In a message dated 5/15/2002, Colin Botha writes:

> Are there any souls out there who can shed some light on possible test
> suites, test equipment, etc. please?
> 


Hi Colin:

It is certainly true that V.32bis modem performance is affected by many other 
impairments besides attenuation.  In fact, attenuation alone would be a very 
poor way to evaluate modem performance.

While the two most widely recognized transmission impairments that can affect 
modems are attenuation and noise, experience has shown that V.32bis modems 
are susceptible to additional impairments such as phase jitter, distortion, 
envelope delay, ADPCM compression, satellite delay, and echo.

To address these impairments in a comprehensive and logical fashion, the TIA 
has published TIA/EIA TSB37A, which describes a comprehensive, objective 
model of the phone network.  All of the relevant impairments are included, 
but a variety of test combinations have been used to make each impairment 
appear in approximately the same percentage of test lines that it appears in 
the actual network.  

Using this network model, it is possible for a modem vendor to objectively 
predict how well their modem will perform over the full range of possible 
connections on the phone network.  On any simulated connections where the 
modem performs poorly, the designer can study the problem in a laboratory 
environment and evaluate the effects of various changes in the design.  

Test equipment for performing tests per TSB 37A and TSB 38 is available from 
Spirent communications (http://tas.spirentcom.com), which recently acquired 
the telecom test manufacturer TAS.  Unfortunately, a system that performs the 
full TSB 38 tests can get expensive.

The TAS 100 (http://www.taskit.com/prod_100.htm) is probably a good candidate 
for basic modem testing, although it can not fully simulate all of the test 
conditions in TSB 37A.

It is also possible to have modem testing performed at an outside lab such as 
Henderson Communications (http://www.henderson-labs.com).   This may seem 
attractive since it involves no capital expenditures for test equipment, but 
the testing fees can quickly add up to more than the cost of some basic test 
equipment.  Also, when using an outside lab there is little opportunity for 
the type of test-modify-retest cycle that designers can use in the comfort of 
their own lab.  In view of this, you might consider purchasing some basic 
test equipment for internal pre-testing, to be followed by more comprehensive 
tests at an outside lab.


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

Reply via email to