Since I was surprised by the answers, I am willing to do some studying on
the issue because I thought we had that one nailed down.
But just for the sake of a starting point, what if we had:

   - a typical 8-layer board with vias that are 5:1 max aspect ratio (13 or
   20mil)
   - 1000 boards a year (low volume, not worried about cost of drill bits)
   - high thermal cycling in automotive environment

Would you leave the internal non-functional pads or remove them if you
wanted the boards to last 20 years?
What is the MOST ROBUST choice?
Sorry for the repeat question, I thought this issue was settled!

Jack (aka "the new guy")

p.s. that's a joke, because every time I think I know everything, I get
slapped!

On Mon, Mar 5, 2018 at 11:47 AM, Jack Olson <[log in to unmask]> wrote:

> I hear ya, brother. and don't think I am arguing with the value of doing
> the work,
> but (here I go anyway)
> One of the major benefits of having the IPC in our industry is that every
> new circuit board designer that comes along, who doesn't know whether to
> put that note on his fabrication drawing or not, can learn from the
> experience of others here. He doesn't have to repeat expensive tests just
> to verify what is already understood by the consensus of other members that
> have already contributed time and effort and resources to study the issue.
> (but I hear ya!)
>
> Jbro
>
> On Mon, Mar 5, 2018 at 10:51 AM, Dwight Mattix <[log in to unmask]>
> wrote:
>
>> Yeah, I blew past the thread to busy to contribute last week.
>>
>> All else being equal retaining non-functional pads tends to increase time
>> to barrel wearout from circumferential cracking at/near the midline.  One
>> notion being that the pads make for a local lower resin content region
>> around the via (less expansive dielectric acting on the copper column to
>> wear it out).
>>
>> Non-func pads don't seem to have much bearing on the actual post
>> interconnect reliability.  Multiwire anyone?  (speaking of dinosaurs).
>>
>> But back to the barrel wearout thing...  All things are rarely equal. So
>> if you actually build it and test it (due diligence? That's crazy talk. Who
>> has time or money for that anymore?), I'd wager you'll get mixed results.
>>  A lot depends on factors like the fabricator's drilling and hole prep
>> skill, the aspect ratio, material involved, copper weights of the included
>> NF-pads etc.
>>
>> Leaving non-func pads in, increases the drilling challenge. That
>> increases things that disrupt the hole and factor in to it's ultimate
>> reliability. For example, a rougher hole wall is very likely to be part of
>> the effect of leaving NF pads in. That introduces stress risers in the
>> holewall topography that can accelerate copper wearout and crack
>> propogation.
>>
>> So all of that say the original question, "It depends."   :)
>>
>> Do your due diligence. If it really matters, build it and test it. Better
>> yet, build it both ways and test it. Even better, build both ways at more
>> than one fab and test it.  You'll be illuminated and smarter at the end of
>> the exercise than 99 of 100 veteran pwb tech people seen walking the floor
>> with tacky polo shirts, broken down posture and done-lops last week at
>> IPC/APEX.
>>
>> cheers,
>> dw
>>
>> -----Original Message-----
>> From: TechNet [mailto:[log in to unmask]] On Behalf Of Jack Olson
>> Sent: Monday, March 5, 2018 7:49 AM
>> To: [log in to unmask]
>> Subject: Re: [TN] Non Functioning pads
>>
>> I must be behind-the-times on this one.
>> Werner Englemaier used to talk about this all the time, and from my
>> memory an analogy might be similar to comparing a simple rivet to a "rivet
>> with ribs". If your goal is a robust product (which a lot of Class 3's are)
>> then it seems like you would want the extra support. (I'm not making a
>> statement, I'm repeating what I was taught). I'm pretty sure I've heard
>> Gary Ferrari recommend keeping them in at least a half a dozen times in his
>> seminars. Aren't the most common failures in boards related to vias?
>>
>> Maybe I have more learnin' to do on this one, but I'm surprised that NO
>> ONE responded in favor of keeping the unconnected internal pads
>>
>> Well, since we are talking about vias, I was also advised to require 1mil
>> hole wall plating, even though the standard is 0.8 (I think). But for the
>> same reason, the stronger we can make our vias, the less "most common"
>> failures we will have, right? (At Caterpillar, we want the most robust
>> product we can get for the money) Am I sounding like an old dinosaur now?
>>
>> but really, is EVERYONE removing unconnected inner layer pads?
>> Is "increasing the longevity of drill bits" the dominant theme now?
>>
>> onward thru the fog,
>> Jack
>>
>>
>> .
>> On Tue, 27 Feb 2018 11:12:22 -0600, Larry <[log in to unmask]> wrote:
>>
>> >Is there any reason I cannot remove non functioning pads on the inner
>> layers for an Class 3 PCB?
>> >
>> >Many thanks,
>> >
>> >Larry
>>
>
>