that's interesting, apparently we will NEVER reach a consensus on this
subject!
just for the sake of comparison, here is what Gerry Gagnon said:
-=-=-=-
Subject: Re: non-functional pad removal
From: Gerald Gagnon <[log in to unmask]>
Date: Tue, 13 Jul 2010 11:36:23 -0400
This topic is design, environment, and fabricator dependant.
In most applications I have worked with, removing NFP improves the
situation.
I just got done with a project with DOEs that cyclically tested a 24 layer
all 2 ounce copper design.
The test samples with NFP removed were much more robust than the samples
with the NFP in place.
The DOE testing also indicated that in this design, retained NFP needed to
be 0.5 mm larger in diameter in order to acheive parity with the DOE testing
for the samples with NFP removed.
This has usually been my experience with designs, environments, and
suppliers I have worked with.
Your mileage may vary.
Brs,
Gerry
-=-=-
I have an interesting quote from Bill Birch at DfR (Design for Reliability),
too, that was given to me by Paul Reid
-=-=-
from Paul Reid <[log in to unmask]>
date Thu, Jul 8, 2010 at 2:12 PM
subject RE: inner-layer pad removal
With smaller vias/higher aspect ratio (.008" / 0.2mm to .020" 0.5mm) the
inclusion of internal lands is negative (10-30% reduction in long term
performance). Conversely, with larger holes/lower aspect ratio (via size -
.020"+ / 0.5mm+) the inclusion of internal lands is positive (10-15%
increase in long term performance).
Our ongoing failure analysis has shown that when .008" - .020"
PTV's contain pads on all or most layers, the barrel crack locations are
nearly always positioned between, or adjacent to the two central pads
(propagating from glass bundles in resin rich areas). With PTV's where pads
are removed, the cracks occur randomly throughout the central region of the
barrel, but not necessarily in the central zone. We believe this situation
is primarily related to the distribution and re-distribution of strain.
We have a hypothesis that when strain is applied to the sidewall of
the .008" - .020" (higher aspect ratio) PTH barrel, the pads in the central
zone create a "focal point" or stress riser for the strain. The pads
"channel" the strain to the point of least resistance, which tends to be the
thinner plated copper in the center of the PTH vias. The small improvements
in performance for products with pads removed (smaller hole size only) is
related to the fact that cracks occur in areas away from the center, where
the PTH copper is slightly thicker.
-=-=-
finally, a post by Werner Engelmaier might be a good compromise
-=-=-
With PTHs/PTVs, there are essentially 2 failure modes: (1) barrel cracks,
and (2) innerlayer separation [post-separation].
Barrel cracks typically occur in the barrel wall near the PCB center where
the PTH is surrounded by a prepreg layer, because of the huge thermal
expansion mismatch between the ED Cu and the surrounding resin above Tg and
the CTE(z) is larger for the prepreg layers than the core layers.
Innerlayer separation occurs near, or at the surface [shoulder crack] of the
PCB [layers 1, 2, 3 n-2, n-1 and n] due to the combination of radial tensile
stresses, as the resin tries to push into the hole, and torque loading due
to land rotation.
The radial tensile loads can be reduced by MORE innerlayer ribs supporting
the barrel wall [think of a submarine]. One can reduce the land rotation
torque loading by NOT removing non-functional lands on layers 1, 2, 3 n-2,
n-1 and n.
-=-=-
Sorry to flood you with so much unrequested history, but I am wondering if
your failures may have been exacerbated by removing ALL of the unused inner
layer pads instead of the "middle third" of the stack, as many are
recommending now. Were your failures randomly distributed in various parts
of the hole? most at the knee? most at the center? Can you share data or is
it proprietary?
Anyway, I'm tempted to go with the note:
NON-FUNCTIONAL INNER-LAYER PADS MAY NOT BE REMOVED FROM LAYERS 1, 2, 3, N-2,
N-1 AND N.
If anyone is interested, DfR is willing to perform an additional study, if
someone is willing to help create the board design.
thanks for sharing your experience,
Jack
.
On Thu, Aug 5, 2010 at 1:47 AM, Paul O Connor <[log in to unmask]
> wrote:
> I carried out thermal shock testing on 12 Layer 2oz cu daisy chain PTHs
> boards, Aspect ratios of 4:1 & 6:1 with 4 different hole wall thicknesses
> (from 22um up to 70um) approx 1300 Test hours
> the results shows the holes with pads removed fail first independent of
> hole
> wall thickness or aspect ratio - so I do allow removal of redundant inner
> layer pads.
>
>
>
>
______________________________________________________________________
This email has been scanned by the MessageLabs Email Security System.
For more information please contact helpdesk at x2960 or [log in to unmask]
______________________________________________________________________
---------------------------------------------------------------------------------
DesignerCouncil Mail List provided as a free service by IPC using LISTSERV 1.8d
To unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF DesignerCouncil.
To temporarily stop/(restart) delivery of DesignerCouncil send: SET DesignerCouncil NOMAIL/(MAIL)
Search previous postings at: www.ipc.org > On-Line Resources & Databases > E-mail Archives
Please visit IPC web site http://www.ipc.org/contentpage.asp?Pageid=4.3.16 for additional information, or contact Keach Sasamori at [log in to unmask] or 847-615-7100 ext.2815
---------------------------------------------------------------------------------