10/100BaseT Impedance Mismatch

G

Guest

Guest
Archived from groups: comp.dcom.lans.ethernet (More info?)

I need to route the 4 pairs of TP ethernet wires from a 10/100BASE-T NIC
card through 2 feet of board traces (approx 50 ohms of impedance) back
out to Cat5E (100 ohm impedance) cable. Or do I need anything at all?

Can I use matching magnetics on either side of the 50ohm traces to
eliminate this disconuity?

such as 1:.7 & 1:1.4?


Thanks in advance
 
G

Guest

Guest
Archived from groups: comp.dcom.lans.ethernet (More info?)

In article <VPFGe.8298$_%4.647@newssvr14.news.prodigy.com>,
JoeG <JoeG@none.com> wrote:

> I need to route the 4 pairs of TP ethernet wires from a 10/100BASE-T NIC
> card through 2 feet of board traces (approx 50 ohms of impedance) back
> out to Cat5E (100 ohm impedance) cable. Or do I need anything at all?
>
> Can I use matching magnetics on either side of the 50ohm traces to
> eliminate this disconuity?
>
> such as 1:.7 & 1:1.4?
>

Why not design your PC etch to be 100 ohms instead of 50? This is quite
simple to do; the impedance is a function of the trace width, the
spacing, and the dielectric constant of the board material. You can
adjust the trace spacing and/or width to achieve a 100 ohm
characteristic impedance. See any decent EE handbook for the equations
(or send me a private e-mail and I can dig them out for you).


--
Rich Seifert Networks and Communications Consulting
21885 Bear Creek Way
(408) 395-5700 Los Gatos, CA 95033
(408) 228-0803 FAX

Send replies to: usenet at richseifert dot com
 
G

Guest

Guest
Archived from groups: comp.dcom.lans.ethernet (More info?)

Rich Seifert wrote:
> In article <VPFGe.8298$_%4.647@newssvr14.news.prodigy.com>,
> JoeG <JoeG@none.com> wrote:
>
>
>>I need to route the 4 pairs of TP ethernet wires from a 10/100BASE-T NIC
>>card through 2 feet of board traces (approx 50 ohms of impedance) back
>>out to Cat5E (100 ohm impedance) cable. Or do I need anything at all?
>>
>>Can I use matching magnetics on either side of the 50ohm traces to
>>eliminate this disconuity?
>>
>>such as 1:.7 & 1:1.4?
>>
>
>
> Why not design your PC etch to be 100 ohms instead of 50? This is quite
> simple to do; the impedance is a function of the trace width, the
> spacing, and the dielectric constant of the board material. You can
> adjust the trace spacing and/or width to achieve a 100 ohm
> characteristic impedance. See any decent EE handbook for the equations
> (or send me a private e-mail and I can dig them out for you).
>
>
> --
> Rich Seifert Networks and Communications Consulting
> 21885 Bear Creek Way
> (408) 395-5700 Los Gatos, CA 95033
> (408) 228-0803 FAX
>
> Send replies to: usenet at richseifert dot com


I wish I could Rich but the 2 feet of ~50ohm impedance is a given -- I'm
familiar with SI and trace impedances but this 2 feet of ~50ohms cannot
be replaced. It's legacy design which would require $$ to replace.
 
G

Guest

Guest
Archived from groups: comp.dcom.lans.ethernet (More info?)

JoeG <JoeG@none.com> writes:

> Rich Seifert wrote:
> > In article <VPFGe.8298$_%4.647@newssvr14.news.prodigy.com>,
> > JoeG <JoeG@none.com> wrote:
> >
> >> I need to route the 4 pairs of TP ethernet wires from a
> >> 10/100BASE-T NIC card through 2 feet of board traces (approx 50
> >> ohms of impedance) back out to Cat5E (100 ohm impedance) cable. Or
> >> do I need anything at all?
> >>
> >> Can I use matching magnetics on either side of the 50ohm traces to
> >> eliminate this disconuity?
> >>
> >>such as 1:.7 & 1:1.4?
> >>
> > Why not design your PC etch to be 100 ohms instead of 50? This is
> > quite simple to do; the impedance is a function of the trace width,
> > the spacing, and the dielectric constant of the board material. You
> > can adjust the trace spacing and/or width to achieve a 100 ohm
> > characteristic impedance. See any decent EE handbook for the
> > equations (or send me a private e-mail and I can dig them out for
> > you).
> > --
> > Rich Seifert Networks and Communications Consulting
> > 21885 Bear Creek Way
> > (408) 395-5700 Los Gatos, CA 95033
> > (408) 228-0803 FAX
> > Send replies to: usenet at richseifert dot com
>
>
> I wish I could Rich but the 2 feet of ~50ohm impedance is a given --
> I'm familiar with SI and trace impedances but this 2 feet of ~50ohms
> cannot be replaced. It's legacy design which would require $$ to
> replace.


If you have traces that have 50 ohm impedance in relation to circuit
board ground, you can use two of them. Connect each of the two
wires from the incoming Ethernet twisted pair wire to one of such
50 ohms trace. This arrangement would give you 100 ohms differential
impedance for the signal traveling on the wire pair, the same as the
UTP cable has.

There are applications where 100 ohms differential signal is
trasnported over pair of 50 ohm coaxial cables instead of
100 ohms twisted pair. It works.



--
Tomi Engdahl (http://www.iki.fi/then/)
Take a look at my electronics web links and documents at
http://www.epanorama.net/