Click to See Complete Forum and Search --> : Coaxial Bandwith Leakage True...or....false
paulie walnuts
November 16th, 2002, 01:37 AM
When joining two pieces of coax do you lose bandwith at the female connector? What Im asking is it like ham radio am I losing signal by lack of insulation, sh$# i dont know I might be getting to diffrent fields mixed up.
slarty
November 16th, 2002, 01:55 AM
I don't know the electrical details, but my gut feeling is that bandwidth on coax 10 megabit network is generally either 100% or nothing.
Sure if you have a very broken network you can lose packets, but in my experience even very dodgy networks have very low packet losses. 10meg coax generally either works perfectly or not at all.
I have read that there is a limit on the number of connectors and length of wire, presumably beyond that it stops working. However, again from experience, once you reach more than three or four machines, the number of connectors gets so high that reliability from dodgy connections falls to almost nothing so you can forget about any theoretical limits.
My university did used to have coax runs with 30 machines which worked, that was amazing. I only assume that they used high quality connectors etc and made their cables properly (unlike my home networks)
magnoon
November 16th, 2002, 02:49 AM
There is some loss which is why the standard is 185m total segment length with maximum of 30 nodes. When joining segments together a repeater is required and counts as a node on each segment.
Und3ertak3r
November 16th, 2002, 01:00 PM
Very similar to radio (RF), You lose signal level over the lenght due ti resistance in the cable as well as the hi frequency attenuation due to capacitance.. you will also lose signal due to insufficent shielding..
The Pulses/RF "Flow" as a EM field between the Core and Shield. If the shield is not 100% signal will be re-radiated...
Also even in the best of Cards the source and Load impeadences will not match correctly, this will cause pulses to be reflected along the coax, this will also cause signal degredation..
BUT to answer your Question..
Any connector will introduce a extra level of resistance and Impedence mismatch to the cable circuit there for causing some form of signal degredation either by resistive attenuation or by pulse reflections distorting the resultant signals..
(bit tired.. I hope that some of this made sence)
Cheers
gghornet
November 16th, 2002, 07:01 PM
I know BNC barrel connectors cause signal attenuation(weakening of the signal),which I imagine could affect the transmission speed and quality due to the recieving server having a hard time reading the signal,and possible need for retransmission of the signal.It's always best,whenever possible,to go with one solid piece of coax.If this is not avoidable,and you have to use a long cable with a number of connectors,I'd recommend a repeater.
Also,if such an arrangement is necessary,it sounds like your network may be a little too big for the particular physical topology you're using.For example if you are using a bus topology,you might want to consider a star or star-bus,that way you can make more effecient use of the shorter pieces of cable.Or you might want to implement a thicknet to thinnet structureto reduce signal attenuation and increase range.
There's a million different ways to solve such networking problems.If I were you,I'd get a Networking essentials book or something along those lines and learn the stuff yourself,because along with the different physical and logical topologies,cabeling types,and different connectors,repeaters,hubs,and routers,you have a budget and that is going to be the main obstacle in most networks.
ammo
November 16th, 2002, 08:10 PM
Hum, gg: coax and BNC connectors == thinnet...
(And thinnet is much deprecated...)
Ammo
gghornet
November 16th, 2002, 09:42 PM
I realize that BNC and coax = thinnet,and that thinnet is deprecated,but that's what it sounds like he's referring to,hence that post,but I would definately recommend going with catagory 5 STP or ideally fiber optic cable as signal integrity is much better than standard thinnet.The fiber optic route is still kinda pricey though.
ammo
November 16th, 2002, 11:22 PM
Yeah ok..
By the way: "oups!" I meant "thicknet is much deprecated", thinnet is only "deprecated" ;)
Indeed as far as interference fiber is the obvious champion... Also, with STP you have to make sure it's properly grounded else it can be worst that UTP, because the shield acts like an atennea...
Ammo
Caine
November 17th, 2002, 02:51 AM
Well, you are all right.
There is several different things to factor for coax.
First, The size (RG-59, RG-6, RG-11) which is for residential houses has a certain bandwidth
for each.
RG-59 = 5 megabit <<Which is not preferred for broadband>>
RG-6 = True 10 megabit <<Which IS preferred for broadband by majority of ISP's>>
RG-11 = 10 to 20 megabit <<Which is very high quality coax, usually used for
apartments, duplexes, and etc>>
Second is the Spectrum that the broandband is ran:
There is different ratios to accuire by:
What MHz your upstream is coming in on
What MHz you downstream is going out on
Most all ISP's usually use for their upstream somewhere between 80Mhz to 1Ghz or higher (usually ranging within 6Mhz worth of bandwidth) On the downstream side anywhere from 5Mhz to 200Mhz with a 6Mhz range.
<<At my company, we use 98Mhz for our downstream and 32 Mhz for our upstream
using a Cisco uBR7114>>
Third is the distance and how far it has to go before it hits a HFC plant amplifier until it's reaches the Headend. Here is a link to look at http://www.radio-ware.com/products/techinfo/coaxloss.htm
There is several more factors to look at when determining bandwidth with coax but
I would be here all day talking about it. Something to look at if you are having problems
with bandwidth with coax is to talk to your ISP and tell them to check your dB levels, check for C/N (Carrier Over Noise), Check for leaks (such as bad connectors or cracks in the coax), Or to see if there is too many splitters in the coax line.
Hope this helps
Caine