo canceller can adapt to the characteristics of the path. Moreover, to meet the increasing desire for higher bit-rates over links of given bandwidth and signal-to-noise ratio, more sophisticated

20 start-up procedures may include provision for measuring other characteristics of the path to enable decisions to be made as to (for example) the bit-rate that may be used, or the optimum transmission powers to be employed (for example as described in our UK patent application no. 8628656). Also

25 our UK patent application no 9116177.8 discusses the generation of test or "probing" multi-frequency signals having spectral lines of uniform amplitude which may be transmitted by one modem so that the other can evaluate, from the received signal, the bandwidth characteristics of the t 30 path and reach a decision on a suitable symbol rate. - - Figure 1 shows identical stations 1, 2, connected by a telephone line or other transmission path 3. Before discussion of the construction of the stations in detail, a start-up procedure to be undertaken by each station at the

35 commencement of a ϋeriod of transmission will first be described. In practice the start up procedure will include provision for echo canceller training;., but this and •other conventional parts of the procedure will not be described. The object of the sequence now to be described is (a) to determine the signal-to-noise ratio of the transmission path at different frequencies.

(b) to determine the signal-to-noise ratio of the transmission path at different transmission powers - thereby determining the relative amounts of noise which are, respectively, dependent on amplitude (eg quantisation noise particularly in nonlinear coding processes such as μ-law coding which may be included in the link) or amplitude independent, or arising as inter odulation distortion.

(c) to determine the round-trip delay - i.e. the sum of the transmission times from station 1 to station 2 and back again.

Optionally, (d) the variation of group delay across the signal bandwidth of interest may also be measured.

The precise use to which this information once obtained, may be used is not an essential part of the present invention, however it is noted that on paths - such as those containing satellite links - with long delays the echo duration may be such as to render the use of a single echo canceller uneconomic and the known expedient of employing a first echo canceller to deal with short-term echo and a second canceller (receiving the locally transmitted signal via a delay line) to deal with long-term echo may be employed, in which case the round-trip delay measurement (c) is needed in order that the delay line may be appropriately adjusted. The other measurements may be used so that the modems (usually in concert) may determine the symbol rate and/or power and/or portion of the available bandwidth to be used (truncation may be desirable if for example the signal- to-noise ratio or group delay characteristics are particularly unsatisfactory close to the band limits).

A timing chart is shown in figure 2. During the sequence depicted, station 2 transmits a multi-frequency probing signal PI which is analysed at station 1 to provide measurements (a). It also repeats this signal at reduced powers (P2, P3) to permit measurements (b) by station 1. The delay measurements necessitate transmission by one station to be acknowledged by the other and station 1 therefore also transmits a signal Q which may be a single frequency tone, or possibly another ultitone signal. The signal Q should be such as to minimise crosstalk with the probing signal. Supposing that the latter contains components at odd multiples of 50Hz then the frequency of the former would ideally be an even multiple (or even multiples) of 50Hz. A single tone of 600Hz is assumed for the purposes of description.

In more detail, at time tl, station 1 begins transmitting unmodulated carrier at 600Hz and station 2 begins transmitting probing signal PI at maximum power. Station 1 receive the probing signal PI and performs signal- to-noise ratio measurements on the various frequency components. When the measurements are complete, station 1 then signals to station 2 at time t2 by a suitable modulation of its 600Hz carrier (eg. by a simple phase reversal indicated as " ø" in figure 2). This transition is received by station 2 (after a delay Δl equal to the transmission delay from station 1 to station 2) at time t3. Station 2 responds (following a fixed delay Δ2) by transmitting the probing signal at reduced power (e. g -1.5dB).

For the purposes of round trip delay and group delay measurement, station 1 must recognise this event; in principle it could do this on the basis of the power change, but, in order that the timing of this event be accurately observed at station 1 it is preferred that it be accompanied by a phase reversal. Thus at time t4 stations 2' s transmission of probing signals PI is replaced by transmission of probing signals P2 which, relative to PI, are phase reversed and 1.5dB down.

After a delay Δ3 (the transmission time from station 2 to station 3) the change is received at station 1 - at times t5. Since the group delay varies with frequency, the value of Δ3, and the times t5, are a function of frequency. The round-trip delay is Δl+Δ3=t5-t2-Δ. Thus, at this point, station 1 is in possession of this information. If the value of t5 at frequency f is t5(f) then the round trip delay at 600Hz is approximately [t5(650)+t5(550) ]-t2-Δ2 and the differential delay for the station 1 - station 2 path at frequency f is t5(f)- [t5(650)+t5(550) ]. Assuming that the differential delay is the same for both directions then the round trip delay at frequency f is 2t5(f) -k [t5(650)+t5(550) ]-t2-Δ2.

After a delay Δ4 to allow any transients from the phase reversal and power reduction to clear, station 1 can then analyse the received signal P2 in the same manner as for PI.

If it is required that station 2 also make a round- trip determination then this can be achieved by providing that station 1, at time t6 (a fixed delay 5 following time t4) execute a further phase reversal of its carrier, which is then recorded at station 2, Δ6 later, at time t7. The round- trip delay is Δ3+Δ5=t7-t4-Δ4.

If desired, a further probing signal P3 at a further reduced power (e.g. -3dB relative to PI) may then be transmitted by station 2 (at time t8) and analysed by station 1 as for PI and P2. Signal P3 may be phase-reversed if this is found desirable to ensure that station 1 does not commence measurement until it is known that signals P3 have commenced.

The sequence described above may be followed by a further sequence in which the roles of the two stations are reversed.

Figure 3 is a block diagram of station 1 (station 2 being identical). Data received at an input 100 feeds a modulator 101 of conventional construction, its modulated output passing via an electronic changeover switch 102, controllable attenuator phase reversal unit 103 and hybrid circuit 104 to the line 3. The changeover switch is switchable to receive signals from the signals from the modulator 101, a 600Hz tone generator 105 or a probing signal generator 106 (eg. constructed as described in our above- noted patent application) . If desired, to provide a probing signal having spectral lines located in regions of particular interest - for example only at the centre of the ^'frequency band and at the band edges - the signals pi etc may be formed as the sum of several signals generated as described in the earlier patent application. In this case, the signal Q would preferably lie in one of the larger intervals of this spectrum.

Received signals from the line 2 pass via the hybrid 104 to a receiver 107 (which may contain equalisation and/or carrier recovery circuits). Its output passes via short-term and long-term echo cancellers 108, 109 (fed with transmitted signal directly and via a delay line 110 respectively) and thence to a demodulator 111 and a data output 112. The received signals from the receiver 107 are also fed to a phase detector 113 (for detecting phase reversals of 600Hz carrier) and a signal-to-noise ratio measuring unit 114 (for measuring the SNR of received probing signals).

A control unit 115 is provided to control the sequence of operations in accordance with figure 2 (specifically the choice of transmission (line 120)), and to receive the outputs of the phase detector 113 and SNR measurement 114. It has control outputs 122, 123 to set the delay value of the line 110, and transmitted power and phase and outputs 124, 125 to control the data rates of the modulator and demodulator. Further connections 126, 127, permit the control unit to communicate signalling data; for example station 2 could sent to station 1 at time t9 a signal SR indicating the symbol rate, carrier frequency and amplitude to be used for subsequent high speed transmission.

Claims

1. A data transmission system comprising first and second stations, the first station including means to transmit a first test signal followed by a second test signal having a different transmitted power from the first, the test signals each comprising a plurality of discrete frequency components; and the second station including means to determine the ^* relative values of signal and noise of the received test signals; wherein the second station further includes means to transit to the first station a signal indicating a time reference and the first station is operable in response to receipt of such signal to control the timing of the transition from the first test signal to the second, the second station having means to recognise the timing of the said transition to determine the duration of the delay between the time reference signal and the transition.

2. A data transmitting and receiving apparatus comprising: (a) means to transmit a first test signal followed by a second test signal having a different transmitted power from the first, the test signals each comprising a plurality of discrete frequency components;

(b) means to determine the relative values of signal and noise of such test signals received from another such apparatus;

(c) means to transmit a signal indicating a time reference;

(d) means operable in response to receipt of such a time reference signal from another such apparatus to control the timing of the transition from the first test signal to the seqond; and (e) means to recognise the timing of such a transition in test signals received from the other such apparatus to determine the duration of the delay between the time reference signal transmitted by the apparatus and the receipt of a transition in the test signals received from the other apparatus.

3. A system according to Claim 1 in which the recognition means of the second station is operable to recognising the transition in each of the discrete components of the test signal so as to determine the group delay of the channel as a function of frequency.

4. An apparatus according to Claim 2 in which the recognition means (e) is operable to recognising the transition in each of the discrete components of the test signal so as to determine the group delay of the channel as a function of freσuencv.