wells@disk.UUCP (Lee Wells) (09/20/90)
As part of our building renovation, we are getting ready to replace
all our phone wiring. We'll be running two cables from the wiring
closet on each floor to each station. Each cable will have 4 twisted
pairs. In most cases, one cable will be used for voice, and the other
cable will be used for 10BaseT Ethernet. Of course in special cases,
we may use both cables for 10BaseT, or both cables for voice.
We've gotten specs on three different types of wiring:
AT&T AT&T Brand
Type 2021 D plenum "X"
Mutual
capacitance
(pF/ft) 12 17 16.2
DC resitance
(ohms/1000ft) 25.7 27.3 25.7
Impedance (ohms)
@1KHZ N/A N/A 600
@1MHZ 122 100 110
Attenuation
(dB/1000ft)
@1KHZ 0.39 0.67 0.44
@.1MHZ 1.8 2.7 N/A
@1MHZ 5.1 8.2 6.4
All solid copper, 24 gauge.
The question, of course, is which wire is "best" for 10BaseT, or are
none of them appropriate? Alternatively, where can one get recommended
wiring specs for 10BaseT?
Please mail your response, will post a summary if appropriate.
--
Lee Wells (pick one address) wells@disk.UUCP
DISK Inc. [a Nixpub machine] att!chinet!mcdchg!ddsw1!corpane!disk!wells
502 968 5401 to 5406 [modems] uunet!ukma!corpane!disk!wells
Louisville, Kentucky U.S.A. arpa!e.ms.uky.edu!corpane!disk!wells brunetti@dumbo.nih.gov (Jim Brunetti) (11/28/90)
Can anyone tell me if the 10BaseT spec has any mention of having multiple
data runs in the same sheath. We have run 4 pair to locations running Starlan
and were wondering if we could use the 2 unused pair for another connection.
Should we use separate cable runs for each connection? Will there be
crosstalk? Also I need to know the model number of the cable that AT&T
recommends for Starlan. Thanks.
Jim Brunetti brunetti@dumbo.dcrt.nih.govcornutt@freedom.msfc.nasa.gov (David Cornutt) (11/28/90)
brunetti@dumbo.nih.gov (Jim Brunetti) writes: >Can anyone tell me if the 10BaseT spec has any mention of having multiple >data runs in the same sheath. We have run 4 pair to locations running Starlan >and were wondering if we could use the 2 unused pair for another connection. You can try, but I wouldn't do it. I've had bad luck with RS-232 runs in multi-pair cable, where the individual pairs weren't sheathed. I worked in one place that did this, and it was crosstalk city on runs of only about 200 feet. Even lowering the rates to 2400 baud didn't help on some lines -- and this is a lot less critical than 10BaseT. -- David Cornutt, New Technology Inc., Huntsville, AL (205) 461-6457 (cornutt@freedom.msfc.nasa.gov; some insane route applies) "The opinions expressed herein are not necessarily those of my employer, not necessarily mine, and probably not necessary."
ncpjmw@brahms.amd.com (Mike Wincn) (12/08/90)
In article <677@nih-csl.nih.gov> brunetti@dumbo.nih.gov (Jim Brunetti) writes: >Can anyone tell me if the 10BaseT spec has any mention of having multiple >data runs in the same sheath. We have run 4 pair to locations running Starlan >and were wondering if we could use the 2 unused pair for another connection. >Should we use separate cable runs for each connection? Will there be >crosstalk? Also I need to know the model number of the cable that AT&T >recommends for Starlan. Thanks. > Jim Brunetti brunetti@dumbo.dcrt.nih.gov It should be possible to use 10Base-T in bundles up to 25 pairs of 24 AWG wire, with any number of combinations of other services used on the other pairs without significant degradation of performance, either on the 10Base-T link or the other services. The standard does stipulate noise characteristics that a 10Base-T receiver is supposed to reject in favor of valid Manchester packets, and those characteristics are believed to be representative of a worst-case network noise environment, whether or not that noise arises from co-resident services, and this includes NEXT, FEXT, impulse noise, ringer hits, and other sources. There is no document that I know of which can perfectly anticipate all imaginable network operating conditions, or all realizable noise sources, and your particular network may or may not exceed conditions outlined in the standard. But insofar as various noise sources in your particular network do not exceed that described in the standard, you should have no trouble establishing a viable co-resident 10Base-T link. Having said that, I have not yet seen any new data on 10Base-T in 25-pair bundles, so I cannot yet comment on the success of the work done in drafting 10Base-T, and "existence proof" is the final arbiter. Mike Wincn ncpjmw@brahms.AMD.COM (408) 749-3156 DISCLAIMER: I speak for myself unless noted otherwise.