Mr. Franck, 
	My earlier response was too hasty.  I apogize.  Indeed, different amounts of deflection 
will be encountered when guaging both sides of the board.  In figure 6, the mass of the part tends 
to lessen the deflection.  In figure 7, a worse case deflection is likely.  Concerning para. 5.4.1 & 
.2, you have a very good point, and it makes me wonder if there was a different intent.  The 
deflection with convex side down, holding 2 corners to the flat surface will always be greater than 
any of the previous procedures.  So, if this production method must meet the same procedures, 
why have the others?  I hope IPC responds to this.  I don't have the answers.

The magnitude of the problem to your customer depends on how the board rests on the conveyor 
at pick and place, or wave solder, etc. whether the bow and twist effects those operations, and if 
there is a potential for mechanical stressing resulting from the tendency of the material to return 
to the "remembered" shape.  The best practice is to find out what your customer expects, 
determine if this is realistic, and adjust your process accordingly.  The test method is best used 
to give you a base line measurement and subsequent yardstick for measuring the effectiveness of 
improvements.





On Wed, 31 Jan 1996 08:21:22 -0500 George Franck X2648 N408 wrote:

> From: George Franck X2648 N408 <[log in to unmask]>
> Date: Wed, 31 Jan 1996 08:21:22 -0500
> Subject: Bow and Twist
> To: [log in to unmask]
> 
> IPC-TM-650 2.4.22 is the test procedure for bow and twist.  It is found in
> various documents, mine is in the back of my trusty IPC-RB-276.  In section
> 5.4 Procedure 4, Production Testing using a "go/no-go" pin gague, there are
> two methods and measurements.  The first, in para. 5.4.1, (fig 6)  show the
> bowed board place cup side down and the measurement is the gap under the
> cup.  In 5.4.2, (fig 7) the board is turned over and the cup is up.  The
> measurement is at the edge or corner of the board.
> 
> My question is this.  The cup side up measurement is about twice the cup
> side down measurement.  Must both measurements meet the percent deviation
> requirement?  Is there a different way of calculating the maximum
> percentage deviation for the two?  I am confused.
> 
> You can see what I am talking about by taking out a business card, and
> curling it a little.  Now put it cup side down, look at the gap.  Turn it
> over and hold down two corners (across the edge which does not have the
> curl in it), the gap is about twice as much.
> 
> Thanks and don't let your boss catch you playing with your cards at your desk!
> 
> 
>   
====================================================================
>                            George Franck
>   PWB Product Assurance                     Phone (703) 560-5000 x2648
>   E-Systems M/S N408                              Fax   (703) 280-4613
>   7700 Arlington Blvd                  E-Mail: [log in to unmask]
>   Falls Church Va 22046                      E-Mail: [log in to unmask]
>   
====================================================================
> 
> 
FRED J. PAUL					FLUKE CORPORATION
Sr. Process Engineer				POB 9090   M/S 55
PCB Operations					EVERETT, WA 98206
[log in to unmask]					206 347-6100
voice:	206 356-5734 (w Voice Mail)
fax:	206 356-6070 
[Ideas and expressions presented herein are my own, and in no way are they implied to 
necessarily represent those of the FLUKE CORPORATION]