Re: [SI-LIST] : Via Capacitances ...

Fred Balistreri (fred@apsimtech.com)
Tue, 13 Jul 1999 11:36:35 -0700

Dave Hoover wrote:
>
> The trace did not connect to the reference planes. And the
> diameter of the Probe Card was less than 12" (Dia) so the trace
> lengths were rather short. So the effect was large proportionately
> speaking with regards to the trace size (length). It definitely
> dropped the impedance. Doug Brooks has an interesting paper
> on Impedance and vias at his webpage.
> try http://www.ultracad.com/articles.htm
> This one seems to hit the nail on the head and then some.
>
> Dave
>
> -----Original Message-----
> From: rbmccammon@mmm.com [mailto:rbmccammon@mmm.com]
> Sent: Tuesday, July 13, 1999 06:31 AM
> To: si-list@silab.eng.sun.com
> Subject: RE: [SI-LIST] : Via Capacitances ...
>
> I think we should consider two cases.
>
> 1. A trace encounters a via and then continues on the same plane.
>
> 2. A trace encounters a via, goes and goes through its reference plane
> and continues on the other side.
>
> **** To unsubscribe from si-list: send e-mail to
> majordomo@silab.eng.sun.com. 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 majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP. si-list archives are accessible at

I guess I don't quite understand your scenario. My discussion was on
vias. As I mentioned the cutouts in the planes effect the capacitance
of the via. Your experience showed exactly that. A via is not like a
coax in that a coax is surrounded by a shield or gnd return for high
frequency operation. The via has no such return. Furthermore it is the
pads that make a via capacitive. If one only has the barrel to contend
with then the via is quite inductive, especially for thick boards. The
reason I bring this up is because Intel and few other are now pushing
micro-vias. A micro-via has a much smaller pad area. And people are
going to be shocked if they believe that all vias are capacitive.
I also cannot believe people who claim that a given structure is
capacitive or inductive. In truth there is "no meaning" to such state-
ments. It's true that the impedance of a given structure may be
different than the trace impedance on a PCB. That does not make the
structure inductive or capacitive. For example:

Suppose I have a PCB trace that is 75 ohm impedance and a via on that
trace that goes through the board and connects to another 75 ohm trace
on the other side. If I tdr the thing I may get a discountinuity due
to the via. I may find that the via drops the impedance to 40 ohms.
So some would conclude that the via is capacitive. However I can design
a PCB trace to have 30 ohms impedance on the same stack up. Now if I
tdr the system the via has not changed so I still get 40 ohms. Only
this time instead of a dip a get a spike. So some pundents would now
claim the via is inductive. Both arguments are absurd. The via has both
inductive, capacitive and even resistive elements to it. Those entities
can be controlled by the physical design of the via. We know very
precisely what to change to adjust the capacitance, inductance and
resistance. The via unlike transmission lines or coax is not a good
distributed system. It is non uniform physically and therefore the
same electrically. Its components are better modeled by lumped circuit
parameters. The pads exhibit high capacitance and low inductance,
where as the barrel exhibits exactly the opposite.

Best Regards,

-- 
Fred Balistreri
fred@apsimtech.com

http://www.apsimtech.com

**** To unsubscribe from si-list: send e-mail to majordomo@silab.eng.sun.com. 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 ****