Eurocircuits Printed circuits blog

What keeps us busy at Eurocircuits, projects we are working at, new idea's, background information and a platform where you can participate, give your opinion and guide us to what is important for you as an electronics developer

Posted by Lengyel Norbert
Lengyel Norbert
Norbert is quality manager in Eurocircuits Kft
User is currently offline
on Tuesday, 31 January 2012
in Technology

Bow and Twist in printed circuits

What is Bow and Twist?

According to the IPC-A-600 standard bow and twist (flatness of the board) is :

“Flatness of printed boards is determined by two characteristics of the product; these are known as bow and twist. The bow condition is characterized by a roughly cylindrical or spherical curvature of the board while its four corners are in the same plane. Twist is the board deformation parallel to the diagonal of the board such that one corner is not in the same plane to the other three. Circular or elliptical boards must be evaluated at the highest point of vertical displacement. Bow and twist may be influenced by the board design as different circuit configurations or layer construction of multilayer printed boards can result in different stress or stress relief conditions. Board thickness and material properties are other factors that influence the resulting board flatness.”

Why is the flatness of a printed circuit board important?

  • During the production of the board the flatness of the panels is important for handling and for positioning the panels on the machines
  • During the assembly process the flatness of the panels is important for correct solder paste deposition and component mounting
  • Flatness is an aspect of the visual quality appearance of the boards.

What are the acceptability criteria for bow and twist?

  • For all boards the bow and twist should be 1.5 % or less
  • For boards using SMD components ( the majority of the boards) the bow and twist shall be 0.75% or less.

How to measure the bow and twist?

The IPC-TM-650 test methods manual describes the method to calculate bow and twist percentages

What can the PCB designer do to avoid bow and twist?

 

 

  • Create a symmetrical copper distribution. As far as possible aim for an even copper distribution across each layer.  For multilayers as far as possible arrange signal and plane layers symmetrically around the centre of the PCB.  If there are areas of very low copper density and areas of high density or full copper it's a good idea to add copper to the low density areas to balance out the copper distribution.
  • Select a symmetrical build-up of cores, pre-pregs and copper thicknesses.

What can a PCB producer do to avoid bow and twist?

  • Select base materials that are suitable for lead-free soldering. We use for instance IS400 from Isola or NP155 from Nanya
  • Use proper pressing parameters for multilayers to reduce stress in the final PCB
  • Do not mix materials from different types or vendors, and lay up the material warp and weft correctly
  • Use horizontal ovens for the curing processes
  • Cool down the panels on a horizontal surface (for instance after hot-air solder-levelling)

What can an assembly house do to avoid bow and twist?

  • Avoid heavy thermal shock during the soldering process by using a suitable soldering profile
  • Organise adequate support during the soldering process.

Your opinion?

Even if the pcb producer and assembly house take the necessary care to avoid bow and twist the deciding factor is the design of the board.
We have been brainstorming at Eurocircuits recently to see if we can develop a tool for electronics designers to predict the risk of bow and twist. This could be in the form of an index or a visual tool in the same way as our latest plating simulation tool .
As we cannot judge how useful this is for electronics developers, we ask you to comment on this post with your opinion. If there is a genuine interest in such a tool, we will (try to) develop it. You can also give your opinion on the plating simulation tool, and suggest to us how we can improve it , Or just give your opinion in our poll below

Bow and Twist simulator

Do you want us to develop a tool to predict bow and twist on printed circuits?

150

Votes

Vote
I would definitely use the tool when it becomes available

Proposed by Administrator on 01.February 2012, 09:23 Clock.

134

Votes

Vote
No, thanks

Proposed by Administrator on 01.February 2012, 09:23 Clock.

126

Votes

Vote
It might be useful, but not for me

Proposed by Administrator on 01.February 2012, 09:23 Clock.

» Go to poll »
3 Votes left

jVoteSystem developed and designed by www.joomess.de.

 

 

 

 

Norbert is quality manager in Eurocircuits Kft

Comments

Guest
Nelson Tuesday, 06 March 2012

What is the acepptance about twist to naked PCBs? The IPCA 610 informes about pcbs with smd components after solder.

Lengyel Norbert
Lengyel Norbert
Norbert is quality manager in Eurocircuits Kft
User is currently offline
Lengyel Norbert Wednesday, 07 March 2012

IPC-A-600 standard is dealing with the bare boards. According this:
“• For printed boards using surface mount components, the bow and twist shall be 0.75% or less.
• For all other boards, bow and twist shall be 1.50% or less.”

Guest
Mugurel Thursday, 10 May 2012

Hello all,
According with IPC-A-600 standard the max. twist accepted is 0.75% or less (for PCB's which use SMT components).
My question is: 0.75% from what?
I try to understand the specification of standard and method 2.4.22 also, but I think something is missing.

Lengyel Norbert
Lengyel Norbert
Norbert is quality manager in Eurocircuits Kft
User is currently offline
Lengyel Norbert Friday, 11 May 2012

Dear Mugurel,
The missing thing is diagonal of the PCB. Please calculate the percentage of twist as follows: (R/2D)x100. Where R= Go/No-Go gauge size, D=Diagonal of PCB. This formula includes a factor of two because, by constraining one corner of the sample, the vertical deflection of twist is approximately doubled.
If needed you can get more information from the web site of IPC:
http://www.ipc.org/4.0_Knowledge/4.1_Standards/test/2.4.22c.pdf

Leave your comment

Guest
Guest Thursday, 17 May 2012