[comp.dcom.modems] Big problems with modems that aren't grounded. What to do?

alexis@panix.uucp (Alexis Rosen) (12/04/90)

I posted a query here a month or so ago, complaining about some very strange
problems with a bunch of 2400bps modems connected to my unix box.

While I'm not sure of this, it now seems likely that the problems are the
result of poor grounding. I think this because the Telebit performs flawlessly,
and the US Robotics nearly so. They both have three-prong (grounded) plugs on
their transformers. The other three modems (a Smart One, a Prometheus, and a
Maxxum) lack grounding, and they frequently die (although the Smart One is
rather less susceptible).

The problem is that the modems and the computer frequently fail to recognize
that DCD and/or DTR control lines have been raised. Thus, the modems may not
answer, and if they do, the getty may not complete the open.

Is there anything I can do to deal with this situation? If not, I suppose I'll
need to get new modems. Which cheap modems have a ground connector? The ones
I know of, which are generally trouble free, don't have these.

Thanks,
---
Alexis Rosen
Owner/Sysadmin, PANIX Public Access Unix, NY
{cmcl2,apple}!panix!alexis

kjh@pollux.usc.edu (Kenneth J. Hendrickson) (12/04/90)

In article <1990Dec4.023427.25652@panix.uucp> alexis@panix.uucp (Alexis Rosen) writes:
>They both have three-prong (grounded) plugs on
>their transformers.
>Which cheap modems have a ground connector?

I think that this is not relevant.  I would be willing to wager that the
DC power from the plug-in power supply is floating, the phone lines are
somehow isolated, and the modem must be "grounded" to the computer via
the RS-232 cable.  I wouldn't wory about it.

-- 
favourite oxymorons:   student athlete, military justice, mercy killing
Ken Hendrickson N8DGN/6       kjh@usc.edu      ...!uunet!usc!pollux!kjh

kreed@telesys.cts.com (Kevin W. Reed) (12/04/90)

>>>alexis@panix.uucp (Alexis Rosen) writes:
>I posted a query here a month or so ago, complaining about some very strange
>problems with a bunch of 2400bps modems connected to my unix box.
[stuff deleted]
>The problem is that the modems and the computer frequently fail to recognize
>that DCD and/or DTR control lines have been raised. Thus, the modems may not
>answer, and if they do, the getty may not complete the open.

>Is there anything I can do to deal with this situation? If not, I suppose 
>I'll need to get new modems. Which cheap modems have a ground connector? 
>The ones I know of, which are generally trouble free, don't have these.

I've run into this problem several times.  I've seen it caused by two
problems:  1) Not enough power to the modem.  If you have a bunch of
modems plugged into the same strip (outlet) and that circuit is being
overdrawn, the ones with the cheap power supplied (or AC/DC converters)
may not have enough power to operate properly (This sometimes happens when 
you use an adapter from another modem type).
2) Sometimes plugging in a phone into a modem causes it to not work properly
for answering the phone.  The phone rings, but the modem doesn't see it.

I doubt it is caused by a ground problem although I'm not an electrician.

>Thanks,
>---
>Alexis Rosen
>Owner/Sysadmin, PANIX Public Access Unix, NY
>{cmcl2,apple}!panix!alexis

-- 
Kevin W. Reed --- TeleSys Development Systems -- PO 17821, San Diego, CA 92177
TeleSys-II BBS & telesys.UUCP +1 619-483-3890 - Telebit PEP Line +619 483 0965
UUCP: {nosc,ucsd}!crash!telesys!kreed -------- Internet: kreed@telesys.cts.com

kaufman@Neon.Stanford.EDU (Marc T. Kaufman) (12/05/90)

In article <1990Dec4.023427.25652@panix.uucp> alexis@panix.uucp (Alexis Rosen) writes:
>I posted a query here a month or so ago, complaining about some very strange
>problems with a bunch of 2400bps modems connected to my unix box.

>While I'm not sure of this, it now seems likely that the problems are the
>result of poor grounding. I think this because the Telebit performs flawlessly,
>and the US Robotics nearly so. They both have three-prong (grounded) plugs on
>their transformers. The other three modems (a Smart One, a Prometheus, and a
>Maxxum) lack grounding, and they frequently die (although the Smart One is
>rather less susceptible).

>The problem is that the modems and the computer frequently fail to recognize
>that DCD and/or DTR control lines have been raised. Thus, the modems may not
>answer, and if they do, the getty may not complete the open.

RS-232 pin 7 is the logic ground reference point for signals (e.g. DCD, DTR).
In my experience, it is preferrable to ISOLATE the modem from system ground
to avoid ground-loop voltage offsets.  In any event, the green wire ground is
not intended to carry current -- it is a protective circuit in case of failures
in the modem power supply.  It should be connected to the case and/or chassis,
but NOT to logic ground in the modem.

A more probable scenario is incorrect wiring of some of your AC outlets --
reversal of the Hot and Neutral leads.  I have seen equipment that couples
logic ground to the presumed Neutral lead (usually through .1 uf and 10-12K
in parallel).  If your plug is on another AC phase and wired incorrectly,
you will see 60 Hz AC superimposed on your modem signals.

Marc Kaufman (kaufman@Neon.stanford.edu)