Bow and Twist of PCB's

Hi there, I'm actually facing a lot of trouble with PCB's presenting up to 3mm of bow /twist at reception i.e. before assembly. The PCB size is 220 by 230mm. The boards are to be assembeled SMD double sides with 05.mm pitch on both sides. In order to discuss with the PCB manufacturer I was looking to find some norms / standards. Could only find IPC-A-600 indicating 1,5% as tolerance. To my experience this is far to much to achieve a good result in assembly (screenprint,placement). Does anybody have any experience and/or norm/standard that I could use in my discussion with the PCB manufacturer? Thanks in advance, Tom

reply »

| Hi there, | I'm actually facing a lot of trouble with PCB's presenting up to 3mm of bow /twist at reception i.e. before assembly. The PCB size is 220 by 230mm. The boards are to be assembeled SMD double sides with 05.mm pitch on both sides. In order to discuss with the PCB manufacturer I was looking to find some norms / standards. Could only find IPC-A-600 indicating 1,5% as tolerance. To my experience this is far to much to achieve a good result in assembly (screenprint,placement). | Does anybody have any experience and/or norm/standard that I could use in my discussion with the PCB manufacturer? | Thanks in advance, Tom It is possible to get to .005" inch per inch. .007" often is specified and achieved. However, a lot rides on DFM. The following is a posting I made recently on the IPC forum: Date: Wed, 12 Aug 1998 07:45:16 -0700 Reply-To: "TechNet E-Mail Forum." , Earl Moon Sender: TechNet From: Earl Moon Subject: PCB Warp / Bow & Twist - ReplyContent-Type: text/plain Your boards are fairly dimensionally stable to begin with because of the low layer counts. Beyond that, the following are basics: 1) Using balanced constructions with respect to reference plane locations (equally distributed in Z axis) and features (openings in metal equal on all planes, etc. to assure equal mass). If not, you have a bi-metal strip situation resulting in potato chips instead of PCB's. 2) Using low resin to glass ratio prepregs and core materials (7628,7629, 2116, as examples) - the more glass the more dimensional stability in all axes with tradeoff being reduced innerlaminar bond strength for the first two glass styles. 3) Using like, or homogeneous glass styles (60's yarn as an example using 106, 1080, 2113, and 2116 materials) so yarn bundles "nest" together and effect little relamination induced stress (later being relieved thermally during soldering operations causing excessive warpage). 4) Using unbalanced constructions with respect to higher layer count MLB's wherein a resin rich glass style (106 or 1080) is laid up with a less resin rich material (2113 or 2116) to provide the best of both worlds as constructions with higher relative dimensional stability and high bond strength. As an example for a 5 core construction, use one ply of 106 with one ply of 2113 instead of using two plies of 1080 and soon. Earl Moon (also at pod@ix.netcom.com) Tom, You also need a bi-directional dialogue capability with your suppliers. I mean, whoever you talk with must understand some basics as well as increasingly demanding requirements and how to achieve them. I have been designing, fabricating, and assembling boards for some time and the one constant in all this is compromise and that the stuff above needs constant relearning/teaching. This ain't rocket science but building PCB's, especially those that are acceptable, is a demanding task. There are more processes and sub processes in this business than most others and with the way new designs are popping up with more demands, it's not possible to build them without some basic knowledge practiced by qualified DFM team members (from the first designer to the last supplier/customer). Earl Moon

reply »

| Hi there, | I'm actually facing a lot of trouble with PCB's presenting up to 3mm of bow /twist at reception i.e. before assembly. The PCB size is 220 by 230mm. The boards are to be assembeled SMD double sides with 05.mm pitch on both sides. In order to discuss with the PCB manufacturer I was looking to find some norms / standards. Could only find IPC-A-600 indicating 1,5% as tolerance. To my experience this is far to much to achieve a good result in assembly (screenprint,placement). | Does anybody have any experience and/or norm/standard that I could use in my discussion with the PCB manufacturer? | Thanks in advance, Tom

Tom, The simplest and easiest way to find what "your" norm should be is to check the SPECS on your assembly equipment. Then use the tightest one in your acceptibility specification to the vendor. Barring anything "unusual" in your board design, or assembly process any reputable vendor should be able to meet or exceed the maximum board warpage spec for any standard piece of assembly equipment on the market today. If not,..find another vendor. Mark j.

reply »