Minimizing Board Warpage During Reflow

Does anybody know if a contraption exists that will minimize a PCB's tendency to become slightly bent after reflow? We have a new product that we're planning to assemble in a couple of months. The board is 9.7" by 13.4" and it is imperitive that we not have any portion of it sag beyond one tenth of an inch from it's plane. Any help that can be offered on this topic would be greatly appreciated. Thank you! Anthony

reply »

| Does anybody know if a contraption exists that will minimize a PCB's tendency to become slightly bent after reflow? We have a new product that we're planning to assemble in a couple of months. The board is 9.7" by 13.4" and it is imperitive that we not have any portion of it sag beyond one tenth of an inch from it's plane. Any help that can be offered on this topic would be greatly appreciated. Thank you! Anthony, As times past, I hate to be the bearer of bad tidings. What you are asking is not possible with normal PCB/MLB materials as epoxy, polyimide, BT, or other resin systems. It's the nature of the beast and the design rules and material selections going into the process and resulting product. If one is compromised (say for electrical characteristics as impedance) another must take its place or suffers. There is so much to say here it takes a book or article to explain it all. I have the article waiting to be published you may contact me about. However, though not knowing your customer's needs, there are some new material types sometimes (again dependent upon designs, constructions, and process capabilities) improving dimensional stability in all axes. Remember, IPC guidelines suggest an acceptable limit of 1.5% overall bow and twist. This number puts your board way out of your needs. Many customers, and some suppliers, require and get .005"/" to .007"/" but everything as suggested above must be exactly right. For a board as large as yours even using a permanently attached stiffner (these may be built into the board during fabrication) will result in "sag" somewhere. If you had the room (it appears you don't), you could do what some R/F design types do and that's place a "web" of aluminum (mode blocks) as an anti warpage deterent. Other than that, you may contact me at my email or phone me at 734-654-0981 and I can fill you in on some important design and material details to make life easier though probably not luxurious. Best wishes on a constant struggle for flatness in the PCB world, Earl Moon | Anthony

reply »

Hi Anthony! Earl's right, things can get rather involved when you start discussing ALL the things that affect whether or not a board is going to warp. The way I look at the issue is try to separate the warp issue into two catagories; 1. The board being flat prior to SMT, and 2. The board being flat after assembly. I think between the two areas, it's easier to deal with the first issue than it is with the second. Obviously, the main enemy of board flatness is heat, be it in the PCB fabrication process, or the assembly process. Just a few things to think about to address minimizing warping in the fabrication process: 1. There are many different laminate materials that are more stable than FR4, but it comes at a price. Almost every other material is going to be more expensive than FR4. 2. Balancing the copper content in the board, distributing it as equally as possible within each layer, and throughout the overall board. If the design allows, using cross-hatch types of patterns, or I've even seen circles distributed in a grid-pattern left in areas where there are no traces or planes to balance the copper. By doing this you not only improve the temperature stability of the board, but it makes things easier for your fab vendor to etch. 3. I know I'll get a few zings from our HASL fans for this one, but almost every other board finish is much less warp inducing than HASL, none of the other board finishes subject the fab to the kind of heat that HASL does. (I gotta get ready for work, I'll come back up later to give a few suggestions about minimizing warp during assembly) -Steve Gregory-

reply »

If your board is single sided you should convey it on a mesh belt conveyor. Unfortunately most assemblies these days are double sided which makes it difficult to convey on the mesh belt. Plus with the increasing population on SM boards the weight is ever increasing. This is why we at ETS developed our Product Support Conveyor for use with the EDGE GRABBER (tm) Conveyor. All ETS Reflow ovens can be equiped with one of two styles of product support conveyors. The first utilizes a #25 chain which is .35" wide enclosed into a rail. The chain moves with the circuit board at the same speed as the conveyor. The rail is adjustable both in height and the location across the width of the process cavity. The second style of product support conveyor is a stainless steel rail which has a series of thin, free rolling disks .035" thick. Again the rail is computer programmable for the x and y axis. This system offers the most flexibility in processing as it requires the thinnest "lane" on the bottom side of the board. We have also conducted a study about the effects of a center support during reflow. I would be happy to share that study with you if you provide me your fax number, or the article can also be found in the September 1990 issue of SMT Magazine. For more information contact Brian Stumm at 509-483-0900 or by e-mail at ets@eznet.com

| Hi Anthony! | Earl's right, things can get rather involved when you start discussing ALL the things that affect whether or not a board is going to warp. | The way I look at the issue is try to separate the warp issue into two catagories; 1. The board being flat prior to SMT, and 2. The board being flat after assembly. | I think between the two areas, it's easier to deal with the first issue than it is with the second. Obviously, the main enemy of board flatness is heat, be it in the PCB fabrication process, or the assembly process. | Just a few things to think about to address minimizing warping in the fabrication process: | 1. There are many different laminate materials that are more stable than FR4, but it comes at a price. Almost every other material is going to be more expensive than FR4. | 2. Balancing the copper content in the board, distributing it as equally as possible within each layer, and throughout the overall board. If the design allows, using cross-hatch types of patterns, or I've even seen circles distributed in a grid-pattern left in areas where there are no traces or planes to balance the copper. By doing this you not only improve the temperature stability of the board, but it makes things easier for your fab vendor to etch. | 3. I know I'll get a few zings from our HASL fans for this one, but almost every other board finish is much less warp inducing than HASL, none of the other board finishes subject the fab to the kind of heat that HASL does. | (I gotta get ready for work, I'll come back up later to give a few suggestions about minimizing warp during assembly) | -Steve Gregory-

reply »

I really appreciate everybody taking the time to help me out with this problem. I'm going to run your input by the design engineer, after that I may be taking you up on that offer to contact you Earl. Again; thank you! Anthony | If your board is single sided you should convey it on a mesh belt conveyor. Unfortunately most assemblies these days are double sided which makes it difficult to convey on the mesh belt. Plus with the increasing population on SM boards the weight is ever increasing. This is why we at ETS developed our Product Support Conveyor for use with the EDGE GRABBER (tm) Conveyor. All ETS Reflow ovens can be equiped with one of two styles of product support conveyors. The first utilizes a #25 chain which is .35" wide enclosed into a rail. The chain moves with the circuit board at the same speed as the conveyor. The rail is adjustable both in height and the location across the width of the process cavity. The second style of product support conveyor is a stainless steel rail which has a series of thin, free rolling disks .035" thick. Again the rail is computer programmable for the x and y axis. This system offers the most flexibility in processing as it requires the thinnest "lane" on the bottom side of the board. We have also conducted a study about the effects of a center support during reflow. I would be happy to share that study with you if you provide me your fax number, or the article can also be found in the September 1990 issue of SMT Magazine. For more information contact Brian Stumm at 509-483-0900 or by e-mail at ets@eznet.com | | | Hi Anthony! | | Earl's right, things can get rather involved when you start discussing ALL the things that affect whether or not a board is going to warp. | | The way I look at the issue is try to separate the warp issue into two catagories; 1. The board being flat prior to SMT, and 2. The board being flat after assembly. | | I think between the two areas, it's easier to deal with the first issue than it is with the second. Obviously, the main enemy of board flatness is heat, be it in the PCB fabrication process, or the assembly process. | | Just a few things to think about to address minimizing warping in the fabrication process: | | 1. There are many different laminate materials that are more stable than FR4, but it comes at a price. Almost every other material is going to be more expensive than FR4. | | 2. Balancing the copper content in the board, distributing it as equally as possible within each layer, and throughout the overall board. If the design allows, using cross-hatch types of patterns, or I've even seen circles distributed in a grid-pattern left in areas where there are no traces or planes to balance the copper. By doing this you not only improve the temperature stability of the board, but it makes things easier for your fab vendor to etch. | | 3. I know I'll get a few zings from our HASL fans for this one, but almost every other board finish is much less warp inducing than HASL, none of the other board finishes subject the fab to the kind of heat that HASL does. | | (I gotta get ready for work, I'll come back up later to give a few suggestions about minimizing warp during assembly) | | -Steve Gregory-

reply »

| Does anybody know if a contraption exists that will minimize a PCB's tendency to become slightly bent after reflow? We have a new product that we're planning to assemble in a couple of months. The board is 9.7" by 13.4" and it is imperitive that we not have any portion of it sag beyond one tenth of an inch from it's plane. Any help that can be offered on this topic would be greatly appreciated. Thank you! | Anthony Boards sag with heat, they relax with heat, the vendor you get your board from could be the affect of board warps (his poor process)... If you have a specific spec for this, restrain the board. Invest in a good set of pallets that keep the board stiff in the process. Check your NEPCON guide for a number of qualified vendors to help you. I hope this is an exception rather than a rule for you! Cheers, Scott

reply »

Contact New Era Technology, Inc. @ (603)669-7971 for custom pallets and titanium stiffiners and masking.

reply »