stone@quest.UUCP (Scott Stone) (06/22/89)
I would like some informed opinions on shielded vs. unshielded cable.
We currently use unshielded twisted pair for most of our data cable
installations, and are considering using shielded twisted pair instead.
The cost of shielded is not considerably more than unshielded, seems to
be a safer option, and would give an upward growth path in the event we
want to run twisted pair ethernet, rather than (or along with) the RS-232
serial we use now.
Have you experienced noise problems on RS-232 using shielded?
Do the current twisted pair ethernet solutions work just as reliably
(or more reliably) on shielded twisted pair?
When you ground one end of the shielded cable, where do you usually
find a convenient ground?
Thanks in advance,
--
*
* Scott T. Stone {rosevax,bungia}!quest!stone
* (612) 894-7752kwe@bu-cs.BU.EDU (kwe@bu-it.bu.edu (Kent W. England)) (06/26/89)
In article <4015@quest.UUCP> stone@quest.UUCP (Scott Stone) writes: > >The cost of shielded is not considerably more than unshielded, seems to >be a safer option, and would give an upward growth path in the event we >want to run twisted pair ethernet, rather than (or along with) the RS-232 >serial we use now. > The cost of shielded *cable* may not be "considerably" more than the cost of unshielded, but the cost of installation is substantially more. You can run Ethernet on unshielded twisted pair (UTP). You can even run 16M token ring on unshielded. Ungermann-Bass offers hardware to do this (at least they have offered a press release :-). So, what are you needing the shielded cable for? Unshielded you can use for everything you have today. As for tomorrow, start thinking about fiber, not shielded twisted pair. I will listen to arguments for thin Ethernet coax vis a vis UTP Ethernet... I think UTP is the way to go, however.
ncpjmw@amdcad.AMD.COM (Mike Wincn) (06/26/89)
In article <4015@quest.UUCP> stone@quest.UUCP (Scott Stone) writes: > > I would like some informed opinions on shielded vs. unshielded cable. [...] > Do the current twisted pair ethernet solutions work just as reliably > (or more reliably) on shielded twisted pair? The current draft standard is intended to describe a system operating over at least 100m of 24 GA unshielded twisted pair cable, as a minimum requirement. If there is no further change that arises out of ballot comments, and once 10BASE-T becomes an approved standard, a "non-standard" (for lack of a better term) cable may be used provided it meets minimum attenuation, impedance, delay, and jitter characteristics. > > When you ground one end of the shielded cable, where do you usually > find a convenient ground? > Wherever ground exists in the MAU or DTE. >* Scott T. Stone {rosevax,bungia}!quest!stone Mike Wincn ncpjmw@amdcad.AMD.COM (408) 749-3156 The opinions expressed are my own, not necessarily those of my employer, nor those of 802.3 Committee or 10BASE-T Task Force.
pat@hprnd.HP.COM (Pat Thaler) (06/28/89)
There are a number of types of twisted pair cable. In unshielded cables the primary variable is wire size. The 10BASE-T draft is designed to give performance comparable to coax media (in terms of BER) when run over 100 m of 24 AWG UTP. For 26 AWG UTP the distance will be shorter. 10BASE-T defines the media in terms of its end-to-end characteristics. These characteristics are attenuation, impedance, jitter, delay, and crosstalk. More than one type of twisted-pair media can meet the requirements for these characteristics for some distance. There should not be significant performance variation for different types when the requirements are met. If when you say shielded twisted-pair you mean IBM Type 1 and 2, there is one problem. It has a characteristic impedance of 150 ohms. UTP is noninally 100 ohms at 10MHz. Some manufacturers recommend using a balun to provide the correct impedance match when using this cable. The IBM STP has lower attenuation than normal 24 AWG UTP and can achieve longer distances than 100m once the proper match is provided. The superior characteristics and greater bulk of the IBM cables are not entirely due to shielding. The wire gauge is larger and the dielectric is thicker. There is also twisted pair which has an overall shield but no individual shields between pairs. This provides protection against external noise sources. Shielded 100 ohm cables are available. Some have lower attenuation than 24 AWG UTP and some don't.
barr@frog.UUCP (Chris Barr) (06/29/89)
In article <4015@quest.UUCP> stone@quest.UUCP (Scott Stone) writes: > > I would like some informed opinions on shielded vs. unshielded cable. 1. Ethernet over Shielded twisted pair cable is rated for *shorter* length that UTP - I believe by about 10%. (Yes, STP max. is *shorter* than UTP.) This is for Lattisnet, whose UTP max length is > 300 ft. 2. UTP is subject to interference. When we had troubles, the vendor would always inquire about flourescent lights' proximity. Our problems were due to loose connections - yet clearly noise can disrupt UTP. I also wonder whether UTP Ethernet has emitted noise that affected other signals, e.g. telephone, async, ..