just treat it as potato chip... you can measure using many techniques if
you wish, laser scan is another, or CMM.  (you can print one of your PCA
white to  minimize the optical color errors due to different reflectivity.
 IMHO
> Yes, believe me many of us have done it.
> There is always a strip of unpopulated real estate along the edges of the
> cards if they are designed properly, and I have never seen a board without
> some areas of unpopulated real estate in the inner areas (has anyone?).
> The reference plane is easily obtained by putting the finished CCA or bare
> PWB into a simple fixture again assuming proper design with at least three
> tooling holes on the outside corners (or really anywhere come to think of
> it). A stainless steel plate with three standoffs of equal height with
> respect to the bench surface is all that is required to hold 99% of all
> CCAs manufactured. You do NOT want to clamp down the CCA/PWB to the
> standoffs, just make sure the bottom surface of the board is making
> contact with all three standoffs. With the CCA thus elevated the
> bottomside components are no longer a concern.
> Then you simply use a dial indicator
> (https://www.mscdirect.com/browse/tnpla/06422976?cid=ppc-google-New+-+Measuring+%26+Inspecting+-+PLA_sW2Cu6rhf___164110844316_c_S&mkwid=sW2Cu6rhf|dc&pcrid=164110844316&rd=k&product_id=6422976&gclid=CKLq1-Le-9ICFRaewAodHc8Eug)
> to measure the amount of deflection (amount of warpage) along the outside
> edges of the CCA, and you can also make measurements at points inside the
> area of the CCA to obtain rates of deflection per cm to determine if the
> CCA meets the requirement of maximum change per cm in any direction. If
> the deflection is greater than the specified rate of change per cm in ANY
> direction ANYWHERE on the CCA/PWB the warpage is OOT (out of tolerance).
> Be sure your reference points are all on soldermask or all on the
> laminate. Do not measure pads, traces, vias, etc.
>
> Every self-respecting engineer should know how to do this; why should only
> some of us have this pleasure?
>
> So have fun with that!
> Odin
>
> -----Original Message-----
> From: TechNet [mailto:[log in to unmask]] On Behalf Of Lum Wee Mei
> Sent: Tuesday, March 28, 2017 10:14 PM
> To: [log in to unmask]
> Subject: Re: [TN] IPC-TM-650, 2.4.22, Bow & Twist
>
> It is this Bow & Twist season again.
> I often wonder whether someone really measured Bow & Twist on assembled
> boards. If yes, wonder how it was done as both sides are already populated
> with components.
>
> ~wee mei~
> ________________________________________
> From: TechNet [[log in to unmask]] on behalf of Joyce Koo [[log in to unmask]]
> Sent: Wednesday, March 29, 2017 3:59 AM
> To: [log in to unmask]
> Subject: Re: [TN] IPC-TM-650, 2.4.22, Bow & Twist
>
> it depends your design. if you have some small parts in the max bow area,
> it may flight out upon installation... (1) your PCB should compliant with
> bow and twist with simulated reflow - how many time you reflow, including
> rework if it required (2) your PCA depend upon what you how you want to
> install or support points and what is density and type of parts... better
> not be Gyro... those need really flat ;-).
>> Fellow TechNetters:
>>
>>    Does the above stated TM applies to raw cards ONLY?   Is there a
>> standard for PCBA?
>>
>> Victor,
>>
>