The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> Circuit Simulators >> Tline and DC in Spectre
https://designers-guide.org/forum/YaBB.pl?num=1231344965

Message started by kevinvinv on Jan 7^th, 2009, 8:16am

Title: Tline and DC in Spectre
Post by kevinvinv on Jan 7^th, 2009, 8:16am

Hey I've got a question about how the TLine model works in Spectre.

I specify Zo=50 and Td=10ns

I am checking things out however, at DC.

The symbol looks something like this

a-----xxxxxxxxxxxx-------b
c-----xxxxxxxxxxxx-------d

At DC, shouldn't a and b actually be connected to each other- and not to c/d?

and shouldn't c and d actually be connected to each other and not to a/b?

This is not how things are looking actually... and it is puzzling to me. I guess I have never looked at tline behavior before at DC but now I need too...

thoughts?

Title: Re: Tline and DC in Spectre
Post by pancho_hideboo on Jan 7^th, 2009, 5:00pm

Your post is usage of simulator not Design issue.
Please post to "The Designer's Guide Community Forum ≫ Simulators ≫ Circuit Simulators".

kevinvinv wrote on Jan 7^th, 2009, 8:16am:

This is not how things are looking actually... and it is puzzling to me.
I guess I have never looked at tline behavior before at DC but now I need too...

You are misunderstanding tline. tline is not differential line(coupled line).
See "spectre -h tline".

Quote:

Only the odd mode is modeled, so only the voltage difference across each port is important.
(The absolute voltage of each terminal is not significant.)
Also, the current into one node of a port exactly equals the current leaving the other node of the port.

Also see Chapter-7 of the following.
http://cp.literature.agilent.com/litweb/pdf/ads2005a/pdf/ccdist.pdf

"TLIN4 (Ideal 4-Terminal Transmission Line)" is most close to "tline of Cadence Spectre".
Here Pins 2 and 4 need to be appropriately grounded for correct usage of this component.
In your case, nodes "c" and "d" need to be grounded.

"tline of Cadence Spectre" is not coupled line such as "CLIN (Ideal Coupled Transmission Lines)".

Attached is a extraction from HSPICE manual.
In your case, "a"->"in", "b"->"out", "c"->"refin", "d"->"refout".

Title: Re: Tline and DC in Spectre
Post by Ken Kundert on Jan 7^th, 2009, 9:52pm

The tline model in Spectre and SPICE is perfectly happy having different DC voltages on the outer conductors at both ends of the lines.

Let
Vi1 = voltage on inner conductor of port 1
Vo1 = voltage on outer conductor of port 1
Vi2 = voltage on inner conductor of port 2
Vo2 = voltage on outer conductor of port 2
Then in DC the transmission line enforces that
Vi1 - Vo1 = Vi1 - Vo2
In other words it models the odd mode, which is RF jargon for the differential mode. It allows Vo1 != Vo2, which means it is not modeling the even mode (or the common mode).

To model both even and odd modes in a coax configuration, use the following subcircuit (note that this model is non-symmetric).

subckt coax (i1 o1 i2 o2)
parameters zinner=50 zouter=50 vinner=1 vouter=1 td=1
inner i1 o1 i2 o2 tline z0=zinner vel=vinner td=td
outer o1 0 o2 0 tline z0=zouter vel=vouter td=td
ends coax

To model both even and odd modes in a twisted-pair configuration, use the following subcircuit (note that this model is symmetric).

subckt twisted (p1 n1 p2 n2)
parameters zodd=50 zeven=50 veven=1 vodd=1 td=1
odd p1 n1 p2 n2 tline z0=zodd vel=vodd td=td
tf1a p1 0 e1 c1 transformer t1=2 t2=1
tf1b n1 0 c1 0 transformer t1=2 t2=1
even e1 0 e2 0 tline z0=zeven vel=veven td=td
tf2a p2 0 e2 c2 transformer t1=2 t2=1
tf2b n2 0 c2 0 transformer t1=2 t2=1
ends twisted

-Ken

Title: Re: Tline and DC in Spectre
Post by pancho_hideboo on Jan 7^th, 2009, 10:14pm

This is symmetrical coupled line using two tline.

Ken's definition of Zeven and Zodd is wrong.
If you refer data sheet of transmission line, Zeven and Zodd must be following.

Zeven=2*Zcommon
Zodd=0.5*Zdiff