E1/T1 E3/T3 cable distances

Unanswered Question
Apr 5th, 2007

Hi,

can anyone tell me the distances obtainable over various cable types for the above network interfaces typicaly installed within 7200, 7300 and 10k routers.

Thanks

Dec

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 4.1 (5 ratings)
Loading.
decklin Thu, 04/05/2007 - 23:44

thannk you for the prompt reply,

Is this the sameaccross all of the interface cards?

Also is it possible to increase the reach by using higher specified cable?

Dec

Paolo Bevilacqua Fri, 04/06/2007 - 04:00

Hi,

the interfaces are built in strict conformance with the ITU standards so all behave similarly. I don't doubt that you can exceeded these values using low loss cable on the other hand these interface are inherently meant to be short haul so you need to exceed these distances you could use HDSL for T1/E1 and fiber for T3/E3. There are converters on the market.

scottmac Fri, 04/06/2007 - 05:15

This is not entirely accurate.

It's possible to go 650 feet(655 actually, I believe) from the NIU/SmartJack, but not over UTP (like Cat3,4,5,5e,6).

Spec for T1 signaling is "Premises Cable," which is individually shielded pair inside a shielded jacket. LIke this stuff:

http://onlinecatalog.anixter.com/SearchResultsServlet?QUERY=KEYWORD-SEARCH&VENDORPARTNUMBER=&APARTNUMBER=&CUSTOMERPARTNUMBER=&SEARCHTYPE=KEYWORD-SEARCH&selCatalogs=NONE&KEYWORD=T1+cable

UTP works OK for short runs, but may still introduce some errors.

The followup questions is usually "Why can't UTP work for 1.5 / 2.0 megabits when it works perfectly for 100 or 1000 megabit?"

The answer is the type of signaling. Ethernet uses differential signaling (Google for Manchester Encoding), T1/E1 is raw pulses, and the voltage levels are different than Ethernet.

The raw pulses are more likely to generate crosstalk / interference.

The difference in encoding schemes also means that Ethernet can discriminate / understand signals, even with some interference (because of the differential signaling).

If you put much noise or crosstalk on a T1/E1 cable, the receiver will not be able to figure out what's valid data and what's junk, and you'll get a loss of frame (LOF).

Our center has had several calls where equipment connected by perfectly in-spec Cat5e was marginal or failed to work .... when it was replaced with Prem cable (Anixter now lists it as "T1 Cable") everything came up and was error-free.

I believe the spec for DS3/coax is also 655 feet.

FWIW

Good Luck

Scott

decklin Sat, 04/07/2007 - 01:41

Thanks for the information so far.

As i thought its not a simple answer, for example i have seen quotes as follows

200 feet over cable type 734

450 feet over cable type 735

also quoted as 12db @ 17.184Mhz

Are the ITU specs freely available?

The customer i am currently looking at would like to deliver mulitple E1 circuits to our routers with a distance of upto 200m between the PTT NTE and our interface. To simplify the cabilng they propose to use a "z" screened multi pare which i asume to be a cable with a screen between TX and RX plus overall screeneing. I have asked for further details be for see lots of probblems.

There are also have some DS3's to be provided over similar distances and i dont think the multi core co-ax will cut the mustard. Hence the question regarding cable types. Would a higher spec cable give more head room ?

Thanks again for the help and any further suggestions you may have.

Dec

scottmac Sat, 04/07/2007 - 14:56

Keep in mind that the "specs" are more like guidelines, such that you can make installation decisions without needing a bunch of high-falootin' equipment to test for compliance.

IF your installer *has* the sophisticated equipment to actually test for the compliance (attenuation @ frequency, crosstalk, skew, etc.) and is willing to certify the cabling job (on paper, with a guarantee, warranty, and penalties for cable faults) then turn 'em loose.

The DS3s shouldn't be too much of an issue, they should be able to extend the drop fiber into the building as far as needed and put the mux in the data center or closet.

For that matter, the copper for the E1 should also be able to be extended into the building (before the demarc) to reduce the distance to the CPE and enhance the probability of keeping the connection to the CPE in-spec.

Either way you end up pulling cabling ... instead of pulling the cable from the demarc to the CPE, you'd be pulling cabling to extend the entry to the demarc.

"Screened" cabling implies a variation of UTP (they call it screened because it'd sound stupid to have "Shielded Unshielded Twisted Pair"...). It stays "screened" versus "Shielded Twisted Pair (STP) because it maintains most characteristics of UTP.

Also, as with any shielded cabling, if they don't terminate the shielding properly, the cable is worse then UTP; you might as well be running twisted barbed wire.

If they're going to pull cable, make it prem cabling and do it right.

My .02

Good Luck

Scott

Actions

This Discussion