larry@uunet.uu.net (Larry Lippman) (08/13/89)
In article <telecom-v09i0288m01@vector.dallas.tx.us> Torsten Dahlkvist <euatdt@euas11g.ericsson.se> writes: > Oh, by the way, the "dB" levels are of course weighted against the old > traditional "1 mW into 600 Ohms" standard, giving a reference voltage > of 0.7746 V. One should exercise caution when discussing and comparing "dB" levels as applied to telecommunication circuits. "dB" may be used to denote comparisons between POWER or VOLTAGE, but the "dB" is not an absolute unit. dB is often referred to as a _relative_ power level with respect to a reference TLP (Transmission Level Point), with say, 0 TLP being 1 milliwatt of signal. What you are really referring to is the "dBm", which is an absolute unit of POWER based upon a power input of 1 milliwatt into an impedance of 600 ohms. It is important to understand that, in general, telecommunication transmission is concerned with POWER levels and not voltage levels, and that to think in terms of "volts" rather than watts (i.e., dBm) is misleading with respect to many of the issues involved. Telecommunication engineering is largely concerned with speaking and listening LEVEL, which is a function of POWER. As a trite, but important example, one rates stereo amplifier output level in watts, not volts! > Does anybody know why the 600 Ohm standard line impedance has been replaced > by "900 Ohm // 30 nF" for modern phones? 900 ohms is probably a more realistic compromise impedance for a subscriber loop than 600 ohms. The impedance of a subscriber telephone loop terminated with a telephone set is a highly complex function of frequency, cable design characteristics and loop length. A typical non-loaded loop of 10 kft in length will have an impedance whose resistive component varies between say, 300 and 1,200 ohms, and whose corresponding reactive component varies between j300 and -j600 ohms. All of these terms vary in a highly non-linear manner over a frequency of 300 to 3,000 Hz. A typical loaded loop of say, 25 kft, shows even more variation in impedance over 300 to 3,000 Hz, with the resistive component varying between 300 and 1800 ohms, and with the reactive component varying between -j300 and -j900 ohms. It is important to understand that the impedance of a subscriber loop is a _characteristic_ impedance, is largely beyond control, and varies greatly depending upon frequency. To be practicable, overall transmission characteristics are usually specified at a _single_ compromise impedance (i.e., the 900 ohms) and at a given frequency (often 1,000 Hz). One of the most critical areas involving loop impedance is when a 4-wire circuit is terminated in a CO to a 2-wire loop. Since a 4-wire/2-wire hybrid requires a balance network approximating the 2-wire impedance, a "compromise" network of a fixed 900 ohms (plus adjustable "build-out" capacitance) is usually used. If the transmission characterics are critical, an equalizer is used to flatten out the transmission loss versus frequency. <> Larry Lippman @ Recognition Research Corp. - Uniquex Corp. - Viatran Corp. <> UUCP {allegra|boulder|decvax|rutgers|watmath}!sunybcs!kitty!larry <> TEL 716/688-1231 | 716/773-1700 {hplabs|utzoo|uunet}!/ \uniquex!larry <> FAX 716/741-9635 | 716/773-2488 "Have you hugged your cat today?"