[RFI] Ethernet RFI
jvpoll at dallas.net
Sun Jan 23 15:32:25 EST 2005
Using fomulas from the theory of 'Permutations and
Combinations' the following equation determines the
number of unique groupings (irrespective of the order
or different permutations such as 1, 2 and 2, 1) or
n_C_r = n! / (n-r)! * r!
Therefore given 8 wires and choosing 2 we have:
8_C_2 = 8! / (8 - 2)! * 2!
8_C_2 = 8! / 6! * 2!
8_C_2 = 40320 / 720 * 2
8_C_2 = 28 unique combinations if my math is correct!
Jim / WB5WPA /
----- Original Message -----
From: "Ed -K0iL" <eedwards at tconl.com>
To: "'RFI Reflector'" <rfi at contesting.com>
Cc: "'John Pelham'" <john at radiophile.com>
Sent: Sunday, January 23, 2005 1:18 PM
Subject: RE: [RFI] Ethernet RFI
> Wouldn't there be 8 to the 8th-power possible combinations for 8 pin
> connections? (Someone may wish to check my math. There are 64
> combinations if you just keep the wires in the same order just shifting
> them over!) But only one of those would place the "twisted pairs" into
> correct EMC signal configuration. Otherwise you might be placing
> data signals together onto the same "twist" and splitting the signal pairs
> up onto different "twists". This would obviously work to some degree and
> for some length, but the EMC benefit of the twisted pair would then be
> lost. It only works correctly for EMC in one configuration out of
> 16,777,216 possible combinations. (if my math is correct).
> -----Original Message-----
> From: John Pelham
> Ford wrote:
> >When you hook up the cables to keep the colors all the same on each
> >pin at each end, you will have conductivity between the correct pins,
> >and the network will work, but you completely lose any shielding benefits
> >of the twisted pairs.
> I don't follow. I assume you mean by "colors all the same on each pin at
> each end," you mean that if at one end the green wire is on pin 6, at the
> other end the green wire is also on pin 6, for example. It seems like
> you're saying that this is the wrong way to do it. I've made zillions of
> Ethernet cables and have always done it this way. I just checked my
> reference and it says to do it this way. (I'm not talking about the
> category of crossover cables, just regular Ethernet wiring.) I don't see
> how doing it this way is wrong electrically. In fact it seems like doing
> any other way would be wrong and would likely not work at all.
> Am I misunderstanding? If not, how would you have us wire Ethernet
> John W1JA
> ----- Original Message -----
> From: "Ford Peterson"
> > I was surprised to find that grounding the EMT to the #4 Cu wire that
> > comes in from outside INcreases the level of these unwanted signals by 2
> > S-units. Any ideas why, and what I could do to get rid of this RFI
> > (until I get the tower up, at which time the feed point will be much
> > farther away from any of the computer equipment)?
> > Alan NV8A
> One of my pet peeves is when computer technicians wire cat5 cables
> incorrectly. There is a right way and a wrong way to do it. The right
> is the most difficult. Guess which way most people do it?
> The 8 conductors in that cable are actually 4 pairs of twisted cables.
> pair is a "hot" and a ground return. When you hook up the cables to keep
> the colors all the same on each pin at each end, you will have
> between the correct pins, and the network will work, but you completely
> any shielding benefits of the twisted pairs. You should not be hearing
> anything from that cable.
> I do not have the specification for those cable ends handy. But as I
> recall, the ground returns are the 4 pins on one end of the connector.
> Before rewiring your house with EMT, I would make sure the connectors are
> installed correctly.
> I have one run here that an 'installation expert' installed for me. I
> queried him explicitly on the correct wiring pattern. A lively discussion
> ensued. Realizing that I was arguing with an idiot with 7th grade shop
> class electrical wiring experience, and that I was paying him $2 a minute
> argue with me, I let him wire it his way. Not surprisingly, that
> 75' run of Cat5 never runs over 10mb in a 10/100 hub. The hub shifts into
> low gear to deal with the noise.
> I'm sure short runs work just fine wired however you want. Wired the
> way, you get 2 pairs carrying all the data and two pairs acting as the
> ground return. I'm sure there is some shielding afforded by lashing them
> into the same run of wire, but they are not acting correctly. It takes 4
> thumbs to wire them correctly. Find the spec and wire them right.
> RFI mailing list
> RFI at contesting.com
More information about the RFI