PCB trace with controlled impedance

BaSed on my empirical experience, provided the surrounding plane is more than 3 track widths clear, it makes a negligible difference to Zo.

Are you trying to make this yourself, or are you having the board made by a faBricator? if it is being made at a fab, tell them the impedance requirements and they will tell you the track widths, spacing and the stackup.

Cheers

PeteS

The buried signal line would be the "stripline" configuration and has a formula similar to the microstrip one. You may want to consider looking at

This structure is sometimes called a "Coplanar Waveguide with Ground". Eq 3.4.3.1 from Waddell says

Zo = 188.5ohm / [ sqrt(er,eff) ( K(k)/K(k') + K(k1)/K(k1') ) ]

where

k = a/b, where a is the trace width and b is the width of the groove (trace plus the two insulating gaps)

k1 = tanh( Pi a /(4 h)) / tanh( Pi b /(4 h)) where h is the substrate thickness

k' = sqrt(1 - k^2) k1' = sqrt(1 - k1^2)

The function K(x) is the elliptic integral.

Regards, Mikko

Search the Agilent website for AppCAD, this free tool does it for you.

Rene

Go to

Hmm... does the link get you TxLine 2003? It knows about coupled (differentla) microstrips and striplines.

For many applications the separation between your differential traces ends up being large enough such that the coupling coefficient between them ends up being pretty small (e.g., less than 5%), in which case you can treat each trace as a single-ended line and not worry too much about the true differential impedances (even and odd modes).

---Joel Kolstad

Thanks for the link, but this tool can only calculate single ended transmission line, how could I use it to calculate differential traces?

Joel,

The TxLine off the Taconic web site is Version 1.1 and has a 1996 copyright. Where can you download TxLine 2003?

Thanks.

Tx2003 seems gone from the AWR website. I could email it to you... it's a 500K zipfile. It's a nice complement to Appcad.

mail me:

jjlarkin/at/highlandtechnology/dot/com/

John