Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


bowtwist

Jason

#10524

bowtwist | 21 July, 1999

I have pcb's in which i am experiencing warping. the boards are 12.8125" on a diagonal measurement. they are out approximatley .010". the boards are combination thru-hole,smt, they also have 3ounce copper.what is acceptable,and what is the propper formula for measuring? any help would greatly be appreciated thank you

Jason MacKinnon

reply »

John Thorup

#10525

Re: bowtwist | 21 July, 1999

| I have pcb's in which i am experiencing warping. | the boards are 12.8125" on a diagonal measurement. | they are out approximatley .010". | the boards are combination thru-hole,smt, they also have 3ounce copper.what is acceptable,and what is the propper formula for measuring? | any help would greatly be appreciated thank you | | Jason MacKinnon | Do you really mean .01" bow/twist? If so I wouldn't worry. IPC spec for an SMT board would be 0.75% - considerably more than you have. (IPC-A-610-B-8.3) If you meant 0.1" is it on the bare FAB? If so, send it back. If it's after reflow or wave, look at the copper distribution. Is it equal on both sides in terms of distribution and amount. Is it 3oz on both sides? Search this forum for much more material. Bottom line - is it causing a problem in placement, soldering, testing or its final use? IPC states " keep in mind form, fit and function without jeopardising reliability".

reply »

Scott

#10526

Re: bowtwist | 21 July, 1999

| | I have pcb's in which i am experiencing warping. | | the boards are 12.8125" on a diagonal measurement. | | they are out approximatley .010". | | the boards are combination thru-hole,smt, they also have 3ounce copper.what is acceptable,and what is the propper formula for measuring? | | any help would greatly be appreciated thank you | | | | Jason MacKinnon | | | Do you really mean .01" bow/twist? If so I wouldn't worry. IPC spec for an SMT board would be 0.75% - considerably more than you have. (IPC-A-610-B-8.3) If you meant 0.1" is it on the bare FAB? If so, send it back. If it's after reflow or wave, look at the copper distribution. Is it equal on both sides in terms of distribution and amount. Is it 3oz on both sides? Search this forum for much more material. Bottom line - is it causing a problem in placement, soldering, testing or its final use? IPC states " keep in mind form, fit and function without jeopardising reliability". | IPC-A-600E Section 2.10.6 has a Flatness spec. "The entire protion of the specimen or production board shall have a maximum bow and twist of 1.5% unless otherwise specified on the master drawing. For automated assembly and surface mounting requirements, more stringent requirements may be necessary (see IPC-D-275).... Note: The stated flatness requirements may not meet surface mount board requirements".

So unless you have your own requirements, the board shop will delivery <1.5% as determined by IPC-TM-650, method 2.4.22 which has four procedures of either cut to size panels or finished rigid PCBs including single-sided, double-sided, and multilayer.

The easy way to measure twist is to apply sufficient pressure to insure that three corners are in contact with the datum surface (flat) and record the height (or thickness) of the board, this is R2. Measure the height of the lifted edge, this is R1. Measure the diagonal dimension, this is Lenght There are four opportunities to measure twist, use the worst example.

Percent twist =100 x ((R1-R2)/(2 x Lenght))

Note, this is the quick method; the real, and more dedious, way is to support the lifted edge until the other three corners touch the datum without applying pressure. Here weight plays into the issue and accentuates the twist. But if you're way under spec using the quick method, using the more tedious method will still pass the board.

Bow has one edge on the datum and measure R2 as above. Measure height at the highest point of the bowed edge, this is R1. Measure the length of the bowed edge, this is L.

Percent bow = 100x (R1-R2)/L

That's what the manuals say. Sorry, I don't have a copy of IPD-D-275 to dig into the specs any further. Plugging in your figures in shows a 0.04% twist, well within the parameters. By working back from 1.5%, you'll be able to measure R1-R2 at at 0.38" before you can reject the board (unless you have more stringent requirements), sorry even 0.1" is acceptable by the standards -- even if you impose the 0.75% rule you're allowed 0.19" twist (I did round down on these dimensions).

Anyway, that's what I pulled from the specs...

That's why it's important to know what you need and make sure it's written into your master documents.

Moonman, any comments?

Scott

reply »

George Verboven

#10527

Re: bowtwist | 22 July, 1999

| I have pcb's in which i am experiencing warping. | the boards are 12.8125" on a diagonal measurement. | they are out approximatley .010". | the boards are combination thru-hole,smt, they also have 3ounce copper.what is acceptable,and what is the propper formula for measuring? | any help would greatly be appreciated thank you | | Jason MacKinnon | Warpage should be as minimal as possible. Board warpage may cause board transfer errors and reduce placement errors. The warp may be conflict with max. warp the PCB-manufacture allows. note: When processing boards through machines with different warpage requirements, design the board using the tightest specifications.

Some machines accept more down warp as up warp.

reply »

Earl Moon

#10528

Re: bowtwist | 22 July, 1999

| | | I have pcb's in which i am experiencing warping. | | | the boards are 12.8125" on a diagonal measurement. | | | they are out approximatley .010". | | | the boards are combination thru-hole,smt, they also have 3ounce copper.what is acceptable,and what is the propper formula for measuring? | | | any help would greatly be appreciated thank you | | | | | | Jason MacKinnon | | | | | Do you really mean .01" bow/twist? If so I wouldn't worry. IPC spec for an SMT board would be 0.75% - considerably more than you have. (IPC-A-610-B-8.3) If you meant 0.1" is it on the bare FAB? If so, send it back. If it's after reflow or wave, look at the copper distribution. Is it equal on both sides in terms of distribution and amount. Is it 3oz on both sides? Search this forum for much more material. Bottom line - is it causing a problem in placement, soldering, testing or its final use? IPC states " keep in mind form, fit and function without jeopardising reliability". | | | IPC-A-600E Section 2.10.6 has a Flatness spec. "The entire protion of the specimen or production board shall have a maximum bow and twist of 1.5% unless otherwise specified on the master drawing. For automated assembly and surface mounting requirements, more stringent requirements may be necessary (see IPC-D-275).... Note: The stated flatness requirements may not meet surface mount board requirements". | | So unless you have your own requirements, the board shop will delivery <1.5% as determined by IPC-TM-650, method 2.4.22 which has four procedures of either cut to size panels or finished rigid PCBs including single-sided, double-sided, and multilayer. | | The easy way to measure twist is to apply sufficient pressure to insure that three corners are in contact with the datum surface (flat) and record the height (or thickness) of the board, this is R2. | Measure the height of the lifted edge, this is R1. Measure the diagonal dimension, this is Lenght There are four opportunities to measure twist, use the worst example. | | Percent twist =100 x ((R1-R2)/(2 x Lenght)) | | Note, this is the quick method; the real, and more dedious, way is to support the lifted edge until the other three corners touch the datum without applying pressure. Here weight plays into the issue and accentuates the twist. But if you're way under spec using the quick method, using the more tedious method will still pass the board. | | Bow has one edge on the datum and measure R2 as above. Measure height at the highest point of the bowed edge, this is R1. Measure the length of the bowed edge, this is L. | | Percent bow = 100x (R1-R2)/L | | That's what the manuals say. Sorry, I don't have a copy of IPD-D-275 to dig into the specs any further. Plugging in your figures in shows a 0.04% twist, well within the parameters. By working back from 1.5%, you'll be able to measure R1-R2 at at 0.38" before you can reject the board (unless you have more stringent requirements), sorry even 0.1" is acceptable by the standards -- even if you impose the 0.75% rule you're allowed 0.19" twist (I did round down on these dimensions). | | Anyway, that's what I pulled from the specs... | | That's why it's important to know what you need and make sure it's written into your master documents. | | Moonman, any comments? | | Scott | You folks have determined the measurement specs and methods. We've talked about all the rest as balanced constructions with respect to internal and external reference planes. These can be any copper weight provided they are closely matched in terms of placement throughout the MLB structure and have essentially the same areas in the same locations.

One thing often not discussed is using homogeneous, unbalanced MLB constructions. As I indicated before, and as shown in the material and constructions part of my article, this means selecting and indicating (on the master drawing) exactly what materials go where. I know you must leave process freedom to board fabricators, but you must know the consequences of using certain glass styles and resin contents without understanding their effects.

With proper metal plane material selections, it is possible to produce extremely flat MLB's. They can exceed all specified requirements during fab, assembly, and operation. For very detailed information, start with your "qualified" board shop and visit a material supplier to see how they make what goes into multilayer boards.

Also, there are many good papers on the subject. One of the best was written by the late Dr. Paul Craven when he was at LTI (now a part of Polyclad). Another was written by Phil Hinton (hangs out on the IPC Technet), and I've written considerably on this important subject.

Earl Moon

reply »

High Throughput Reflow Oven
Benchtop Fluid Dispenser
Manufacturing Software

ICT Total SMT line Provider