Re: [SI-LIST] : PCB impedance variation v.s. layout pattern

Chris Padilla ([email protected])
Fri, 27 Aug 1999 13:56:02 -0700

Allen,

What is the rise-time of the TDR you are using? You may be seeing
effects of the higher frequency content of the signal. I have used
some HP scopes with a 20 ps rise-time and that is roughly equivalent
to a 3 dB cut-off about 25 GHz! [1/(2*rise-time)] Unless you are
running signals through your traces with that kind of frequency
content or that fast, then you need to worry about such things.

Each bend or corner tends to build up charge and that is like adding
in a bunch of little femto-farad caps from your trace to its reference.
Your TDR should have several little "bumps" in it indicating each turn
in the trace which shows the charge build-up, i.e. capacitance. You should
also notice the propagation time slowing down for the serpentine trace
as well as an impedance change.

Others have pointed out coupling between the serpentining and that
could certainly cause impedance changes as well. These need to be
spaced accordingly--see Scott McMorrows's reply.

Hope this helps.

Chris Padilla
Cisco Systems

At 10:17 AM, Allen Wang wrote:
>Dear all:
>
>We are designing a Rambus related product which requires the PCB to
>maintain
>28 Ohm +/- 10% trace impedance.
>However, what we have found that a straight line trace will get a higher
>impedance (through TDR, 1 Ohm or more) than
> a trace with many turns on it. We know a turn on the trace will change
>the
>impedance. However, is it the root cause
>that make our trace impedance not consistent?
>
> Two traces are on the same layer, with the same trace width.
>
>
>+---------------------------------------------------------------------------
>------------------------------------+
> Straight line
>
> +-----------+ +------+ +---- \\\\ -----+
>+------------------------+
> | | | | | |
> | | | | | |
> +-----+ +----+ +---+
> line with turns
>
>
>Allen
>
>**** To unsubscribe from si-list: send e-mail to
>[email protected]. In the BODY of message put: UNSUBSCRIBE
>si-list, for more help, put HELP. si-list archives are accessible at
>http://www.qsl.net/wb6tpu/si-list ****
>

**** To unsubscribe from si-list: send e-mail to [email protected]. In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP. si-list archives are accessible at http://www.qsl.net/wb6tpu/si-list ****