Re: [SI-LIST] : Plating Tails at High Frequencies

About this list Date view Thread view Subject view Author view

From: Scott McMorrow (scott@vasthorizons.com)
Date: Thu Mar 29 2001 - 03:41:50 PST


wayne,

I'm back from vacation!

Actually, there is current flowing through the plating tails. From a purely
transmission line standpoint, the tail is a transmission line stub hanging
out off the via. For example, when a wave edge is launched from the die,
when it reaches the ball via, there are two paths it can take. One down the
ball to the PCB trace, and the other down the plating tail. The wavefront does
not know apriori that the plating tail is not connected to it. That wave will split,
part going down the ball path and part going down the plating tail path, based
upon the relative impedance of both paths. Should they have nearly equal
impedance, then half of the wave will travel along the plating tail until it hit the
end (which is an infinite impedance) and reflect back with a reflection coeficient
of 1. Since there is wave propagation, there is current flow.

The effect of the plating tail will depend upon its electrical length relative to
the risetime and frequency components of the signals being driven off the die.
Since some plating tails can easily be over 250mils in length or more, they can
present significant discontinuities to fast edge rate or high frequency signals. For
example, a trace in BT epoxy and encapsulated in plastic has a velocity of
propagation of about 180 to 200ps per inch. So, a 250mil plating tail would have
an electrial length of 45ps. The end of the tail would not be "seen" by waveforms
passing through the ball until one round trip or 90ps after the first incident waveform
arrives. That plating tail would also exhibit a quarter wave resonance at 5 GHz, and
would have an adverse effect on OC48 (2.5Gbps) data streams. At lower data rates,
it will cause jitter and edge distortion of the signal transmitted through the package.

To approximate the plating tail as just a capacitor misses the fact that it as a
waveguide stub.

Also, for those interested, multiple plating tails of the same length (and there are many
in a package) will be tuned to the same resonant frequency. They can actually
"sing" together if excited with fast enough edges, or high enough data rates. A full wave
simulation can show these sorts of eigenmodes ... at least for simplified cases.

best regards,

scott

--
Scott McMorrow
Principal Engineer
SiQual, Signal Quality Engineering
18735 SW Boones Ferry Road
Tualatin, OR  97062-3090
(503) 885-1231
http://www.siqual.com

wayne.nunn@philips.com wrote:

> In a PBGA substrate there are plating tails that project from the last via to ball connection to the edge of the substrate. Since there is no current flowing through these traces, inductance shouldn't be a problem at any frequency. Are there capacitive > issues at high frequencies, say above 500mHz? > > Thanks, > > --Wayne > ------------------------------------------------------------------------------------------------------------- > Philips Semiconductors - ATO Innovation San Jose > Wayne A. Nunn > Staff Engineer > > 1101 McKay Drive > San Jose, Ca 95131 < Voice: 408.474.5620 Fax: 408.474.5505 > > ------------------------------------------------------------------------------------------------------------ > > **** To unsubscribe from si-list or si-list-digest: send e-mail to > majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE > si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP. > si-list archives are accessible at http://www.qsl.net/wb6tpu > ****

**** To unsubscribe from si-list or si-list-digest: send e-mail to majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP. si-list archives are accessible at http://www.qsl.net/wb6tpu ****


About this list Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Thu Jun 21 2001 - 10:11:21 PDT