[SI-LIST] Re: The return of nickel!
- From: jeff_latourrette@xxxxxxxxxxx
- To: yost@xxxxxxxxxxxxxxxxxxx
- Date: Fri, 11 Apr 2003 16:09:33 -0600
Boris:
Am assuming difference is significant and well above your measurement
repeatability and above unit-to-unit variations. On most EMI tests, we don't
call things really different until we get into ~5 dB or more, although we are
usually able to get repeatability of a dB or so. If test time and $ are not
too constrained, test each unit 5 times and then test 5 of each unit and you
can get a better idea if difference is real.
Is board being scanned at an OATS range or chamber and maximized over all
angles, antenna heights ?? Repeatability of test facility/operators ??
Fixturing, cabling or Test Chassis and their positioning ?? If radiated beams
are very narrow, sometimes differences occur due to resolution/steps/variation
in those variables. Automated test routines or manual ??
Given what you describe, it's hard to conclude that there'd be significant
shifts in loss or impedance from the two fab processes. Nickel has lower
conductivity than gold or copper and has a non-unity magnetic permeability
constant, which will reduce both Q (higher loss) and skin depth (how deep RF
energy penetrates conductive layer). Higher frequency energy travels more on
the outside (top and bottom) of traces and only penetrates a skin depth or two.
Skin depth over your frequencies of interest (assuming they'd be tenths of a
mil) can be calculated for both cases to see whether energy is in gold, nickel
or copper. Assuming Au & Ni plating are very thin wrt Cu, energy is probably
mostly in thicker Cu for both cases, making losses ~ equal, unless you are at
very high frequency or using very thick (tenths of a mil) Au or Ni plating. A
much lossier board (very thick surface nickel, lower Q) should radiate less.
Material differences and their effect on skin depth and impedance can be
simulated using Linecalc/Momentum (part of Agilent ADS) or whatever simulation
tools you use. It may be much harder to predict corresponding EMI predictions
wrt to these differences and they could well be insignificant. Without seeing
your board, I think you would have to see a very large difference in impedance
or loss to see significant changes in EMI, so I'd look more at what's on your
board.
Something practical to try: Can you get an SMA, SMB or other RF connector on
to one of your larger traces and externally excite both board types (bare and
loaded) with EMI frequencies of interest, measuring radiation ? This might
help you to eliminate any loading differences or component variations and show
you real difference in boards, if any. For instance, difference might be in
the clock/crystal or IC (or their bonding/packages) that drive the traces in
the board. Testing more samples will also highlight this kind of variation.
Hope that can help,
Jeff LaT.
-----Original Message-----
From: Boris Yost [mailto:yost@xxxxxxxxxxxxxxxxxxx]
Sent: Friday, April 11, 2003 11:23 AM
To: Si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] The return of nickel!
Dear SI-ers:
I have two boards. One of them is louder than the other w.r.t. EMI at
high
frequencies. They were made at different times from different lots from the
same artwork at the same factory. I'm trying to track down the differences
between them. One real obvious one (a marketing guy pointed it out) is that
they are a different shade of green....
The reason that they are different colors is that one board has
copper->immersion gold->solder mask, and the other has copper->solder
mask->immersion gold. So therefore on one board all the length of all the
traces on the outer layers has nickel on the surface; the other one has Ni
just on the pads. The one on which they laid the gold (and associated
nickel) down first is the noisier one.
Is this significant, or should I forget this one.
Boris
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List archives are viewable at:
//www.freelists.org/archives/si-list
or at our remote archives:
http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List archives are viewable at:
//www.freelists.org/archives/si-list
or at our remote archives:
http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Other related posts:
