Alication Note 52 ADSL line qualification tests Auth: Peter Ziemann (e-mail: eter.ziemann@wago.de) Asymmetric Digital Subscriber Line (ADSL), a modem technology, converts existing twisted-air telehone lines into access aths f multimedia and high seed data communications. As the name imlies, ADSL transmits an asymmetric data stream, with much me data going downstream to the subscriber and much less coming back. A modem must therefe be installed at both ends of the line. The maximum downstream rate deends on the length of the line between the central office (CO) and the subscriber. Tyical values are: U to 5.0 miles (8 km) 1.544 Mbit/s (T1) 4.5 miles (7 km) 2.048 Mbit/s (E1) 3.0 miles (5 km) 6.312 Mbit/s (DS2) 2.5 miles (4 km) 8.448 Mbit/s Measurements on existing coer lines are end-to-end, with generat at one end and receiver at the other end of the twisted air. Crosstalk Crosstalk is unwanted transfer of energy from one wire to an adjacent wire caused by electromagnetic couling. Crosstalk attenuation is defined in terms of ower level differences as follows: Near-end crosstalk NEXT ADSL uses discrete multi-tone coding (DMT), dividing the transmission channel into 255 subchannels. Each subchannel is QAM modulated with a 4.3 khz bandwidth. This tye of transmission is very similar to frequency division multilex (FDM) systems. The lower frequency range (u to 20 khz) is used f POTS. The range from 20 khz to 1.1 MHz is used f dulex data transmission and is most imtant f line qualification tests. Verification of coer line arameters: Attenuation Attenuation is the main fact limiting cable length. It deends on the frequency of the transmitted signal. Nmally, attenuation increases with increasing frequencies and causes unwanted reduction in the signal strength. Far-end crosstalk FEXT As cables are never homogeneous, NEXT and FEXT should both be measured at both ends of the line. Longitudinal conversion loss Longitudinal conversion loss (LCL) signal balance ratio is defined in ITU-T Rec. O.9 as the logarithmic ratio of the longitudinal voltage (V L ) to the transverse voltage resulting from it. age 1
LCL is imtant in balanced systems when it comes to reducing the effects of commonmode voltages to ground, e.g. resulting from external interference fields. Imedance towards ground Transverse voltages (high imedance to ground) occur in balanced signal transmission. Damaged insulation, humidity effects and the caacitance between the twisted air and ground result in a variable imedance / frequency resonse towards ground, resulting in higher attenuation disttion and lower common-mode rejection. The measurement frequencies of interest are below 100 khz, to reduce the influence of shield caacitance. Descrition of test rocedures: The test setus shown in this alication note are f ADSL with a line imedance of 100 ohms. They are also suitable f other technologies, such as HDSL, ISDN, 2 Mbit/s PCM and leased lines with Z = 120, 135 150 ohms. In such cases, baluns are unnecessary and the balanced connects of the instruments can be used. Matching resists are always required to match the system imedance. The PSM-137-139 range of instruments is recommended (receiver + generat). Together with the LevelPRO control software, they fm a owerful test system f fast measurements and documentation. The searate SPM-33A receiver and PS-33A generat may also be used f tests at u to 2 MHz. Nmalization is required in der to achieve best accuracy. In the following, the term nmalization means: - Set the reference memy to the dislayed value (coy ABS to REF) - Dislay the nmalized value ABS-REF (sets the dislay to 0 db) Line attenuation test setu Characteristic line imedance The transmission characteristic of a coer line mainly deends on the characteristic imedance Z 0. This arameter is determined by measuring the imedance of the line with one end oen-circuit and then sht-circuit. Z 0 is easily determined by finding the frequency oints f which Z oen is equal to Z sht. There is no need to determine the hase angles with this method. The PSM-137 AUTOSTEP mode allows synchronized increments in the generat and receiver frequencies in u to 100 stes. The generat oerates as master, with the receiver as a time-controlled slave. The receiver is triggered by ressing the generat start key. The far end instrument can be connected via modem (V.24) f PC-controlled measurements with LevelPRO. age 2
Near-end crosstalk test setu Instrument dislay of attenuation vs. frequency NEXT is a single-end measurement. Users can select either AUTOSTEP the faster SWEEP mode f determining the frequency resonse. The result dislay and the comensation f insertion loss are the same as described f line attenuation. LevelPRO dislay of attenuation vs. frequency If a generat level of 0 dbm is used, the receiver dislays the attenuation with inverted sign. An easy way to comensate f the insertion loss of the baluns (< 0.5 db), is to increase the TX level. Longitudinal conversion loss test setu Far-end crosstalk test setu SWEEP mode rovides a dislay of LCL vs. the desired frequency range. Befe starting the measurement, the test setu must be nmalized to comensate f the intrinsic attenuation of the SDZ-30 bridge. The setu f nmalizing is similar to the test setu, with the line under test relaced by a reference element LCL = 0 ( a shtcircuit between oint a b and ground). The test rocedure is the same as f line attenuation. At low frequencies, crosstalk attenuation may be very high. Setting FSTART to f max and FSTOP to f min will ensure enough signal strength f the AUTOSTEP trigger. After nmalization (see above) at one frequency oint (in LEVEL mode), the test setu is ready f LCL measurements. The lower frequency limit f the SDZ-30 is 10 khz. The uer limit deends on the instrument used (PSM-139 u to 32 MHz). age 3
LevelPRO rovides convenient menu-driven nmalization f measurements using external bridges. The figure below shows an examle f the BMB-30 bridge. The instrument dislay is in db. LevelPRO shows the results in ohms. P -21 102-31 345-41 1112-51 3538-22 116-32 388-42 1249-52 3971-23 131-33 437-43 1403-53 4457-24 148-34 491-44 1575-54 5002-25 168-35 552-45 1768-55 5613-26 190-36 621-46 1985-56 6300-27 214-37 698-47 2229-57 7069-28 241-38 784-48 2502-58 7933-29 272-39 881-49 2808-59 8903-30 306-40 990-50 3152-60 9990 IMB-30 conversion table Imedance towards ground test setu Characteristic imedance test setu SWEEP mode yields a dislay of imedance towards ground vs. the desired frequency range. Befe starting the measurement, the test setu must be nmalized to comensate f the intrinsic attenuation of the IMB- 30 bridge. During nmalization, the line under test should be connected and the button on the IMB-30 should be ressed. After nmalization (see above) at one frequency oint (in LEVEL mode), the test setu is ready f imedance measurements. The lower frequency limit f the IMB-30 is 50 Hz, the uer limit deends on the measured imedance (max. 1 MHz at 1kΩ, max. 100 khz at 10 kω). LevelPRO makes it very easy to determine the frequency oints, where Z oen and Z sht are equal. The rocedure starts with menudriven nmalization of the test setu. The results of the first measurement (e.g. with oen end) are sted as TRACE B. The next ste is to measure with the end sht-circuit and to dislay the result of TRACE A&B. It is then a simle rocedure to use the marker to find the oints of intersection where the dislayed imedance cresonds to the characteristic line imedance. The next figure shows the result grah f a twisted air line with Z = 75 ohms. age 4
Determination of characteristic line imedance using LevelPRO: The black trace (high imedance f low frequencies) shows the oen-end values, the red trace shows the sht-circuit values. Another interesting alication using the same test setu is the estimation of the cable length. The length can be calculated if the λ/4 resonant frequency of the line has been measured and the NVP of the cable is known. The resonant frequency is the first minimum in the imedance trace of the oen-end air the first maximum of the sht-circuit air. The examle above shows a first resonance at 0.428 MHz. The resulting cable length is 87 m (if ε r =4 and NVP=0.5). Length = (NVP * c) / (4 * f res ) Length /m = (NVP * 300) / (4 * f res /MHz) Length /m = 300 / (4 * f res /MHz * SQR(ε r )) NVP = nominal velocity of roagation (0.5 to 0.9 c) ε r = dielectric constant of the cable insulation Test equiment PSM-137 Level test setu (8 MHz) BN 2203/15 with tracking generat (2 required) PSM-138 Level test setu (18 MHz) BN 2203/16 with tracking generat (2 required) PSM-139 Level test setu (32 MHz) BN 2203/17 with tracking generat (2 required) SPM-33A Selective level meter (2 MHz) BN 2033/01 PS-33A Level generat (2 MHz) BN 2071/01 LevelPRO BN 2203/93.01 (control and evaluation software) SMZ-100/75 Balun 1 khz to 3 MHz BN 2078/14 (2 required f 100 Ω systems, e.g. ADSL) SDZ-30 Signal balance ratio bridge BN 2234/01 Reference element LCL=0 f SDZ-30 BN 2234/01.01 IMB-30 Imedance bridge BN 2234/20 BMB-30 Imedance bridge BN 2234/30 Reference element 100 Ω f BMB-30 BN 2234/30.01 (may also be used as a matching resist) age 5