FreeLists / si-list / [SI-LIST] Re: Dielectric loss modeling

["Sogo Hsu" <sogo.hsu@xxxxxxxxxxx>]

Hi, Vladimir,

I truly appreciate your fruitful comment. I think it is an old story 
in Si-List. What's the behavior and mechanism in time domain for a 
frequency dependent component? Telegraph's equation is a special 
case of Transmission line equation in sinusoidal steady state. 
Regarding a frequency dependent component, we even can not derive 
Telegraph's equation and characteristic impedance accordingly. That 
is said, time domain is real life, frequency domain is only a 
mathematical expression. 

Besides, I am very interested in the physical meaning of imaginary 
part of G(f). As I know, dielectric material ideally is free 
conductive. In other words, G(f) means the imaginary part of 
dielectric constant. the model of dielectric material. As a result, 
the imaginary part of G(f) should be the equivalent real part of C
(f).

Finally, I wonder the configuration in your case 1. Could you pls 
describe it in detail? Thank you in advance.

Best Regards,

Sogo Hsu, Ph. D.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Simulation center/Foxconn
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


        "Dmitriev-Zdorov, Vladimir" <vladimir_dmitriev-
zdorov@xxxxxxxxxxx>
        2004/12/02 01:06 AM
                 
                 =A6=AC=A5=F3=A4H=A1G <sogo.hsu@xxxxxxxxxxx>
                 =B0=C6=A5=BB=A7=DB=B0e=A1G 
                 =A5D=A6=AE=A1G [SI-LIST] Dielectric loss modeling


Hi Sogo,
 
Although it=A1=A6s not a direct answer to your question, I think this 
info would be useful.
Unfortunately, the very model for dielectric losses used in HSPICE
 
G(f) =3D G0 + freq * Gd
 
is not mathematically correct. It is non-causal since it only 
describes the real part of dielectric conductance and assumes 
imaginary part being zero. However, such imaginary part of G(f) 
should be nonzero, in fact, real and imaginary part must obey Kronig-
Kramer relation.
 
Non-causality is not an abstract requirement. It leads to some 
unpleasant effects normally seen as a discrepancy between AC and 
transient result. I heard reports on problems of the following type:
 
1. Take a line of considerable length, 20in or so, with considerable 
dielectric loss. Terminate the line at both ends. Apply voltage 
through termination resistance. In AC mode, find the line=A1=A6s 
attenuation as a ratio of the voltage at output and input end. Then, 
repeat experiment in transient mode. To do that, apply sine voltage 
and solve for considerable time so that magnitudes become constant. 
Measure the same ratio. The difference reaches 3 dB or more.
2. Again, take a long line with considerable dielectric losses, also 
terminated. Apply a step voltage and find the response. Repeat this 
experiment when dividing the line into N =3D 2, 4, 8, 16 identical 
shorter sections (each of length L/N) connected in line. Place all 
responses on the same plot. You will see the delay and even the form 
of the response depend on N.
 
Note that both discrepancies disappear if you remove dielectric loss.
 
I would be thankful if you know any other data relevant to the 
matter. 
 
Vladimir

 
 
 
 
 





------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field

or to administer your membership from a web page, go to:
http://www.freelists.org/webpage/si-list

For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field

List FAQ wiki page is located at:
                http://si-list.org/wiki/wiki.pl?Si-List_FAQ

List technical documents are available at:
                http://www.si-list.org

List archives are viewable at:     
                http://www.freelists.org/archives/si-list
or at our remote archives:
                http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
                http://www.qsl.net/wb6tpu