Re: [SI-LIST] : Transmission Line Conductors

alterra@adnc.com
Thu, 18 Sep 1997 21:07:04 -0700

In response to Hans' assertion that one can always ignore permeability at
high frequencies, I would like to mention that Alloy 42 and Kovar are
sometimes used as conductors in controlled impedance situations, usually for
thermal compatibility with a ceramic package or substrate. Both materials
are ferromagnetic. I have never seen any high frequency permeability data
for Kovar, but published measurments for Alloy 42 (in an IC package) show
significant relative permeability up to 500 MHz. To improve the high
frequency performance of these materials they are sometimes plated with a
better, non-magnetic conductor such as gold. However, this makes analysis
even more complicated since one has to deal with a layered geometry as well
as frequency dependent material properties. My experience so far is that it
is safe to assume a relative permeability of 1 for any metal operated above
1 GHz; below 1 GHz this is not a safe assumption, especially for
ferromagneitc metals.

Bottom line is, Alastair has a legitimate question concerning frequency
dependent material properties in transmission lines which is relevant to
practical situations.

I don't know of any analytical code that can handle this problem directly
and produce a time domain model although one can certainly solve the problem
at a number of discrete frequencies to create a freqeuncy domain model.
Some SPICE codes allow one to input tabular frequency domain data and then
do a time domain simulation.

I hope this is helpful. (...and if anyone has any kovar permeability data
vs frequency, please let me know)

---------------------------------------------------------------
Eric Wheatley Ph.D. (760) 942-9426 (phone)
Alterra Technology Co. (760) 942-2366 (fax)
Encinitas, CA 92024 alterra@adnc.com
---------------------------------------------------------------

At 04:07 PM 9/18/97 -0700, HANS_MELLBERG@NON-HP-Cupertino-om5.om.hp.com wrote:
> As far as the permeability is concerned, you can safely assume a value
> of 1.0 for the mu-sub-r. If you are using nickel, iron or cobalt as
> your conductor material or part of your alloy,then you need to worry
> about a non-unity value. The conductivity (real value) all it does is
> generate ohmic loss. The value you need to concern with is the
> frequency dependant skin effect wich essentially increases the ohmic
> resistance with frequency (R-sub-s).
> So, to answer you question, changing the material down the conductor
> path will essentially be insignificant except for ohmic losses. The
> impedance will stay the same (assuming the diameter is unchanged)
> Hans Mellberg
>
>
>______________________________ Reply Separator
_________________________________
>Subject: [SI-LIST] : Transmission Line Conductors
>Author: Non-HP-AHardie (AHardie@compuserve.com) at HP-ColSprings,mimegw5
>Date: 9/18/97 3:19 PM
>
>
>All the transmission line impedance formulae that I can find do not take
>into account the
>conductor's properties such as conductivity and frequency dependant
>permeability (mu).
>Does anyone know how a change in transmission line material part of the way
>along it's length
>affects the transmission line's performance?
>
>Alastair Hardie.
>
>
>