[comp.dcom.lans] ANSWERS: Voice and Ethernet over the same 4 pair wire

John_Robert_Breeden@cup.portal.com (07/24/89)

I posted this message about two weeks ago:

  Has anyone out their had any experience with 10mb twisted pair ethernet
  (ie;Synoptics, Cabletron, HP, UB, AT&T etc) co-existing with voice 
  (ie: PBX and/or Centrex) in the same wire bundle (ie; voice/data in the same
  4 pair PDS wiring or 25 pair).

  I'm most interested in experience with Synoptics, I've been led to believe
  that, because their signalling is only 1v pp that the Ring Indicator on the
  voice circuit would drive it crazy.

Some of you asked if I'd post the answers, so here they are.

The vendors that I've gotten answers to this question (both on and off
the net) are: AT&T, UB, Synoptics, Cabletron and 3Com and the answer with
all the vendors is, in fact, running voice and data through the same 
wire bundle (4pair or 25 pair) seems to work! (if you try it and it 
dosn't....well. don't blame me).

Another interesting point. The only vendors equipment that can interoperate
on a twisted pair level (ie. plug an RJ45 into one vendors hub and plug the
other end into another vendor's hub) are those vendors that conform to the
IEEE 10baseT draft TODAY (<- key word). Those vendors are AT&T, UB, HP and
Networth - all others are propriatary.

Sorry for the editorial comments but I'm pro-10baseT.
--------------------------- next message --------------------------- 
From: smb@ulysses.att.com

Let me strongly suggest that if you're purchasing new stuff, that
you get one of the products compatible with the proposed 10BaseT
standard.  These include AT&T (my company; obvious disclaimers apply),
HP, and some others.

[ AT&T tested Starlan10 (10baseT) and voice together at Lockheed and it
worked - John B.]
--------------------------- next message --------------------------- 
PG&E in San Francisco has an ethernet with both Synoptics (1000 series) and
Ungerman Bass (10baseT) with over 3000 nodes. They have run both voice 
and data thru the same 4 pair (well, with PG&E 5 pair - so much for PDS) 
with both Synoptics and UB (some runs in excess of 300 feet between hub 
and end point) with no problems.

--------------------------- next message --------------------------- 
From: johne@pwcs.StPaul.GOV (John A Erickson)

We've got a TIE systems ONYX (digital) phone system installed in the same 
cable bundle as a Synoptics, with no problems. We do, however, have what the
voice people called "4 X 4" cable: two sets of four pairs, one set of four
inside a shield for isolation. We're running the voice pairs inside the shield,
and, like I said, no problems.
-- 
John A. Erickson,City of St. Paul (johne@pwcs.StPaul.GOV)
                                  (...!{amdahl|hpda}!bungia!pwcs!johne)
Wiker's law:
    Government expands to absorb all available revenue, and then some.

--------------------------- next message --------------------------- 
From: apple!ames!frog.UUCP!barr (Chris Barr)

Synoptics died only when we actually plugged a telephone
into a concentrator card.  This killed heartbeat.  We spent some time
(this was at an enduser MIS shop) getting the connectors crimped right,
and trying to keep phone rj45 males out of our Ethernet rj45 cubicle sockets.

Our bundles came out of the  ceiling to 4 punch-down blocks, 2 for phone,
2 for data which we converted from async terminal to Synoptics Ethernet.
So the phone lines were pretty close, in fact bundled at the cubicle
end through umbilical cords to the ceiling (1 per 8 wkstations) and 
run parallel for ~50-100 feet.

Could you post a brief summary of responses?

 - Chris Barr

--------------------------- next message --------------------------- 
From: Reinier_Tuinzing@dsd.3Mail.3Com.COM

In your request for information regarding running Ethernet in the same bundle 
with voice data.

3Com has a solution, MultiConnect PairTamer, which is compatible with the 
10Base2 standard, that allows users to run Ethernet on 1 pair of wire in a 
bundle.  Tests with the phone ringing show no performance degradation because 
of the voaltage spikes.  On 24AWG unshielded twisted pair wire the maximum 
segment distance allowed is 500 ft.  The configuration is to an install a 
PairTamer at either end of a twisted pair run.  The client side connects to an 
Ethernet station while the other end usually connects to the wiring closet 
where a MultiConnect repeater resides.  The MultiConnect repeater is an 
Ethernet repeater platform which allows users to run Ethernet on twisted pair, 
fiber optic, or coax cables.  A versatil solution from one platform.

If you would like more information I would be happy to send this to you if you 
would provide me with your address.

/regards
Reinier Tuinzing

[Note: The PairTamer is a balun coil and therefore NOT 10baseT standard. The
draft supports ONLY the AUI, which is an active, retiming, regenerating
repeater. Also, how can a twisted pair be 10base2, doesn't the 10base2
standard call for thin net coax? - John B.]
--------------------------- next message --------------------------- 
From: uunet!nuchat!steve (Steve Nuchia)

Works fine for me.  More precicely, for my clients who
have Synoptics sharing wire bundles with voice, token ring,
rs232, bisync, and a few other things.
-- 
Steve Nuchia	      South Coast Computing Services
uunet!nuchat!steve    POB 890952  Houston, Texas  77289
(713) 964 2462	      Consultation & Systems, Support for PD Software.
---
Steve Nuchia	      South Coast Computing Services
uunet!nuchat!steve    POB 890952  Houston, Texas  77289
(713) 964 2462	      Consultation & Systems, Support for PD Software.

--------------------------- next message --------------------------- 
From: rsm@math.arizona.edu (Robert Maier)

I haven't, but I'm afraid we folks at the University of Arizona will
have a great deal of experience in the near future.

This campus is installing a 5ESS and lots of accompanying wiring,
including many building-to-building fibre optic trunks.  The original
plan was for voice and data connectivity to each office on campus
(`data' meaning 9.6kb, switched).  Midway through the design stages
the plan was changed: our telecommunications people realized that
researchers now need 10 Mb/s Ethernet speeds, and that 9.6kb serial
lines are obsolete.  Unfortunately -- you guessed it! -- offices will
get access to the 10 Mb/s data speeds on the inter-building fibre only
through the medium of the new intra-building twisted pair.  I.e. the
wiring that was to have carried 9.6kb terminal-to-host traffic.  The
plan was too far along to be changed.

This lack of foresight is likely to lead to real trouble.  Like you
I'm very concerned about running 10 Mb/s twisted pair Ethernet through
the same wire bundle carrying voice.  Fortunately the technically
advanced departments here, including my own, have already installed a
decent coaxial Ethernet.

By all means pass on any information you collect; I'm quite
interested...

--
Robert S. Maier   | Internet: rsm@math.arizona.edu	[128.196.128.99]
Dept. of Math.    | UUCP: ..{allegra,cmcl2,hao!noao}!arizona!amethyst!rsm
Univ. of Arizona  | Bitnet: maier@arizrvax
Tucson, AZ  85721 | FAX: +1 602 621 8322
U.S.A.            | Voice: +1 602 621 6893  /  +1 602 621 2617

--------------------------- next message --------------------------- 
From: mark@central.cis.upenn.edu (Mark Foster)

Yup for Cabletron.  When I was at Oregon Graduate Center, in
Beaverton, Oregon, I installed Cabletron's MMAC twisted pair
system.  We tested it pretty thoroughly the first few weeks,
esp. for immunity to ringing voltages (since we had many
lines where data pair and phone pair were inside the same
jacket).  We were also worried about crosstalk of adjacent
TP ether connections.  We had absolutely no problems.
I highly recommend Cabletron.

By the way, I think Cabletron's solution is much better and
more flexible than Synoptic's.  When I was initially deciding
on whether to get Cabletron or Synoptics gear, I got a much
better feeling from talking to Cabletron customers and 
Cabletron tech support (before, during, after sale).
I'm also confident that because of their design strategy,
Cabletron is much more likely to have 10BASET-conformant
gear, once the standard is finalized.  It looks like Synoptics
may have to do some serious redesign.

----
Mark Foster
CIS Research Computing
Univ. of Pennsylvania
Philadelphia, PA
   mark@central.cis.upenn.edu

[ Yea, but today you can choose from vendors that ALREADY conform to the
10baseT draft - then you don't have to worry who "is much more likely" to
conform. Besides, you can mix-and-match vendors - John B.]