In a message dated 10/17/2002, Alan Shaw writes:

> So, with regard to European CTR21, is the complex Reference Z equal to the
> telephone line impedance or does represent the complex conjugate so as to
> provide a power match when I design the TE's input impedance to provide a
> good return loss in the measuring bridge? 
> 


Hi Allen:

Yes, the complex reference Z in CTR 21 is intended to represent the impedance 
of a "typical" phone line as seen by the terminal device.  It is not the 
complex conjugate of the phone line impedance.  The goal in specifying return 
loss is to minimize reflections.  So, if you want to achieve good return loss 
with respect to the reference Z, your interface should present an impedance 
that is similar to the reference Z.

It sounds like you are trying to understand what it takes to meet the CTR 21 
return loss requirement.  You might be interested to know that even a pure 
600 ohm impedance will meet the requirement, although some shunt capacitance 
will improve the compliance margin at 4000 Hz.  Some purists may insist that 
better matching is needed, but I am not convinced.

The fact is that the actual impedance seen looking into a phone line varies 
over a tremendous range.  I have scatter plots of measurements on a large 
number of actual lines, plotted on the complex plane.  It looks like someone 
threw food at the wall.  When studying these plots it is possible to conclude 
that 600 ohms is sort of in the middle of the mess.  It is also possible to 
conclude that a complex network such as the one called out in TBR 21 
represents a closer approximation (which makes sense, since phone cables have 
some distributed capacitance).  

Even though the phone networks in various countries are fairly similar, 
different engineers have come up with different fixed-compromise 
representations for the reference impedance.  That's why some countries use a 
600 ohm reference and some have created their own complex reference.  So, 
that is all the TBR 21 network is: somebody's idea of a fixed compromise 
approximation of the "center" of the range of possible impedances experienced 
in the real world.

My point here is that for most applications, it is not productive to strive 
for extraordinary matching to whatever reference impedance is provided by a 
particular standard.  If the end product will be used on an arbitrary PSTN 
line, the actual matching achieved on a given line is likely to be rather 
poor.



Joe Randolph
Telecom Design Consultant
Randolph Telecom, Inc.
781-721-2848
http://www.randolph-telecom.com

Reply via email to