Premier IT Outsourcing and Support Services within the UK

User Tools

Site Tools


From Lynzie's Motherboard —

–Message #282– From: Jim 853 To: ALL Date: 08-31-88 13:22:17 Subject: LINE NOISE

Most line noise that has troubled me has been input through the telephone lines of Pac Tel. Here are two possible corrective measures. Go to Radio Shack with $6.95 plus tax. Get a snap on choke, aka toroid. Cat. No. 273-104. It is a ferrous core that snaps apart and snaps toegether again. Snap it open. Wind your telephone cord around it up to seven times. Place it close to the modem. The plastic blister pack comes with two cores. If you need more suppression, use both. I have found it quite effective - not perfect, but real good - on the Everex Evercom 920 1200 b internal modem that I am using right now. Before installing it, I could not have possibly used this modem during the daytime hours. It is VERY simple. Even you can do it. There is no wiring at all. Just wrap your phone cord - be careful about the direction of winding and snap it together again. The second approach is one which has benefitted a member of this board. This involves a friendly Pac Tel repairman who came up with a "trap." To me, a "trap" is a notch filter. I have looked at it. It is a small plastic item about the size of a book of matches. It has two wires in and two wires out. It is placed in the phone junction box (where the phone line enters the house). It has three two pole switches on it. By setting the switches, it is possible to screen out noise in a particular frequency range. The member who benefitted from this has the computer/modem about two blocks from a cluster of 5 tall AM radio transmitting towers. Before the installation, it was absolutely impossible to use the modem until late at night when the station powered down. Now there is no noise during the day when the station is powered up. This installation was a "favor." The member has checked (with the wife of the repairman) and been informed that the trap is nowhere available. By report, you can't even PAY Pac Tel for them. But they have them. So why don't some of you clever folks approach your friendly Pac Tel repairmen and find out how to get them? And when you do, please let me know. Meanwhile, the Radio Shack toroid is onehellofalot better than nothing. P.S. -Don't screw the goose who laid the golden egg. The member who got this really did get a favor. The phone co. does not know about it. So please don't say "So and so has one."

–Message #308– From: Bill 876 To: Jim 853 Date: 09-04-88 16:19:58 Subject: LINE NOISE –Why your inexpensive gagdet from Radio Shack helps

Jim, you told us a few messages back about a trick for reducing noisy communications problems with a Radio Shack gadget. For the benefit of those who may not have noticed or who may remain skeptical of the merits of such a simple thing (everything good has to be expensive, right?), the following is diagram of what you've done and an explanation of why it really does help.

  1. - - - - - -
    P1    + /\ /\ /\ /\      M1    |      |\
        ____\/_\/_\/_\/___________________| \       Input

Phone _ | / circuitry Line _ |/ of

        ____  _  _  _  __________________     |\    modem
            /\ /\ /\ /\                  \____| \
    P2    + \/ \/ \/ \/      M2               | /
                                   |          |/
        Bifilar transformer         -------------

The "+" signs are the polarity marks for the windings (the two wires of the phone line in this case); they identify which terminals of the transformer will, instantaneously, have the same polarity of induced voltage (EMF). The effect of the transformer is to reject common-mode signals while transmitting differential signals.

The desired information is in the differential signal, and is thus not attenuated. Noise sources such as leakage or inductive coupling from nearby power lines and voltages induced by lightning strikes and the like often have a large part of their energy in the common-mode signal, which is rejected. The net effect is an improvement of signal-to-noise ratio, which is what you reported to us.

A few words for those who may want to understand the language of the field a bit better for purposes of following up these ideas:

   The differential-mode signal is the difference in voltage

between two points. It is VP1-VP2 at the phone line and VM1-VM2 at the modem input (where Vxy denotes the voltage at point xy).

   The common-mode signal is the mean voltage at two points.  It is

(VP1+VP2)/2 at the phone line and (VM1+VM2)/2 at the modem input.

   The action of the bifilar transformer is to enforce that

(VP1-VM1), the voltage induced in the upper winding, is equal to (VP2-VM2), the voltage induced in the lower winding. The rejection of common-mode and the acceptance of differential-mode signals then follows as a direct consequence.

Modems normally attempt to do their *own* job of rejecting common mode signals. Often, however, this is done by circuitry which either only partially accomplishes this, or which accomplishes it up to a certain value of common-mode voltage, after which the circuitry no longer effectively does its rejection job. Thus the bifilar transformer gives a hand in both cases.

Thanks, Jim, for bringing this easily-implemented trick to our attention.


/data/webs/external/dokuwiki/data/pages/archive/bbs/noise.txt · Last modified: 2000/12/17 14:53 (external edit)