From a large lot, 20 populated boards with components of varied heights on both sides are warped. What methods can they use to flatten them? Jim Hall and Phil Zarrow, The Assembly Brothers, discuss this situation and offer their own suggestions. Board Talk
Board Talk is presented by Phil Zarrow and Jim Hall of ITM Consulting.
Process Troubleshooting, Failure Analysis, Process Audits, Process Set-up CEM Selection/Qualification, SMT Training/Seminars, Legal Disputes
Phil Zarrow
With over 35 years experience in PCB assembly, Phil is one of the leading experts in SMT process failure analysis.
He has vast experience in SMT equipment, materials and processes.
Jim Hall
A Lean Six-Sigma Master Blackbelt, Jim has a wealth of knowledge in soldering, thermal technology, equipment and process basics.
He is a pioneer in the science of reflow.
Transcript
Phil
And welcome to Board Talk with Jim Hall and Phil Zarrow of ITM Consulting, also known as the Assembly Brothers. Place and pick. In our never-ending quest to help people solve their assembly processes, situations and hopefully not to confuse them.
Today we have a question from E.J. E.J. says from a large lot, we have 20 populated boards with components on both sides that are warped. Components are varied height. What methods can we use to flatten them?
Jim
Syntax issues here, Phil. Is he saying the components are warped or the boards are warped? If it was components, it would probably have to big things like BGAs. Frankly they would each have to be dealt with individually and I don’t know any way to flatten components that have been warped. I suspect that for some reason 20 of the PCBs during the double-sided assembly process, the boards have become warped.
Phil
The first question it begs, as we put on our FA hats, is 20 populated boards in the process, what changed? What happened here? Was it the way the boards were stored or handled? Was it your reflow oven? Did something change there? Somebody lacking on support? That is one thing. I don’t know if there is a way with some fixturing that we can iron out these boards.
Jim
Going back to the causes, typically you think about warpage coming from inadequate board support, improper reflow profiles causing warpage during the process. But 20 boards, the thought I had Phil was one package of boards that wasn’t packaged properly and came in with warpage, the bare PCB coming in. I’m thinking 20, maybe that is the size of a package of boards that are coming in from your vendor, from your PCB fab.
Maybe one package was damaged, got moisture in it or something else so that you got 20 raw boards coming in that either had warpage or had some variation in them such as moisture absorption which caused them then to warp during the assembly process.
Phil
Yeah. There is a lot of things we don’t know. We aren’t even sure if he is talking about boards or components that are warped either. But running on that assumption.
Jim
You have to accept that fact that it may be impossible to flatten out. If 20 of these PCBs are warped, it may be impossible to flatten them out and still maintain the reliability of your product.
To straighten them out, as Phil said, you can envision some sort of fixture much like a board support fixture that you would use to prevent warpage with pins that would contact the board in between, the components that are mounted on both sides, with fixtures that could put pressure on them to flatten out the board and leave them for some time, hoping that would flatten out or reduce the warpage.
That may be possible. In any case, if you try to do that you have to be conscious that you have solidified solder joints in that warped configuration. It may be the reality that if you try to straighten you are going to crack some of those solder joints. It may not be possible to salvage these boards if you require a certain flatness in the final assembly.
Phil
I wouldn’t be surprised if some of our listeners chime in with some methodologies, homeopathic or otherwise that they have used to de-warp boards. It will be interesting to see on this one. It is a tough one. I just wish we had more information.
Jim
That is a great lead in, Phil. All of you faithful followers of Board Talk, when we do this initially and these are presented on Circuitnet it is just us, our ideas good or bad. But over time people send in comments. We have a lot of very conscientious listeners who share their experiences.
This is a good case where someone may have had this problem, might have come up with a really good solution that doesn’t come to our minds right at the moment. E.J., or other people with similar problems, wait a week or a couple weeks and go back online and pull up this session on the archives and see if anyone has posted a suggestion. There have been in the past many good suggestions. Corrections, for mistakes we made and so forth by our listeners. We thank you all for your active participation.
Phil
And on that note, you have been listening to Board Talk with Phil and Jim, the Assembly Brothers. If you go forth, no matter how warped you might be, please don’t solder like my brother.
Jim
And don’t solder like my brother.
Comments
Here is the military procedure to restore a warped PCB Diagram the CCA features by noting the component locations, types, values, orientation, and part numbers before removing any components or conductors from the CCA. Note: Technical manuals or manufacturer’s specification drawings usually provide this information. If drawings are not available, make a sketch or take a digital photograph of the assembly showing component locations, conductor locations, and CCA markings.
Remove any conformal coating from the repair area using one of the Conformal Coating Removal Methods in WP 006 00 (paragraph 06-5.4). Remove the components necessary to provide access to the repair area for a bowed CCA or remove all components for a twisted CCA (WP 007 00 for Through Hole Components or WP 018 00 for Surface Mount Devices). Set the oven temperature to 257°F (125°C). Clamp the bowed edges between two rigid metallic devices.
Clamp the bowed and twisted CCA between two rigid metallic surfaces. C-ClampsFlat SurfaceLaminate. Place the clamped CCA into the oven. Bake the clamped CCA for one hour. After the one-hour bake cycle, turn off the oven leaving the clamped CCA inside. Note: This allows the CCA to slowly cool to room temperature improving stress relief.
After the oven and the clamped CCA have returned to room temperature, remove the clamped CCA from the oven. Remove the clamp(s) and the metallic surfaces. Inspect to determine if the warp has been reduced sufficiently to restore functionality. Repeat steps 7 through 13 until the warp has been reduced sufficiently to restore functionality. Clean the CCA with isopropyl alcohol and blot dry with a clean, lint-free tissue. Replace the components removed to facilitate this repair.
Frank Honyotski, STI Electronics Inc.
It is a given that reliability will be totally lost. The only way to approach a flat board result requires running the assembly through the re-flow oven while attached to a custom fixture that applies very light pressure to cause the PCB edges to relax to match the fixture.
The very best thing to do is scrap the assemblies. Otherwise you expend time and money to knowingly create a completely unreliable assembly.
Jaye Waas, Renkus-Heinz
I read your web site religiously every time I see it in my Email account. The issue on warped PCBs was especially interesting to me.
You can not unwarp/straighten a PCB after the soldering operation.Thin 4 or 5 mil traces will break. Stiffeners must be installed prior to soldering to keep the board flat. It does not matter whether it is a large board with heavy components or a panel of tiny boards. Warped boards with Smt components on 2 sides present a problem with the pick and place and solder paste machines. Stiffening is required prior to soldering. My Company, Compufab Inc. has provided PCB Stiffeners to electronic companies world wide for over 34 years with satisfactory results. Over 2 million stiffeners to date.
Check out the Compufab web site at www.compufab.com or contact me at sales.compufab.com.
The Compufab PCB stiffener is a low cost alternative to machined metal and G10 fiberglass parts.
Otto Steiner, Compufab Inc
It is not uncommon to have 1 warped board in a pack from the board supplier. The it can be in the middle of a pack say of 50 boards. It will be flat in the pack, but once removed will spring back to a warped board. Providing it does not come off of a conveyor it will get populated and go into the reflow oven.
From my experience, the board usually climbs up onto the chain during reflow. This means it is again not flat in the oven so it remains Warped.
Flattening a warped FR4 board is a simple process, however is complicated by the presents of components. First of all you need to know the Glass transition temperature (Tg)of the laminate. Typical temperature is 120C for non RoHS boards and 150C for RoHS that are reflowed and 170C for RoHS boards that are wave soldered.
Non populated boards you can clamp between 2 flat metal plates, place in an oven +/- 10C above the Tg temperature. This will allow the board to soften. Turn the oven off and the boards will retain the shape they are held in when passing back below the Tg.
Populated boards you can do the same, however I have used titanium stiffeners that used to be used on the leading and trailing edge of a board in a wave soled process when carriers are not used. All you really need to do is clamp all the sides of the board between 2 straight surfaces that will not warp themselves at high temperature. Again heat to 10C above Tg. hold there for a period time typically 20 to 30 min to allow the fixture to heat up, then allow to slowly cool below the Tg.
Once it falls below the Tg it will hold the shape it is in at that point.
Some important things to remember you will be stressing the components. very important is if there is non SMD electrolytic capacitors on the board the plastic will shrink or melt and they will definitely dry out as most of them do not tolerate temperatures over 105C. I would recommend removing and replacing later.
Moisture in a board from the manufacturing process will usually result in blisters, bubbles or delaminating during the reflow or wave process, as apposed to warping. Please also check the manufacturers specification on any connectors and critical type components.
Les Watts, Testerion
"It may be the reality that if you try to straighten you are going to crack some of those solder joints"
<< The customer should also be warned of the potential for cracking and latent failures of the mounted MLCCs (ceramic chip capacitors) if they try to straighten the warped PCBs >>
Jim Wright, NIC COMPONENTS CORP.
Thanks for your channel and improving my pcb knowledge. I think it might happen when there is a mixed-up layer-stack, if it is laid by hand.
Frank Rockenstiehl, KARL STORZ SE & Co. KG
I suspect E.J. has 20 CCAs that are warped, as this is an issue that would not show up until sometime after reflow. The probable cause is that the board designer attempted to get as many fabricated PWBs out of a large flat of laminate as possible, and thus most of the PWBs are oriented one way but a row of the boards were oriented 90 degrees from the rest to fit along one of the edges of the flat. Design guidelines for orienting the general long axis of the individual PWBs and the general copper features with the long axis of the fiberglass strand orientation were ignored, as is often the case with no issues. Some designs however are susceptible to severe warping during reflow when this is done, and that is how you end up with some quantity of boards with severe warping and others with none.
There are three ways to check this, the first is to simply ask the fabricator for the number of flats and the flat layout that were used to produce the entire lot. If the fabricator is not very forthcoming, then the second method is to compare the strand orientation of the warped PWBs with those that are not warped under magnification. The third is to determine if the total number of warped PWBs is a multiple of the total quantity in the lot. In this case, for example, if the vendor states that each flat has 20 PWBs, and 5 flats were used to produce a total of 100 PWBs, and the flat layout was a 4x4 array with 4 oriented at 90 degrees from the rest, well 4 times 5 is....BINGO!
Richard Stadem, Kongsberg Defence AS
Typically the warp stress only poses a risk to larger leadless ceramic components. Consider, for example, a 6" square pwb with a total deflection of .600" across the width of the CCA has a deviation per inch of .100”, much greater than the .075” per inch limit allowed by IPC guidelines. The amount of warp across the length of even a larger 1206 cap is then approximately .00084”, which is probably not putting much if any stress on the component when you consider that even SAC305 has a certain relaxation factor. The rest of the larger components other than any large leadless BTCs, etc, typically have compliant-leaded parts taking up the stress.
So even a larger BTC of 30mm X30mm is not going to see an awful lot of stress if the CCA itself were put into a clampdown window-pane style of fixture utilizing perhaps 6 adjustable DeStaCo clamps around the periphery with a center support. A rework process can be derived by using such a fixture and clamping it with minimal force initially, bake in the fixture at 105C for 4 hours, then add a little more pressure to the clamps and then spray-flux and re-reflow the CCA and allow to cool while still clamped in the fixture.
This assumes all components can go through the reflow a second time and the CCA is not conformal-coated. Granted, after reflow the PWB will still be exerting stress on the components from wanting to spring back to its original memory, but it is usually a very minimal amount, and if the CCA is then clamped into its cabinet fairly quickly the SJ stress will be greatly reduced and what remains will be gone after just a few days as the solder relaxes over time. For Sn63 solder, the Pb37 percentage adds much more of a SJ relaxation factor, typically within 24 hours. The actual stress levels can be measured using the methods within IPC-TM-650 for lateral stress/CTE measurement.
Keep in mind the above is a really-worst-case scenario. For class 3 CCAs I would not recommend this for cases of such extreme warpage, but for applications where the typical warpage is .075” per inch or less, there may be little or no effect on reliability even if NO rework was attempted, provided the CCA can be clamped into its final-build cabinet or housing without causing any clearance issues.
Theodore Kleever, Wallace Grommet and Bushing, Inc.
Never attempted with assembled, only with bare PCB's, but if you are able to stack metal sheets or another heat tolerant heavy enough material (scrap PCB's) with enough weight on top of the assembled board so you could flatten it without damaging it, you could try placing them in an oven at a temperature above the Tg of your laminate (if a leaded board maybe above 140C, if ROHS maybe above 170C but below the melting point of your solder) for an hour or so, maybe the epoxy resin of the laminate may flow enough to improve your situation. Caution regarding any temperature sensitive components or possible delamination of the PCB that may occur if you leave the board in for too long. Maybe another reader can comment on safe temperature/time setting or it might just be trial and error.
