Modem Follies -- A number of people wrote in about Mark Anbinder's article in TidBITS #176 concerning a strange line noise problem. It seems that this problem was big news in Australia some time back, as Ian MacColl <macoll@qut.edu.au> reported, and some of the theories there included some phones drawing too much power from the line, a capacitor charging to maintain stored numbers, or the affected phones reporting to their superiors at Telecom Australia Headquarters (a popular choice, since the problem was cyclical).
Ed Segall <edward.segall@n3.sp.cs.cmu.edu> proposed an alternate theory based on a problem he had and solved. Apparently, if the phone creates Radio Frequency Interference (RFI), the RFI can wreak havoc on modem connections. Ed said the simplest solution (short of buying a new phone) is a $20 AT&T Radio Frequency filter.
John Harkin <jh@nbn.com> had the best sounding theory, suggesting that the problem is caused by "the nonlinearities of the input impedances caused by cheap transformers." I don't know what it means, but I like the sound of it.