Method, apparatus, and system for time synchronization of xdsl

p00001Method, equipment and system for time synchronization in xDSL

p00002Field of the Invention

p00003The present invention relates to the field of communications and, more particularly, to a method, a device and a system for synchronizing the time of a Digital Subscriber Line (DSL).

p00004Background of the invention

p00005The number of Femtocells is increasing to meet the needs resulting from the emergence of 3rd generation mobile communication (3G) and other advanced digital mobile communications technologies. In the Femtocell a time synchronization with great precision is necessary. In general, a clock recovery module is included in a network terminal. In this way, clock synchronization can be easily provided for the Femtocell (this is frequency synchronization). However, it is very difficult to provide time synchronization. It is necessary to solve some technical questions. FIG. 1 is a schematic diagram showing a scheme to achieve time synchronization with precision proposed in the art. Suppose that Offset is a lag between a slave clock and a master clock, Delay1 is a propagation delay from the master clock to the slave clock and Delay2 is a propagation delay from the slave clock to the master clock. In this case, from FIG. 1 it can be deduced that:

p00006Ts0 = Tm1 + Offset +

p00007Ts1 - Ts0 = Delay1

p00008therefore, Offset = Ts1 - Tm1 - Delay1

p00009equivalently, Tm2 = Ts2 - Offset + Delay2

p00010so that, Offset = Ts2 - Tm2 + Delay2

p00011If the delay from the master clock to the slave clock is equal to the delay from the slave clock to the master clock, that is, Delay1 = Delay2, then

p00012Offset = (Ts1 + Ts2 - Tm1 - Tm2) / 2 (1)

p00013In this way, the offset between the slave clock and the master clock is obtained so that the slave clock can be precisely synchronized with respect to the master clock.

p00014However, in the event that a device of the Subscriber's Line xDigital (xDSL) functions as a mobile hub, the Master corresponds to a switchboard (CO) device, and the Slave corresponds to a Device in the Customer Facilities (CPE). The channel between the CO equipment and the CPE is complicated, and passes through an analog circuit of the CO equipment, a cable, an analog circuit of the CPE and also digital signal processing circuits in the CO equipment and in the CPE. As a result, the downlink delay from the CO equipment to the CPE may not necessarily be the same as an uplink delay from the CPE to the CO equipment; This is, in general, Delay1 ≠ Delay2. According to some empirical results, the difference between Delay1 and Delay2 is greater than 1μs. Therefore, the offset between the CO clock and the CPE clock cannot be obtained directly from the formula (1).

p00015As shown in FIG. 2, a downlink delay includes a delay! T1 of a digital transmission circuit of the CO, a delay! T2 of an analog transmission circuit 203 of the CO, a delay! T3 of the downlink of a twisted pair 90 , a delay! t2 'of an analog reception circuit 205 of the CPE, and a delay! t1' of a digital reception circuit 80 of the CPE; and an uplink delay includes a delay! t4 of a digital reception circuit of the CO, a delay! t5 of a circuit 2005 of analog reception of the CO, a delay! t6 of the uplink of a twisted pair 90, a delay! t5 'of an analog transmission circuit 2003 of the CPE, and a delay! t4' of a digital transmission circuit 85 of the CPE. In general, Delay1 =! T1 +! T2

p00016+! t3 +! t2 '+! t1' ≠ Delay2 =! t4 +! t5 +! t6 +! t5 '+! t4', and the difference between the two delays is, in general, greater than 1μs.

p00017An xDSL receiver detects a frame limit and implements frame synchronization during initialization. In real cases there may be a small error with the synchronization algorithm, and the precision of the synchronization is limited by the sampling rate and the error of the frame synchronization can affect the precision of the time synchronization. If a transmitter records the beginning of a specified frame as a time stamp Tm1 (on the side of the CO) or a time stamp Ts2 (on the side of the CPE), an error is entered when a receiver records the time stamp Ts1 (on the CPE side) or a Tm2 timestamp using an algorithm

p00019for frame synchronization. Due to the frame synchronization error, the error introduced by the registration of Ts1 on the CPE side or the Tm2 on the CO side will be very large. In particular, the error will be even greater when the CO records the Tm2 in the upstream direction with a low sampling rate.

p00020Delay1 can also be obtained directly by measuring a delay of the downlink channel. In this way, a gap between the CO and the CPE can be obtained directly, that is, Offset = Ts1 - Tm1 - Delay1. However, at present, the measurement of the delay of the xDSL channel (especially the twisted pair) is not accurate enough, particularly when the length of the loop is too large, there is a lot of noise in the loop or there are splices in the loop .

p00021The publication of Sungwon Lee "An Enhanced IEEE 1588 Time Synchronization Algorithm for Asymmetric Communication" in IEEE COMMUNICATIONS LETTERS, VOL. 12, NO. from September 9, 2008, pages 687-699, discloses an improved synchronization algorithm to calculate the asymmetric proportion of an xDSL communication link, and the algorithm improves the timing accuracy.

p00022Summary of the Invention

p00023An object of the present invention is to accurately obtain a delay of one channel, to ensure that the CO equipment and the CPE correctly read the clock time, to achieve time synchronization between the CPE and the CO equipment by the calculation of a gap between the CPE clock and the CO equipment clock.

p00024The invention is defined in the claims.

p00025In accordance with the present invention, the problem of a poorly defined frame limit can be solved when the frame limit is retrieved using an algorithm of the receiving terminal; a synchronization error can be calculated between a reception terminal and a transmission terminal according to a specific symbol transmitted by the transmission terminal, and then, according to the synchronization error, an error of the marking of the time caused by the badly defined frame limit. At the same time, a gap between a CPE clock and a CO equipment clock can be obtained by calculating a delay of a channel so that synchronization between the CPE clock can be precisely achieved based on the offset and the clock of the team of the CO.

p00026Brief description of the drawings

p00027FIG. 1 is a schematic diagram illustrating the principle of time synchronization defined in IEEE 1588v2;

p00028FIG. 2 is a schematic diagram of a propagation delay of the downstream and a propagation delay of the upstream;

p00029FIG. 3 is a flow chart of a synchronization method according to a first embodiment of the present invention;

p00030FIG. 4 is a schematic diagram that identifies the elements that constitute a downward propagation delay;

p00031FIG. 5 is a schematic diagram that identifies the elements that constitute an upward propagation delay;

p00032FIG. 6 is a flow chart of a synchronization method according to a second embodiment of the present invention;

p00033FIG. 7 is a schematic diagram of a system according to the present invention; and

p00034FIG. 8 is a schematic diagram of an equipment according to the present invention.

p00035Detailed description of the embodiments

p00036Hereinafter the present invention is clearly described with reference to the accompanying drawings.

p00037A first embodiment of the present invention provides a method for time synchronization of xDSL. The method includes the following steps:

p00038transmit, by a first device, a first symbol to a second device, and obtain the time Ts2 indicating the moment at which the first symbol is transmitted;

p00039receive, by the first team, a second symbol transmitted by the second team, and obtain the time Ts1 3

p00041which indicates the moment at which the second symbol is received;

p00042obtain, by the first team, the time Tm2 that indicates the moment at which the first symbol is received by the second team and the time Tm1 that indicates the moment at which the second symbol is transmitted by the second team ;

p00043calculate, by the first team, a gap between a clock of the first team and a clock of the second team based on Ts1, Ts2, Tm1, Tm2 and a delay of the first team; and

p00044adjust, by the first team, the clock of the first team based on the offset to achieve synchronization. In the following embodiments, the first equipment is considered as a CPE and the second equipment is considered as a CO; however, those skilled in the art may understand that the first team can also be a CO and the second team can also be a CPE.

p00045When an uplink delay is not equal to a downlink delay, the offset between the CPE clock and the CO clock is obtained using a certain mathematical relationship between the Delay1 delay of propagation of the downstream and the Delay2 delay of upstream propagation so that the CPE (or CO) can adjust the local clock based on this offset.

p00046The method for time synchronization according to the first embodiment works so that the CPE first transmits a synchronization symbol and then the CO transmits a synchronization symbol, the specific process of which is shown in FIG. 3.

p00047In step 10, the CPE transmits the first symbol, and obtains the time Ts2 indicating the moment at which the first symbol is transmitted.

p00048In xDSL a discrete multi-carrier modulation (DMT) scheme is used so that in a DMT frame a signal is transmitted. In this case, time synchronization in xDSL is also achieved in DMT frames. Therefore, the first symbol transmitted by the CPE can be a DMT frame, and the specific frame to be chosen can be determined by negotiation between the CPE and the CO.

p00049During initialization, the CPE transmits the first symbol. When a particular position of the first symbol or a D / A module is written to an intermediate memory from the buffer, the CPE records the corresponding Ts2 time of its local clock.

p00050Through negotiation between the CO and the CPE a specific point is also determined at which the time stamp registration is activated. Any position of the first symbol can be used. In the following embodiments, an initial position of the first symbol is taken as an example.

p00051In step 20, the CO receives the first symbol transmitted by the CPE, and obtains the time Tm2 indicating the moment at which the first symbol was received.

p00052The CO receives the first symbol transmitted by the CPE. When the CO records in the buffer a sample at the start position of the first symbol or an A / D module reads from the buffer the sample at the start position of the first symbol, the CO records the corresponding time Tm2 'of its local clock (that is, an action is initiated to obtain a timestamp). Because the CO obtains a frame limit by calculating with a certain algorithm, when the starting position is calculated with the algorithm an error can be introduced. In this case, the CO has to correct the time Tm2 '.

p00053According to a phase difference between a phase of the reception point and a phase of the check point of a sinusoidal signal (or a cosinusoidal signal) of the first symbol, the CO corrects the time Tm2 'to obtain a time Tm2 where the Tm2 time is the instant of time that indicates the moment at which the CO should receive the checkpoint. The reception point is a point of the signal at which the CO receives the first symbol initially, and the checkpoint is a point of the signal at which the CPE transmits the first symbol initially.

p00054When the CO corrects the time Tm2 'according to a sinusoidal signal in the first symbol:

p00055a phase of a corresponding point of the sinusoidal signal is set (for example, 0º, 45º, 90º or any other angle) when the CPE initiates an action to obtain the time frame so that this point can be taken as a point test and the checkpoint phase is obtained when the CO corrects the time Tm2 '. In the following embodiments, 0 ° is taken as an example.

p00056The CO obtains a position of the sinusoidal signal in which the CO starts obtaining a time stamp, where the position is a receiving point where the CO receives the first symbol, and calculates the time that must elapse from the phase of the reception point to the phase of the checkpoint. Then, depending on the time, the CO adjusts the time Tm2 'to the time Tm2.

p00058The CO can also perform the correction using a plurality of sinusoidal signals in the symbol. When the CPE records in the buffer the sample at the start position of the first symbol or the sample is read from the buffer at the start position of the first symbol, each of the sinusoidal signals in the first symbol is exactly at a specific point The CO takes these points as checkpoints and determines the respective phases of the checkpoints in the sinusoidal signals when the CPE takes the timestamps. For example, a checkpoint in one of the sinusoidal signals is at 0 °, a checkpoint is at 90 °, a checkpoint is at 45 °, and so on.

p00059After receiving the first symbol, the CO obtains a corresponding reception point for each of the sinusoidal signals and obtains the phase of the reception point. The CO then calculates the time that elapses from the phase of the reception point to the phase of the checkpoint. Time is a lag of a timestamp that obtains CO for each of the sinusoidal signals. The phases of these sinusoidal signals can be obtained by fast Fourier transform (FFT) in the DMT system. To improve the accuracy of the estimate and reduce the influence of noise, the offset can be the average of multiple calculations, or it can be estimated with the FEQ coefficient of an equalizer in the trained frequency domain (FEQ) placed after the FFT because the FEQ can perform offset offset offset. Because an error can be introduced in the DMT frame synchronization process, there may be a gap between these angles obtained by the CO and the CPE. The offset has a linear relationship with the frequencies of the sinusoidal signals, and the slope of the linear relationship directly reflects the frame synchronization error. The offset of each of the sinusoidal signals can be represented on a coordinate system and then these offset are joined by a straight line. The slope of the straight line is precisely the lag of the timestamp taken by the CO due to the synchronization error. Affected by factors such as noise, these angle errors obtained by real calculations may not be strictly on a straight line. The CO can obtain an optimal straight line by approximation based on a certain optimization algorithm (for example, the least squares method) so that the CO can calculate the error of the timestamps measured at the other end and based on This error corrects the Tm2 'timestamp to obtain the Tm2 timestamp.

p00060Considering the characteristics of the xDSL system, these angle errors can also be obtained using information from the FEQ, and then the time Tm2 'is adjusted similarly to the time Tm2.

p00061In step 30, the CO transmits a second symbol, and obtains the time Tm1 indicating the moment at which the second symbol is transmitted.

p00062The CO transmits a second symbol, which can also be a DMT frame. When the CO records in the buffer a sample at a start position of the second symbol or a D / A module of the CO reads from the buffer a sample at the start position of the second symbol, the CO obtains a time value of a local clock on the side of the CO (that is, an action is initiated to obtain the timestamps) and obtains the time Tm1. A specific point at which the action is initiated to obtain time is also determined by negotiation between the CO and the CPE, and any position of the second symbol can be used as a specific point. In the following embodiments, the starting position of the second symbol is taken as an example.

p00063In step 40, the CPE receives the second symbol transmitted by the CO, and obtains the exact time Ts1 indicating the instant at which the second symbol is received.

p00064When the buffer is recorded in the buffer or an A / D module reads from the buffer the sample at the start position of the second symbol, the CPE initiates an action to obtain the timestamps and records as time Ts1 'the time value of the local clock on the CPE side. Because the CPE also calculates the frame limit by a certain algorithm, an error can be introduced in the determination of the starting position of the second symbol and the CPE also has to correct the obtained Ts1 'time.

p00065Depending on a phase difference between the phase of the reception point and a phase of the checkpoint of a sinusoidal signal (or a cosinusoidal signal) in the second symbol, the CPE corrects the time stamp Ts1 'to obtain the time Ts1 where the time stamp Ts1 is the time that indicates the instant at which the checkpoint should be received. The receiving point is a signal point at which the CPE initially receives the second symbol, and the checkpoint is a signal point at which the CO initially transmits the second symbol.

p00066When the CPE uses a sinusoidal signal in the second symbol, a phase of a corresponding point in this sinusoidal signal is corrected when the CO initiates the time stamp registration action, so that this point of the sinusoidal signal can be taken as a checkpoint and a phase of the point is obtained, for example 0 °. In this way, the CPE can make a correction based on this checkpoint.

p00067The CPE takes as the reception point the corresponding point of the sinusoidal signal indicating the moment at 5

p00069that the CPE receives the second symbol, and obtains a phase from this point. The CPE then calculates the time that elapses from this phase to a phase of a closer checkpoint, and adjusts the time Ts1 'to the time Ts1 based on that time.

p00070The CPE can also use a plurality of sinusoidal signals in the second symbol. The CPE has obtained the phases of the corresponding points of these sinusoidal signals when the CO has generated the timestamps; for example, a corresponding point of one of the sinusoidal signals is at 0 °, one is at 90 °, one is at 45 °, and so on. In this way, the CPE can take as a checkpoint the corresponding point of each sinusoidal signal. After receiving the second symbol, the CPE obtains the position in which the CPE makes a time stamp on each of the sinusoidal signals and takes these points as reception points. The CPE then calculates the time that elapses from the phase of a reception point to the phase of a checkpoint. Time is exactly a lag of the timestamp made by the CPE in each of the sinusoidal signals. The angles of these sinusoidal signals can be obtained using the FFT in the DMT system. To improve the accuracy of the estimate and reduce the influence of noise, the offset can be obtained by averaging multiple calculations or by training an equalizer in the frequency domain (FEQ) after the FFT. Because the FEQ performs offset offset angle, the trained FEQ coefficient can also be used to estimate the angle offset of each of the sinusoidal signals. Because the synchronization of the DMT frame may have an error, there may be a gap between these angles obtained by the CPE and the CO. These offsets have a linear relationship with the frequencies of the sinusoidal signals, and a slope of the linear relationship directly reflects the frame synchronization error. The offset of each of the sinusoidal signals can be represented on a coordinate system, and these offset are joined by a straight line; and a slope of the straight line is precisely the offset of the timestamps obtained by the CPE due to the synchronization error. Affected by factors such as noise, these angle errors obtained by real calculations may not be strictly on a straight line. Consequently, the CPE can calculate the optimal straight line by approximation based on a given optimization algorithm (for example, the least squares method). In this way, the CPE calculates the offset of the timestamps obtained by the CPE and, depending on the offset, corrects the time Ts1 'to obtain the time Ts1.

p00071In step 50, the CPE obtains the time Tm2 and the time Tm1 from the CO.

p00072The CO transmits the time Tm1 and Tm2 to the CPE through a message channel.

p00073The CPE obtains a propagation delay of the CO and a propagation delay of the CPE.

p00074In FIG. 4 a propagation delay from CO to CPE is shown and includes:

p00075a delay of a digital transmission circuit of the CO denoted by! t1, which includes a delay of a BUF 201 and a delay of a D / A 202; and a delay of a CPE digital reception circuit denoted by! t1 ', which includes a delay of a BUF 207 and a delay of a D / A 206. In some systems, the delays! t1 and! t1' are fixed and can be read directly from the team. In the calculation of the delay both delays must be included. In some other systems, the delays! T1 and! T1 'are not fixed so that they must be excluded from the calculation. It may also be possible that part of both delays is fixed, and then, during the calculation, only the fixed part of the delay is included;

p00076a delay of an analog transmission circuit 203 of the CO denoted by! t2 and a delay of an analog reception circuit 205 of the CPE denoted by! t2 '. Both delays! T2 and! T2 'are due to the devices, and can be obtained at the factory or by exchanging information between the CPE and the CO; and

p00077a delay of a symbol in a twisted pair 204 from the CO to the CPE denoted by! t3, which is unknown.

p00078In FIG. 5 a propagation delay from the CPE to the CO is shown and includes:

p00079a delay of a digital transmission circuit of the CPE denoted by! t4, which includes a delay of a 2001 BUF of a CPE and a delay of a 2002 D / A of a CPE; and a delay of a digital reception circuit of the CO denoted by! t4 ', which includes a delay of a 2006 D / A of the CO and a delay of a BUF 2007. In some systems, both delays! t4 and! t4 'are fixed and can be read directly from the equipment. In some other systems, none of the delays are fixed, and then none of the delays are included during the calculation;

p00080a delay of a 2003 analogue transmission circuit of the CPE denoted by! t5 and a delay of a 2005 analogue reception circuit of the CO denoted by! t5 '. Because both delays! T5 and! T5 'are due to the devices, they can be obtained at the factory or by exchanging information between the CPE and the CO; and

p00081a delay of a signal in a twisted pair 2004 from the CPE to the CO denoted by! t6, which is unknown.

p00082The CO transmits the delays! T1,! T2,! T4 'and! T5' to the CPE through a message channel or the CPE obtains data

p00084previously stored

p00085In step 60, the CPE calculates a gap between a CPE clock and a CO clock, and sets the CPE clock to offset function. The CPE calculates the offset between the CPE clock and the CO clock according to:

p00086Offset = Ts1 - Tm2 - Delay1, and

p00087Offset = Tm2 - Ts2 + Delay2. During the calculation process, the CPE establishes a calculation model and separates Delay1 and Delay2. The CPE stores the mathematical relationship between Delay1 and Delay2, for example, the proportion of! T3 = 0.9! T6 or! T6 = 0.9! T3. The

p00088proportion can be obtained by statistics. The offset can be obtained with the following equations:

p00089Offset = Ts1 - Tm1 - Delay1 = Ts1 - Tm1 - (! T1 +! T2 +! T3 +! T1 '+! T2')

p00090Offset = Ts2 –Tm2 + Delay2 = Ts2 - Tm2 + (! T4 +! T5 +! T6 +! T5 '+! T4')

p00091or

p00092Offset = Ts1 - Tm1 - Delay1 = Ts1 - Tm1 - (! T2 +! T3 +! T2 ')

p00093Offset = Ts2 - Tm2 + Delay2 = Ts2 - Tm2 + (! T5 +! T6 +! T5 ') The delays! T3 and! T6 are approximately identical or have a proportion ratio. It is assumed that the delays! T3 and! T6 are approximately identical. The offset can be estimated using the following equation:

p00094Offset = (Ts1 - Tm1 - (! T1 +! T2 +! T1 '+! T2') + Ts2 - Tm2 + (! T4 +! T5 +! T4 '+! T5')) / 2

p00095or

p00096Offset = (Ts1 - Tm1 - (! T2 +! T2 ') + Ts2 - Tm2 + (! T5 +! T5')) / 2 Delay1 and Delay2 delays can be obtained with the estimated offset:

p00097Delay1 = Ts1 - Tm1 - Offset

p00098Delay2 = Ts2 - Tm2 + Offset After obtaining the offset between the CPE clock and the CO clock, the CPE obtains the local clock time value and adjusts the local clock time based on the local clock time obtained and the lag.

p00099In the embodiments described above, the CPE first transmits a symbol, and then the CO receives the symbol and subsequently transmits a symbol. In the actual monitoring process, it is also possible for the CO to transmit a symbol and then the CPE receives the symbol and subsequently transmits a symbol. The last case will be described in the second embodiment below, shown in FIG. 6 a specific process of it.

p00100In step 15, the CO transmits a second symbol, and obtains the time Tm1 indicating the moment at which it

p00101It transmits the second symbol. During initialization, the CO transmits a second symbol. When the CO records in a buffer or a D / A module of the CO reads from the buffer a sample at a specific position of the second symbol, an action is initiated to obtain timestamps to read a time value of a clock Local CO and get a time stamp Tm1. The second symbol can be a DMT frame. Through negotiation between the CO and the CPE a specific point is also determined at which the action of obtaining the timestamps is activated. Any position in the second symbol can be used as a specific point. From now on, in this embodiment, a starting position of the second symbol will be taken as an example.

p00102In step 25, the CPE receives the second symbol transmitted by the CO, and obtains the exact time Ts1 indicated by the

p00103instant in which the second symbol is received. When the CPE records in the buffer or an A / D module reads from the buffer a sample of the starting position of the second symbol, an action is initiated to obtain timestamps in order to obtain a clock time value CPE local, denoted by Ts1 '. Because the CPE calculates a frame limit using a given algorithm, an error can be introduced when calculating the starting position using that algorithm. In this case, the CPE has to correct the time Ts1 '. The correction method used

p00105here it is the same as that used by the CPE in the first embodiment.

p00106In step 35, the CPE transmits a first symbol, and obtains the time Ts2 indicating the instant at which the CPE transmits the first symbol.

p00107During initialization, the CPE transmits a first symbol, which can also be a DMT frame.

p00108When a buffer is read in the buffer memory or a D / A module reads from the buffer a specific position of the first symbol, an action is initiated to obtain timestamps in order to read a time value of the local clock of the CPE, denoted by Ts2. Through negotiation between the CO and the CPE a specific point is also determined at which the action of obtaining the timestamps is initiated. Any position in the first symbol can be used. From now on, in this embodiment, a starting position of the first symbol will be taken as an example.

p00109In step 45, the CO receives the first symbol transmitted by the CPE, and obtains the exact time Tm2 that indicates the moment at which the first symbol was received.

p00110The CO receives the first symbol transmitted by the CPE. When recording in the buffer or an A / D module reads from the buffer a sample at the start position of the first symbol, an action is initiated to obtain timestamps to read a time value of the local CO clock , denoted by Tm2 '. Because the CO calculates a frame limit using a certain algorithm, the CO has to correct the time stamp Tm2 '. The correction method used here is the same as that used by the CO in the first embodiment.

p00111In step 55, the CPE obtains the time Tm1 and Tm2 obtained by the CO.

p00112The CO transmits the time Tm1 and Tm2 to the CPE through a message channel.

p00113The CPE obtains a CO delay and a CPE delay:

p00114a delay of a digital transmission circuit of the CO denoted by! t1, which includes a delay of a BUF 201 of the CO and a delay of a D / A 202; and a delay of a CPE digital reception circuit denoted by! t1 ', which includes a delay of a BUF 207 of the CPE and a delay of a D / A 206. In some systems, both delays are fixed and can be read directly from the team. In the calculation of the propagation delay, both delays must be included. In some other systems, none of the delays are fixed so that they must be excluded during the calculation. It may also be possible that a part of both delays is fixed, and then, during the calculation, only the fixed part of the delay is included;

p00115a delay of an analog transmission circuit 203 of the CO denoted by! t2 and a delay of an analog reception circuit 205 of the CPE denoted by! t2 '. Both delays! T2 and! T2 'are due to the equipment, and can be obtained at the factory or by exchanging information between the CPE and the CO;

p00116a delay of a symbol in a twisted pair 204 from the CO to the CPE denoted by! t3, which is unknown.

p00117In FIG. 5 a propagation delay from the CPE to the CO is shown and includes:

p00118a delay of a digital transmission circuit of the CPE denoted by! t4, which includes a delay of a 2001 BUF of the CPE and a delay of a D / A 2002 of the CPE; and a delay of a digital reception circuit of the CO denoted by! t4 ', which includes a delay of a 2006 D / A of the CO and a delay of a BUF 2007. In some systems, both delays are fixed and are They can read directly from the team. In some other systems, none of the delays are fixed, and then none of the delays are included during the calculation;

p00119a delay of a 2003 analog transmission circuit of the CPE denoted by! t5 and a delay of a 2005 analogue reception circuit of the CO denoted by! t5 '. Because both delays! T5 and! T5 'are due to the devices, they can be obtained at the factory or by exchanging information between the CPE and the CO;

p00120a delay of a signal in a twisted pair 2004 from the CPE to the CO denoted by! t6, which is unknown.

p00121The CO transmits the delays! T1,! T2,! T4 'and! T5 to the CPE through a message channel; or alternatively, the CPE obtains previously stored data and thus the CO may also not transmit the information.

p00122In step 65, the CPE calculates a gap between a CPE clock and a CO clock, and adjusts the CPE clock time based on this offset.

p00123The CPE calculates the offset according to the following equations:

p00124Offset = Ts1 - Tm2 - Delay1, and 8

p00126Offset = Ts2 - Tm2 + Delay2.

p00127During the calculation process, the CPE establishes a calculation model and separates Delay1 and Delay2. The CPE stores the mathematical relationship between Delay1 and Delay2, for example, the proportion of! T3 = 0.9! T6 or! T6 = 0.9! T3. The specific mathematical relationship can be obtained through statistics. The offset is obtained with the following equations:

p00128Offset = Ts1 - Tm1 - Delay1 = Ts1 - Tm1 - (! T1 +! T2 +! T3 +! T1 '+! T2')

p00129Offset = Ts2 - Tm2 + Delay2 = Ts2 - Tm2 + (! T4 +! T5 +! T6 +! T5 '+! T4')

p00130or

p00131Offset = Ts1 - Tm1 - Delay1 = Ts1 - Tm1 - (! T2 +! T3 +! T2 ')

p00132Offset = Ts2 - Tm2 + Delay2 = Ts2 - Tm2 + (! T5 +! T6 +! T5 ')

p00133Since the delays! T3 and! T6 are approximately identical or have a ratio ratio, the offset can be estimated.

p00134After obtaining the offset, the delays Delay1 and Delay2 can be obtained:

p00135Delay1 = Ts1 - Tm1 - Offset

p00136Delay2 = Ts2 - Tm2 + Offset

p00137After obtaining the offset between the CPE clock and the CO clock, the CPE obtains a local clock time from the CPE, and adjusts the local clock time based on the local clock time obtained and the estimated offset.

p00138A third embodiment of the present invention provides a method for time synchronization of an xDSL. The method is applicable to the case in which the delays Delay1 and Delay2 can be obtained by SELT or DELT or in other ways. The method includes the following steps.

p00139In step 1, the CO transmits a symbol, and obtains the time Tm1 indicating the instant at which the symbol is transmitted (or the CPE transmits a symbol, and obtains the time Ts2 indicating the moment at which the symbol is transmitted symbol), and this symbol can be a DMT frame.

p00140In an initialization stage, the CO transmits the symbol. When the CO records data sampled at a specific position of this symbol in a buffer memory or a CO D / A module reads from the buffer the data sampled at the specific position of this symbol, the CO initiates an action to obtain the timestamp, reads the local clock time of the CO and obtains the time Tm1. By negotiation between the CO and the CPE a specific point is also determined at which the action to obtain the timestamp is initiated, and any position in this symbol can be used. From now on in this embodiment, a starting position of this symbol is taken as an example.

p00141In step 2, the CPE receives the symbol transmitted by the CO, and obtains the time Ts1 of reception (or the CO receives the symbol transmitted by the CPE, and obtains the time Tm2 indicating the time at which the CP was received. symbol).

p00142When the CPE writes in the buffer memory data sampled at the start position of this symbol or an A / D module reads from the buffer the data sampled at the start position of this symbol, the CPE initiates the mark obtaining action of time and read the local Ts1 'time of the CPE. Because the CPE calculates a frame limit by a certain algorithm, the starting position calculated by the algorithm may have an error. In this case, the CPE has to correct the time Ts1 '. The correction method is the same as that of the CPE of the first embodiment.

p00143In step 3, the CPE obtains the time Tm1 transmitted by the CO (or the CPE obtains the time Tm2 transmitted by the CO).

p00144The CO transmits to the CPE the time Tm1 (or the time Tm2) through a message channel.

p00145In step 4, the CPE calculates a gap between a CPE clock and a CO clock according to Offset = Ts1 - Tm2 - Delay1 or Offset = Ts2 - Tm2 + Delay2.

p00146As Delay1 (or Delay2) has been measured, the offset can be calculated.

p00147In step 4, the CPE obtains a local clock time value, and adjusts the local clock time according to the time obtained from the local clock and the offset.

p00148A fourth embodiment of the present invention provides a method for time synchronization of 9

p00150a DSL. As there is a delay due to the processing of the equipment, the equipment delay must be taken into account when calculating the propagation delay of a symbol. Thus, the delay of the CO equipment may not be necessary when the CPE calculates the offset. The specific steps are as follows:

p00151In a first step, the CO transmits a second symbol, and obtains the time that indicates the moment in which the second symbol is transmitted.

p00152During initialization, when the CO records in a buffer or the CO reads it from the buffer a sample at a specific position of the second symbol, an action is initiated to obtain the timestamps to read the time Tm1 of the local clock.

p00153The CO equipment obtains a delay! T1 of digital transmission and a delay! T2 of analog transmission of the CO, and processes the time when the equipment of the CO transmits the second symbol. Specifically, Tm1 = Tm1 +! T1 +! T2; and if the digital transmission delay is not fixed, it can be excluded, and in this case, Tm1 = Tm1 +! t2.

p00154In a second step, the CPE receives the second symbol, and obtains the time that indicates the moment in which the CPE receives the second symbol.

p00155When the CPE records in the buffer or an A / D module reads from the buffer a sample at the start position of the second symbol, the CPE initiates an action to obtain the timestamps to read a local clock time value of the CPE denoted by Ts1 '. Because the CPE calculates a frame limit using a certain algorithm, an error can be introduced when the starting position is calculated using the algorithm. In this case, the CPE has to correct the time Ts1 'and the correction method used here is the same as that of the CPE of the first embodiment.

p00156In a third step, the CPE transmits a first symbol, and obtains the time that indicates the moment in which the first symbol is transmitted.

p00157During initialization, the CPE transmits the first symbol. When the CPE records in the buffer a sample at a specific position of the first symbol or a D / A module reads from the buffer a sample at a specific position of this symbol, an action is initiated to obtain the timestamps for reading Ts2 time of the local clock.

p00158In a fourth step, the CO receives the first symbol, and obtains the time indicated by the moment at which the first symbol is received.

p00159The CO receives the first symbol transmitted by the CPE. When the CO records in the buffer a sample at the start position of the first symbol or an A / D module reads from the buffer a sample at the specific position of this symbol, an action is initiated to obtain the timestamps for read the value of the time Tm2 'of the local clock. Because the CO calculates the frame limit by a certain algorithm, an error can be introduced when the starting position is calculated by the algorithm. In this case, the CO has to correct the time Tm2 ', and the correction method used here is the same as that of the first embodiment.

p00160Through negotiation between the CO and the CPE a specific point is determined at which the action is initiated to obtain the timestamps. Any position of the first symbol can be used, for example, the starting position of the first symbol.

p00161The CO obtains a delay of the digital reception circuit, denoted by! T4, and a delay of the analog reception circuit, denoted by! T5, of the CO, and processes the timestamps indicating the instant at which the equipment The CO receives the first symbol. Specifically, Tm2 = Tm2 -! T4 -! T5. If the digital reception delay is not fixed, it can be excluded and, therefore, Tm2 = Tm2 -! T5.

p00162In a fifth step, the CO transmits to the CPE the time Tm1 and the time Tm2 through a message channel, and the CPE calculates a gap between a clock of the CPE and a clock of the CO.

p00163The CPE obtains a delay! T1 'from the digital reception circuit, a delay! T2' from the analog reception circuit, a delay! T4 'from the digital transmission circuit, and a delay! T5' from the analog transmission circuit of the CPE.

p00164The CPE calculates the offset according to:

p00165Offset = Ts1 - Tm1 - Delay1 = Ts1 - Tm1 - (! T1 '+! T2' +! T3), and

p00166Offset = Ts2 - Tm2 + Delay1 = Ts2 - Tm2 + (! T4 '+! T5' +! T6)

p00167Alternatively, the digital reception delay and the digital transmission delay are not fixed and therefore

p00169exclude, and in this case the CPE calculates the lag according to:

p00170Offset = Ts1 - Tm1 - Delay1 = Ts1 - Tm1 - (! T2 '+! T3), and

p00171Offset = Ts2 - Tm2 + Delay1 = Ts2 - Tm2 + (! T5 '+! T6)

p00172In this process, the CPE can also process the timestamp Ts2 indicating the instant of transmission of the second symbol and the timestamp Ts1 indicating the instant of receipt of the first symbol. For example, Ts1 = Ts1 -! T1 '-! T2' or Ts1 = Ts1 -! T2 '; Ts2 = Ts2 -! T4 '-! T5' or Ts2 = Ts2 -! T5 '.

p00173In this way the CPE calculates the offset according to:

p00174Offset = Ts1 - Tm1 - Delay1 = Ts1 - Tm1 -! T3, and

p00175Offset = Ts2 - Tm2 + Delay1 = Ts2 - Tm2 +! T6.

p00176The lag is estimated based on the ratio between! T3 and! T6 or the assumption that the delays! T3 and! T6 are approximately identical.

p00177In a sixth step, the CPE adjusts the CPE clock based on the offset.

p00178The CPE obtains the local clock time value, and adjusts the local clock time based on the local clock time obtained and the estimated offset.

p00179In the embodiments described above, it is the CPE that adjusts the local CPE clock so that the CPE clock synchronizes with the CO clock. In practice, the CO can also set the local clock of the CO so that the local clock of the CO is synchronized with the clock of the CPE, in which case the synchronization method is similar to the synchronization method in which the CPE Set your local clock.

p00180The method described in the embodiments indicated above takes into account the influence of the sampling rate, and can be executed a plurality of times.

p00181An embodiment of the present invention provides an xDSL communication system. As shown in FIG. 7, the communication system includes a CO 100 and a CPE 200.

p00182The CPE 200 transmits a symbol and obtains the time Ts2 indicating the instant at which the first symbol is transmitted. The CPE 200 transmits the first symbol, which is a DMT frame agreed by negotiation between the CO 100 and the CPE 200 at an initialization stage. The CO 100 and the CPE 200 agree by negotiation a point of the first symbol as a reference point, which can be any position of the first symbol. From here on, an initial position of the first symbol is taken as an example.

p00183When the CPE 200 writes in a buffer memory data sampled at the start position of the first symbol, or reads from the buffer the data sampled at the start position, the CPE 200 starts an action to obtain the timestamps and reads the Ts2 time value of the local clock of the CPE 200.

p00184The CO 100 receives the first symbol transmitted by the CPE 200, and obtains the time Tm2 that indicates the moment at which the first symbol was received. When the CO 100 records in the buffer the sampled data at the start position of the first symbol or reads from the buffer the sampled data at the start position of the first symbol, the CO 100 starts the action to obtain the timestamps and read the time value Tm2 'of the local clock. Because the CO 100 recovers a frame limit using a certain algorithm, an error may occur when determining the starting position of the first symbol and, therefore, the CO 100 has to correct the time.

p00185According to a phase difference between a phase of the reception point and a phase of the check point of a sinusoidal signal (or a cosinusoidal signal) of the first symbol, the CO 100 corrects the time stamp Tm2 'to obtain the mark of Tm2 time that indicates the moment at which the CO 100 should receive a checkpoint. The reception point is a signal point at which the CO 100 receives the first symbol initially, and the aforementioned check point is a signal point at which the CPE 200 transmits the first symbol initially.

p00186When the CO 100 corrects the time Tm2 'according to a sinusoidal signal in the first symbol:

p00187when the CPE 200 starts obtaining a time stamp so that during the correction process a phase of a corresponding point of the sinusoidal signal is set (for example, 0º, 45º, 90º or any other angle), so that The CO 100 can take this point as a checkpoint and obtain a checkpoint phase. In the following embodiments, 0 ° is taken as an example.

p00189The CO 100 obtains a position of the sinusoidal signal in which the CO 100 starts obtaining the time stamp (said position being a receiving point in which the CO 100 receives the first symbol), and calculates the time that elapses from the phase of the reception point to the phase of the checkpoint. Then, depending on the time, the CO 100 adjusts the time Tm2 'to the time Tm2.

p00190The CO 100 can also perform the correction using a plurality of sinusoidal signals in this symbol. When the CPE 200 records in the buffer the starting position of the first symbol or reads from the buffer the starting position of the first symbol, each of the sinusoidal signals in the first symbol is at a specific point. The CO 100 takes these points as checkpoints and determines the respective phases of the checkpoints in these sinusoidal signals when the CPE 200 took the timestamps. For example, a checkpoint in one of the sinusoidal signals is at 0 °, one is at 90 °, one is at 45 °, and so on.

p00191After receiving the first symbol, the CO 100 obtains the corresponding reception point for each of the sinusoidal signals and obtains the phase of the reception point. Next, the CO 100 calculates the time that elapses from the phase of the reception point to the phase of the checkpoint. The time is a lag of the timestamp calculated by the CO 100 for each of the sinusoidal signals. The phases of these sinusoidal signals can be obtained by FFT in the DMT system. To improve the accuracy of the estimate and reduce the influence of noise, the offset can be obtained by averaging after multiple calculations or training an FEQ after FFT. Because the FEQ compensates for an angle offset, the coefficient of the trained FEQ can also be used to estimate the angle offset of each of the sinusoidal signals. Because the synchronization of the DMT frame may have an error, there may be mismatches between these angles obtained by the CO 100 and the CPE 200. These offset have a linear relationship with the frequencies of the sinusoidal signals, and a slope of the relationship linear directly reflects the frame synchronization error. The offset of each of the sinusoidal signals can be represented on a coordinate system and these offset are connected by a straight line; and a slope of the straight line is precisely the offset of the timestamps taken by the CO 100 due to the synchronization error. Affected by factors such as noise, these angle errors obtained by real calculations may not be strictly on a straight line. Consequently, the CO 100 can calculate an optimal straight line by approximation according to a certain optimization algorithm (for example, the least squares method). In this way, the CO 100 calculates the error of the timestamps taken by the CPE and, based on this error, corrects the timestamp Tm2 'to obtain the timestamp Tm2.

p00192Considering the characteristics of the xDSL system, these angle errors can also be obtained using information from the FEQ, and then the time Tm2 'is adjusted similarly to the time Tm2.

p00193The CO 100 transmits a second symbol and obtains the time Tm1 indicating the moment at which the second symbol is transmitted. When the CO 100 records in a buffer memory sampled data in a starting position of the second symbol or reads sampled data in the starting memory from the buffer, the CO 100 initiates an action to obtain timestamps and reads the value of time Tm1 of the time of the local clock of the CO 100. A specific point at which the action is initiated to obtain the timestamps is also determined by negotiation between the CO and the CPE, and any position of the second symbol can be used. In this embodiment, the starting position of the second symbol is taken as an example.

p00194The CPE 200 receives the second symbol transmitted by the CO 100, and obtains the time Ts1 indicating the instant at which the second symbol is received. When the CPE 200 records in the buffer the sampled data at the start position of the second symbol or reads from the buffer the sampled data at the start position of the second symbol, the CPE 200 starts the action to obtain the timestamps and read the time value Ts1 'of the local clock time. Because the CPE 200 recovers a frame limit by a certain algorithm, the CPE 200 corrects the time Ts1 'to obtain time Ts1 in a similar way as the CO does

p00195100.

p00196The CO 100 transmits to the CPE 200 the time Tm1 and the time Tm2 through a message channel. If the CPE 200 does not store the transmission delay and the reception delay of the CO 100, the CO 100 transmits to the CPE 200, by interaction with the CPE 200, the transmission delay and the reception delay of the CO 100 through of a message channel.

p00197The transmission delay and the reception delay of the CO 100 include a delay! T1 of the digital transmission circuit, a delay! T2 of the analog transmission circuit, a delay! T5 'of the analog reception circuit, and a delay! T4 'of the digital reception circuit.

p00198The CPE 200 obtains the transmission delay and the reception delay of the CPE 200, which includes a delay! T1 'of the digital transmission circuit, a delay! T2' of the analog transmission circuit, a delay! T5 of the reception circuit analog, and a delay! t4 of the digital reception circuit. These delays can be read directly from the CPE 200.

p00200The CPE 200 calculates a lag between a clock of the CPE 200 and a clock of the CO 100 based on Ts1, Ts2, Tm1, Tm2, ​​the delay of the CO 100, and the delay of the CPE 200.

p00201Specifically, the CPE 200 calculates the offset according to:

p00202Offset = Ts1 - Tm1 - Delay1 = Ts1 - Tm1 - (! T1 +! T2 +! T3 +! T1 '+! T2')

p00203Offset = Ts2 - Tm2 + Delay2 = Ts2 - Tm2 + (! T4 +! T5 +! T6 +! T5 '+! T4')

p00204or

p00205Offset = Ts1 - Tm1 - Delay1 = Ts1 - Tm1 - (! T2 +! T3 +! T2 ')

p00206Offset = Ts2 - Tm2 + Delay2 = Ts2 - Tm2 + (! T5 +! T6 +! T5 ')

p00207where the CPE 200 stores in its interior the mathematical relationship between the delay Delay1 and the delay Delay2.

p00208Specifically, it can be statistically determined that delay! T3 and delay! T6 are approximately equal to each other or have a proportional relationship, for example,! T3 = 0.9! T6 or! T6 = 0.9! T3.

p00209After obtaining the offset, the CPE 200 obtains a local clock time value and adjusts the local clock time based on the local clock time and the offset obtained.

p00210In the communication system described above, it is the CPE 200 that adjusts the time of the local clock so that the local clock of the CPE 200 is synchronized with the clock of the CO 100. Alternatively, the CO 100 can also set the clock of the CO 100 so that the clock of the CPE 200 is synchronized with the clock of the CO 100; The synchronization process of this alternative is the same as the synchronization process in which the CPE 200 clock is set.

p00211The present invention also provides an xDSL device, which can be used for the CO and for the CPE. As shown in FIG. 8, the equipment includes a transmission unit 300, a reception unit 400 and a processing unit 600.

p00212The transmission unit is configured to transmit a first symbol and obtain the time Ts2 indicating the moment at which the first symbol was transmitted.

p00213The receiving unit is configured to receive a second symbol transmitted by a second device and obtain the time Ts1 indicating the moment at which the second symbol was received; and obtain the time Tm2 that indicates the moment in which the second device has received the first symbol and the time Tm1 that indicates the moment in which the second device has transmitted the second symbol.

p00214The processing unit is configured to obtain a delay of the DSL equipment, calculate a lag between a clock of the DSL equipment and a clock of the second equipment based on Ts1, Ts2, Tm1, Tm2 and the DSL equipment delay, and adjust the DSL device clock depending on the offset.

p00215Specifically, the transmission unit 300 transmits the first symbol, and obtains the time Ts2 indicating the instant at which the first symbol was transmitted. The first symbol may be a training signal transmitted during an initialization stage, and this signal may be a DMT frame.

p00216When the transmission unit 300 writes in a buffer memory data sampled at a start position of the first symbol, or reads from the buffer memory the data sampled at the start position of the first symbol, the transmission unit 300 initiates an action to obtain Timestamps and reads the local Ts2 time.

p00217The receiving unit 400 receives the second symbol transmitted by the other terminal and obtains the time Ts1 indicating the instant at which the second symbol was received. The second symbol may be a training signal transmitted during the initialization stage.

p00218The receiving unit 400 also includes an acquisition module and a correction module. The obtaining module receives the second symbol, obtains the time Ts1 'from the DSL device clock, and obtains the time Tm2 that indicates the moment in which the second equipment has received the first symbol and the time Tm1 that indicates the moment in the that the second device has transmitted the second symbol.

p00219The correction module, based on a phase difference between a phase of the reception point and a phase of the checkpoint of a signal of the second symbol, corrects the time stamp Ts1 'to obtain the time stamp Ts1 indicating the moment in which the obtaining module should receive a checkpoint, and obtains the time stamp Ts1 for use as a time that indicates the time at which the module

p00221Obtaining has received the second symbol. The reception point is a signal point at which the obtaining module initially receives the second symbol, and the check point is a signal point at which the second device initially transmits the second symbol.

p00222When the obtaining module records in the buffer memory sampled data at a starting position of the second symbol or reads from the buffer the sampled data at the starting position of the second symbol, the obtaining module initiates an action to obtain the marks of time and read the time value Ts1 'of the local clock time. Because a frame limit of the second symbol is retrieved using a certain algorithm, an error may occur when the limit is positioned. Consequently, the correction module, according to a phase difference between a phase of the reception point and a phase of the checkpoint of a signal in the second symbol, corrects the time stamp Ts1 'to obtain the time stamp Ts1 indicating the moment at which the acquisition module should receive a checkpoint.

p00223The correction module obtains a position in a sinusoidal signal in which the module starts obtaining the time stamp, takes this position as a reception point, and calculates the time that elapses from the phase of the reception point to the phase of the checkpoint. Then, depending on this time, the correction module corrects the time Ts1 'to obtain the time Ts1.

p00224The correction module can also use a plurality of sinusoidal signals in the second symbol. The correction module determines the respective angles of the corresponding points (that is, the checkpoints) in these sinusoidal signals when the second equipment takes the timestamps. For example, a checkpoint in one of the sinusoidal signals is at 0 °, one is at 90 °, one is at 45 °, and so on. Therefore, after the obtaining module has received the second symbol, the correction module obtains the positions in which the obtaining module takes the timestamps, which are reception points, and calculates the time that elapses. from the phase of each reception point to the phase of a corresponding checkpoint. The angles of these sinusoidal signals can be obtained by the FFT in the DMT system. To improve the accuracy of the estimate and reduce the influence of noise, the offset can be obtained by averaging after multiple calculations or training an FEQ after FFT. Because the FEQ performs offset offset angle, the coefficient of the trained FEQ can also be used to estimate the angle offset of each of the sinusoidal signals. Because the synchronization of the DMT frame may have an error, there may be mismatches between the angles obtained by the correction module and the angles obtained by the other equipment. These offsets have a linear relationship with the frequencies of the sinusoidal signals, and a slope of the linear relationship directly reflects the frame synchronization error. Therefore, the correction module can represent the offset of each of the sinusoidal signals on a coordinate system and connect these offset using a straight line, its slope is precisely the offset of the timestamps taken by the CPE due to the sync error Affected by factors such as noise, these angle errors obtained by real calculations may not be strictly on a straight line. The CPE can calculate an optimal straight line by approximation based on a certain optimization algorithm (for example, the least squares method). The correction module calculates the error of the CPE timestamp and, based on this error, corrects the time Ts1 'to obtain the time Ts1.

p00225The correction module can also be located in the communication equipment, being independent of the receiving unit 400.

p00226The receiving unit 400 can also receive, by means of a message, information of the channel transmitted by the second device, including the time Tm2 that indicates the moment in which the second device has received the first symbol, the time Tm1 that indicates the instant in which the second device has transmitted the second symbol, and the transmission delay and the reception delay of the second equipment. The transmission delay and the reception delay of the second device include: a delay! T1 of the digital transmission circuit, a delay! T2 of the analog transmission circuit, a delay! T5 'of the analog reception circuit and a delay! T4' of the digital reception circuit.

p00227The second device can also process the time Tm2 and the time Tm1 using the delay data of the second device. In this way, the second device only has to transmit to the DSL the time Tm1 and the time Tm2 that are processed; for example, Tm1 = Tm1 +! t1 +! t2 or Tm1 = Tm1 +! t2, Tm2 = Tm2 -! t5 -! t4 or Tm2 = Tm2 -! t5.

p00228DSL equipment can also process time Ts1 and time Ts2; that is, Ts1 = Ts1 -! t1 '-! t2' or Ts1 = Ts1! t2 '; Ts2 = Ts2 -! T4 '-! T5' or Ts2 = Ts2 -! T5 '.

p00229The processing unit 600 obtains the DSL equipment challenge, calculates a gap between the local equipment clock and the second equipment clock based on the time Ts2 obtained by the transmission unit, Ts1, Tm2, ​​Tm1 obtained by the receiving unit , and the delay of the DSL equipment, and adjusts the clock of the DSL equipment according to said offset.

p00230The DSL equipment delay includes: a delay! T1 'of the digital transmission circuit, a delay! T2' of the analog transmission circuit, a delay! T5 of the analog reception circuit and a delay! T4 of the digital reception circuit, all which can be obtained directly when the DSL equipment is sent from the factory.

p00231The processing unit 600 calculates the offset according to:

p002325 Offset = Ts1 - Tm1 - Delay1 = Ts1 - Tm1 - (! T1 +! T2 +! T3 +! T1 '+! T2') Offset = Ts2 - Tm2 + Delay2 = Ts2 - Tm2 + (! T4 +! T5 +! T6 +! T5 '+! T4') or Offset = Ts1 - Tm1 - Delay1 = Ts1 - Tm1 - (! T2 +! T3 +! T2 ') Offset = Ts2 - Tm2 + Delay2 = Ts2 - Tm2 + (! T5 +! T6 +! T5 '); or

p0023310 The DSL device and the second device, after processing the transmitted / received symbol, calculate the offset according to: Offset = Ts1 - Tm1 - Delay1 = Ts1 - Tm1 -! t3, and Offset = Ts2 - Tm2 + Delay1 = Ts2 - Tm2 +! T6 The processing unit 600 reads the local clock time and adjusts the local time based on the local clock time 15 and the offset.

p00234The second device can be the CO or the CPE, and the DSL device can also be used as the CO or as the CPE. As can be seen from the embodiments described above, in accordance with the present invention, by correcting the local time corresponding to the timestamps, the receiver can accurately read the local time, and the offset between the CPE clock and the CO clock so that the CPE clock

p00235twenty It can be adjusted depending on the offset to achieve synchronization between the CO clock and the CPE clock. Those with a normal experience in the art can understand that all or part of the steps included in the methods of the above embodiments can be performed by means of a program that runs on associated hardware. The program can be stored in a storage medium readable by a computer, including a read-only memory (ROM), a random access memory (RAM), a magnetic disk or a compact disc (CD). The description made above includes only several embodiments of the present invention. However, the present invention is not limited to said embodiments.