Measuring Line Characteristics on the Huawei

Unlocking the Huawei gains access to a command line tool in the firmware called xdslcmd.  The tool has numerous functions. Those examined below use xdslcmd for the measurement of line characteristics.

Burakkucat (Alan) and Little_Bird (Pauline), two fellow Huawei hackers, have built some useful graphing scripts. These use GNUPlot to display the line data from the Huawei. The graphs are very insightful but they still need some explanation.

The scripts can currently plot four different graphs of line statistics.

These illustrate, on a per subcarrier basis, Bit Loading (the number of bits carried by a tone), Signal to Noise Ratio (SNR), Quiet Line Noise (QLN), and the logarithmic channel characteristic function Hlog(f).

The Bit Loading metric is a simple concept to grasp, but the last three functions, SNR, QLN, and Hlog are less understood, particularly the Hlog function.

Below, is an excerpt from the ITU-T Recommendations document G.992.3 (July 2002) where the measurement functions are defined.

  • H(f) can be used for analyzing the physical copper loop condition;
  • QLN(f) can be used for analyzing the crosstalk;
  • SNR(f) can be used for analyzing time dependent changes in crosstalk levels and line attenuation (such as due to moisture and temperature variations);
  • The combination of H(f), QLN(f) and SNR(f) can be used for trouble-shooting why the data rate cannot reach the maximum data rate of a given loop.

For reference, the function definitions for SNR, QLN and Hlog are in Clause 8.12.3.1 of G.992.3 which is entitled Digital Line Systems – Access networks (ADSL2).  [1]

Methods for obtaining QLN, SNR and HLOG (and HLIN) data from the Customer Premises Equipment are described in Technical Report TR-138 – “Accuracy Tests for Test Parameters” published November 2009 by the Broadband Forum. [2]

Below is a fuller excerpt from Clause 8.12.3.1 of G.992.3:

[T]est parameters are measured by the transmit or receive function of the PMD (Physical Media-Dependent sublayer) and shall be reported on request to the near-end management entity using the Management.Defect.indicate primitive.

Test parameters allow to debug possible issues with the physical loop and to check for adequate physical media performance margin at acceptance and after repair verification, or at any other time following the execution of initialization and training sequence of the ADSL system.

The following test parameters shall be passed on request from the receive PMD transmit function to the near-end management entity:

  • Channel Characteristics Function H(f) per subcarrier (CCF-ps);
  • Quiet Line Noise PSD QLN(f) per subcarrier (QLN-ps);
  • Signal-to-Noise Ratio SNR(f) per subcarrier (SNR-ps);
  • Line Attenuation (LATN);
  • Signal Attenuation (SATN);
  • Signal-to-Noise Margin (SNRM);
  • Attainable Net Data Rate (ATTNDR);
  • Far-end Actual Aggregate Transmit Power (ACTATP).

The following test parameters shall be passed on request from the transmit PMD transmit function to the near-end management entity:

  • Near-End Actual Aggregate Transmit Power (ACTATP).

The purposes of making the above information available are:

  • a) H(f) can be used for analyzing the physical copper loop condition;
  • b) QLN(f) can be used for analyzing the crosstalk;
  • c) SNR(f) can be used for analyzing time dependent changes in crosstalk levels and line attenuation (such as due to moisture and temperature variations);
  • d) The combination of H(f), QLN(f) and SNR(f) can be used for trouble-shooting why the data rate cannot reach the maximum data rate of a given loop.

This enhances the ADSL service maintenance and diagnostics defined in ITU-T Rec. G.992.1 by making diagnostic information available from both ends of the loop during active operation of the service.

The most detailed diagnostic information H(f) and QLN(f) would be useful during showtime, however, requesting this would place an undo computational burden on the ADSL modems.

Thus, the combination of complete information on the channel (H(f) and QLN(f)) during initialization combined with initialization and showtime SNR(f) is provided as a reasonable compromise.

This combination of data will allow greater analysis of the line conditions than traditional methods and will reduce interruptions of both the ADSL and the underlying service that traditional diagnostic methods require.

Channel Characteristics Function per subcarrier (CCF-ps)

The channel characteristics function H(f) is a quantity that is related to the values of the (complex) source and load impedance. A simplified definition is used in which source and load are the same and equal to a real value RN.

The channel characteristics function H(f) is associated with a two-port network, normalized to a chosen reference resistance RN, shall be defined as a complex value, equal to the U2/U1 voltage ratio (see Figures 8-21 and 8-22).

The channel characteristics function is the result of the cascade of three functions:

  • the transmitter filter characteristics function;
  • the channel characteristics function;
  • the receiver filter characteristics function.

The objective is to provide means by which the channel characteristics can be accurately identified. Therefore, it is necessary for the receive PMD function to report an estimate of the channel characteristics. This task may prove to be a difficult one given the fact that the receive PMD function only observes the cascade of all three elements of the channel.

The passband part of the reported H(f), which is most essential to debug possible issues with the physical loop, is not expected to significantly depend upon the receiver filter characteristics (not including receiver AGC). The receive PMD function shall therefore undo the gain (AGC) it has applied to the received signal and do a best effort attempt to remove the impact of the near-end receiver filter characteristics.

The result is then a best estimate of how the receiver views the passband channel characteristics plus the transmitter filter characteristics. Because the in-band portion of the spectrum is also expected not to significantly depend upon the transmitter filter characteristics, this result is considered a sufficient estimate of the channel characteristics for desired loop conditioning applications.

If the channel characteristics are reported to the CO-MIB (Central Office-Management Information Base) the ATU-C (ADSL Transceiver Unit at Central Office) shall do a best effort attempt to remove the impact of the near-end transmit filter characteristics from the channel characteristics measured at the ATU-R (ADSL Transceiver Unit at Remote Consumer End).

If the channel characteristics are reported to the RT-MIB (Remote Terminal-Management Information Base) the ATU-R shall do a best effort attempt to remove the impact of the near-end transmit filter characteristics from the channel characteristics measured at the ATU-C.

Two formats for the channel characteristics are defined:

  • Hlin(f): a format providing complex values in linear scale;
  • Hlog(f): a format providing magnitude values in a logarithmic scale.

The Hlin(f) shall be measured by the receive PMD function during diagnostics mode in a REVERB transmitter state. The Hlin(f) shall be sent to the far-end management entity during diagnostics mode and shall be sent on request to the near-end Management Entity during diagnostics mode.

The Hlog(f) shall be measured by the receive PMD function during diagnostics mode and initialization. The measurement shall not be updated during showtime. The Hlog(f) shall be sent to the far-end management entity during diagnostics mode and shall be sent on request to the near-end Management Entity. The near-end Management Entity shall send the Hlog(f) to the far-end Management Entity on request during showtime (see 9.4.1.10).

In diagnostics mode, both Hlin(f) and Hlog(f) shall be measured, because there may be a difference in up to what extent the receiver and/or transmitter filter characteristics can be undone in Hlin(f) versus Hlog(f).

The PMD receive function shall measure Hlin(f) and Hlog(f) with the PMD transmit function in a REVERB state. The Hlin(f) and Hlog(f) shall be measured over a 1 second time period in diagnostics mode. The ATU shall do a best effort attempt to optimize Hlog(f) measurement time in initialization, however, measuring over at least 256 symbols, with indication of the measurement period to the far-end Management Entity (in symbols, represented as 16-bit unsigned value), see 9.4.1.10).

The channel characteristics function Hlin(i × ∆f), shall be represented in linear format by a scale factor and a normalized complex number a(i) + j × b(i), where i is the subcarrier index i = 0 to NSC – 1. The scale factor shall be coded as a 16-bit unsigned integer. Both a(i) and b(i) shall be coded as a 16-bit 2′s complement signed integer. The value of Hlin(i × ∆f) shall be defined as Hlin(i × ∆f) = (scale/215) × (a(i) + j × b(i))/215. In order to maximize precision, the scale factor shall be chosen such that max(|a(i)|, |b(i)|) over all i is equal to 215 – 1. This data format supports an Hlin(f) granularity of 2–15 and an Hlin(f) dynamic range of approximately +6 dB to –90 dB. The portion of the scale factor range above 0 dB is necessary to accommodate that short loops, due to manufacturing variations in signal path gains and filter responses, may appear to have a gain rather than a loss.

An Hlin(i × ∆f) value indicated as a(i) = b(i) = –215 is a special value. It indicates that no measurement could be done for this subcarrier because it is out of the PSD mask passband (as relevant to the chosen application option – see annexes) or that the attenuation is out of range to be represented.

The channel characteristics function Hlog(f) shall be represented in logarithmic format by an integer number m(i), where i is the subcarrier index i = 0 to NSC – 1. The m(i) shall be coded as a 10-bit unsigned integer.

The value of Hlog(i × ∆f) shall be defined as Hlog(i × ∆f) = 6 – (m(i)/10). This data format supports an Hlog(f) granularity of 0.1 dB and an Hlog(f) dynamic range of approximately +6 dB to –96 dB. An Hlog(i × ∆f) value indicated as m(i) = 210 – 1 is a special value. It indicates that no measurement could be done for this subcarrier because it is out of the PSD mask passband (as relevant to the chosen application option – see annexes) or that the attenuation is out of range to be represented.

Quiet Line Noise PSD per subcarrier (QLN-ps)

The quiet line noise PSD QLN(f) for a particular subcarrier is the rms level of the noise present on the line, when no ADSL signals are present on the line.

The quiet line PSD QLN(f) per subchannel shall be measured by the receive PMD function during diagnostics mode and initialization. The measurement shall not (i.e., cannot) be updated during showtime. The QLN(f) shall be sent to the far-end transmit PMD function during diagnostics mode (see 8.15.1) and shall be sent on request to the near-end Management Entity. The near-end Management Entity shall send the QLN(f) to the far-end Management Entity on request during showtime (see 9.4.1.10).

The objective is to provide means by which the quiet line PSD can be accurately identified. Therefore, it would be necessary for the receive PMD function to report an estimate of the quiet line PSD. This task may prove to be a difficult one given the fact that the receive PMD function observes the noise through the receiver filter. The passband part of the reported QLN-ps, which is most essential to debug possible issues with the physical loop, is not expected to significantly depend upon the receiver filter characteristics (not including receiver AGC). The receive PMD function shall therefore undo the gain (AGC) it has applied to the received signal and do a best effort attempt to remove the impact of the near-end receiver filter characteristics. The result is then a best estimate of how the receiver views the passband quiet line PSD. This result is considered a sufficient estimate of the quiet line PSD for desired loop conditioning applications.

The receive PMD function shall measure the QLN(f) in a time interval where no ADSL signals are present on the line (i.e., near-end and far-end transmitter inactive). The quiet line PSD QLN(i × ∆f) shall be measured over a 1 second time interval in diagnostics mode. In initialization, the ATU shall do a best effort attempt to optimize QLN(f) measurement time, however measuring over at least 256 symbols, with indication of the measurement period to the far-end Management Entity (in symbols, represented as 16-bit unsigned value, see 9.4.1.10).

The quiet line PSD QLN(i × ∆f) shall be represented as an 8-bit unsigned integer n(i), where i is the subcarrier index i = 0 to NSC – 1.

The value of QLN(i × ∆f) shall be defined as QLN(i × ∆f) = –23 – (n(i)/2) dBm/Hz. This data format supports a QLN(f) granularity of 0.5 dB and an QLN(f) dynamic range of –150 to –23 dBm/Hz.

An QLN(i × ∆f) value indicated as n(i) = 255 is a special value. It indicates that no measurement could be done for this subcarrier because it is out of the PSD mask passband (as relevant to the chosen application option – see annexes) or that the noise PSD is out of range to be represented.

Signal-to-Noise Ratio per subcarrier (SNR-ps)

The signal-to-noise ratio SNR(f) for a particular subcarrier is a real value which shall represent the ratio between the received signal power and the received noise power for that subcarrier. The signal-to-noise ratio SNR(f) per subchannel shall be measured by the receive PMD function in diagnostics mode and initialization. The measurement may be updated autonomously and shall be updated on request during showtime. The SNR(f) shall be sent to the far-end transmit PMD function during diagnostics mode (see 8.15.1) and shall be sent on request to the near-end Management Entity. The near-end Management Entity shall send the SNR(f) to the far-end Management Entity on request during showtime (see 9.4.1.10).

The receive PMD function shall measure the signal-to-noise ratio SNR(f) with the transmit PMD function in a MEDLEY or showtime state. The signal-to-noise ratio SNR(f) shall be measured over a 1 second time interval in diagnostics mode. In initialization and showtime, the ATU shall do a best effort attempt to optimize SNR(f) measurement time, however measuring over at least 256 symbols, with indication of the measurement period to the far-end Management Entity (in symbols, represented as 16-bit unsigned value, see 9.4.1.10).

The signal-to-noise ratio SNR(i × ∆f) shall be represented as an 8-bit unsigned integer snr(i), where i is the subcarrier index i = 0 to NSC – 1. The value of SNR(i × ∆f) shall be defined as SNR(i × ∆f) = –32 + (snr(i)/2) dB. This data format supports an SNR(i × ∆f) granularity of 0.5 dB and an SNR(i × ∆f) dynamic range of –32 to 95 dB.

An SNR(i × ∆f) value indicated as snr(i) = 255 is a special value. It indicates that no measurement could be done for this subcarrier because it is out of the PSD mask passband (as relevant to the chosen application option – see Annexes) or that the signal-to-noise ratio is out of range to be represented.

Loop Attenuation (LATN)

The loop attenuation (LATN) is the difference in dB between the power received at the near-end and that transmitted from the far-end over all subcarriers, i.e., the channel characteristics function H(f) (as defined in 8.12.3.1) averaged over all subcarriers. LATN shall be defined as:

with NSC the number of subcarriers (see 8.5) and H(f) represented by Hlin(f) in diagnostics mode and Hlog(f) in initialization (with conversion of log to linear values for use in the above equation).

If one or more H(f) values could not be measured because they are out of the PSD mask passband (as relevant to the chosen application option – see annexes) (see 8.12.3.1), then the LATN shall be calculated as an average of H(f) values over a number of subcarriers that is less than NSC. The loop attenuation shall be calculated by the receive PMD function during diagnostics mode and initialization. The calculation shall not be updated during showtime. The loop attenuation shall be sent to the far-end transmit PMD function during initialization and diagnostics mode (see 8.15.1) and shall be sent on request to the near-end Management Entity. The near-end Management Entity shall send the LATN to the far-end Management Entity on request during showtime (see 9.4.1.10).

The loop attenuation LATN shall be represented as an 10-bit unsigned integer latn, with the value of LATN defined as LATN = latn/10 dB. This data format supports an LATN granularity of 0.1 dB and an LATN dynamic range of 0 to 102.2 dB.

An LATN value indicated as latn = 1023 is a special value. It indicates that the loop attenuation is out of range to be represented.

Signal Attenuation (SATN)

The signal attenuation SATN is defined as the difference in dB between the power received at the near-end and that transmitted from the far-end. Received signal power in dBm shall be defined as the received subcarrier power, summed over the subcarriers in the MEDLEYset. During initialization and diagnostics mode, the transmit PSD for subcarriers in the MEDLEYset is at the REFPSD level. Therefore, the received signal power shall be fine-tuned with the gi values for each subcarrier in the MEDLEYset to estimate the signal power that will be received during showtime. During diagnostics mode, the fine tuning shall be restricted to using gi values 0 (for subcarriers to which no bits can be allocated) and 1 (for subcarrier to which at least one bit can be allocated).

Transmitted signal power shall be defined as the nominal aggregate transmit power (NOMATP), lowered by the power cutback (PCB, see 8.5). During diagnostics mode, only gi values 0 (for subcarriers to which no bits can be allocated) and 1 (for subcarrier to which at least one bit can be allocated) shall be used.

The signal attenuation shall be measured by the receive PMD function during diagnostics mode and initialization (i.e., estimate the signal attenuation at the start of showtime with the negotiated control parameter settings). The measurement may be updated autonomously and shall be updated on request during showtime. The signal attenuation shall be sent to the far-end transmit PMD function during initialization and diagnostics mode (see 8.15.1) and shall be sent on request to the near-end Management Entity. The near-end Management Entity shall send the SATN to the far-end Management Entity on request during showtime (see 9.4.1.10).

The attenuation SATN shall be represented as a 10-bit unsigned integer satn, with the value of SATN defined as SATN = satn/10 dB. This data format supports an SATN granularity of 0.1 dB and an SATN dynamic range of 0 to 102.2 dB.

An SATN value indicated as satn = 1023 is a special value. It indicates that the signal attenuation is out of range to be represented.

Signal-to-Noise Ratio Margin (SNRM)

The signal-to-noise ratio margin is the maximum increase (in dB) of the received noise power, such that the ATU can still meet all the target BERs over all the frame bearers. The signal-to-noise ratio margin shall be measured by the receive PMD function during initialization and diagnostics mode. The measurement may be updated autonomously and shall be updated on request during showtime. The signal-to-noise ratio margin shall be sent to the far-end transmit PMD function during initialization and diagnostics mode (see 8.15.1) and shall be sent on request to the near-end Management Entity. The near-end Management Entity shall send the SNRM to the far-end Management Entity on request during showtime (see 9.4.1.10).

To determine the signal-to-noise ratio margin (SNRM), the receive PMD function must be able to first determine the bits and gains table. During diagnostics mode, the receive PMD function may measure the SNRM value, or alternatively, may use the special value to indicate that the SNRM value was not measured.

The signal-to-noise ratio margin shall be represented as an 10-bit 2′s complement signed integer snrm, with the value of SNRM defined as SNRM = snrm/10 dB. This data format supports an SNRM granularity of 0.1 dB and an SNRM dynamic range of –51.1 to +51.1 dB.

An SNRM value indicated as snrm = –512 is a special value. It indicates that the signal-to-noise ratio margin is out of range to be represented. During diagnostics mode, the special value may also be used to indicate that the SNRM value was not measured.

Attainable net data rate (ATTNDR)

The attainable net data rate is the maximum net data rate that the receive PMS-TC and PMD functions are designed to support, under the following conditions:

  • Single frame bearer and single latency operation;
  • Signal-to-Noise Ratio Margin (SNRM) to equal or be above the SNR Target Margin;
  • BER not to exceed the highest BER configured for one (or more) of the latency paths;
  • Latency not to exceed the highest latency configured for one (or more) of the latency paths;
  • Accounting for all coding gains available (e.g., trellis coding, RS FEC) within latency bound;
  • Accounting for the loop characteristics at the instant of measurement.

To accurately determine the attainable net data rate (ATTNDR), the receive PMD function must be able to first determine the bits and gains table. Therefore, during diagnostics mode, the ATTNDR value shall be defined as an estimate of the line rate (without coding), calculated as:

with SNR(i × ∆f) in dB as defined in 8.12.3.3, snrgap = 9.75 dB (see Note). The function [x] is equal to 0 for x < 0, is equal to BIMAX for x > BIMAX and rounding to the nearest integer for 0 ≤ x ≤ BIMAX. The values of BIMAX and TARSNRM are defined in Table 8-48.

NOTE – The snrgap value is defined for a 10–7 bit error ratio on 4-QAM, in accordance with [B11]. The attainable net data rate shall be calculated by the receive PMS-TC and PMD functions during diagnostics mode and initialization. The measurement may be updated autonomously and shall be updated on request during showtime. The attainable net data rate shall be sent to the far-end transmit PMD function during initialization and diagnostics mode (see 8.15.1) and shall be sent on request to the near-end Management Entity. The near-end Management Entity shall send the ATTNDR to the far-end Management Entity on request during showtime (see 9.4.1.10).

The attainable net data rate shall be represented as a 32-bit unsigned integer attndr, with the value of ATTNDR defined as ATTNDR = attndr bit/second. This data format supports an ATTNDR granularity of 1 bit/s. No special value is defined.

Actual Aggregate Transmit Power (ACTATP)

The actual aggregate transmit power (ACTATP) is the total amount of output power delivered by the transmit PMD function to the U reference point at tip-and-ring (in dB), at the instant of measurement. Therefore, it would be necessary for the transmit PMD function to take into account the transmit filter function. This task may prove to be a difficult task. Because the actual aggregate transmit power is expected not to significantly depend upon the transmit filter characteristics, the transmit PMD function shall take the nominal aggregate transmit power (NOMATP, see 8.5), lowered by the power cutback (PCB, see 8.5), as a best estimate of the near-end actual aggregate transmit power and do a best effort attempt to remove the impact of the near-end transmitter filter characteristics. The ACTATP should also include discretionary transmit power possibly applied during showtime to some subcarriers not in the MEDLEYset (see 8.10).

The receive PMD function is not aware of the far-end transmit filter characteristics, nor of the far-end discretionary power levels. Therefore, the receive PMD function shall take the nominal aggregate transmit power (NOMATP, see 8.5), lowered by the power cutback (PCB, see 8.5), as a best estimate of the far-end actual aggregate transmit power.

The near-end and far-end actual aggregate transmit power shall be calculated by the PMD function during initialization (i.e., the estimated aggregate transmit power at the start of showtime with the negotiated control parameter settings). The measurement may be updated autonomously and shall be updated on request during showtime. The near-end and far-end actual aggregate transmit power shall be sent on request to the near-end Management Entity. The near-end Management Entity shall send the near-end and far-end ACTATP to the far-end Management Entity on request during showtime (see 9.4.1.10).

To determine the near-end actual aggregate transmit power (ACTATP), the transmit PMD function must first receive the bits and gains table from the receive PMD function. Therefore, during initialization and diagnostics mode, only the far-end actual aggregate transmit power is exchanged.

The actual aggregate transmit power shall be represented as an 10-bit 2′s complement signed integer actatp, with the value of ACTATP defined as ACTATP = actatp/10 dBm. This data format supports an ACTATP granularity of 0.1 dB, with an ACTATP dynamic range of –31 to + 31 dBm. An ACTATP value indicated as actatp = –512 is a special value. It indicates that the actual aggregate transmit power is out of range to be represented.

[1] http://www.analytic.ru/articles/lib26.pdf (local copy 07/2002, local copy 04/2009)

[2] http://www.broadband-forum.org/technical/download/TR-138.pdf

About these ads

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s