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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.

PCMCIA QFP Process Issue



PCMCIA QFP Process Issue | 21 May, 2000

Currently manufacturing double-sided PCMCIA cards with a nickel/gold finish, which are 0.0235thou in thickness, using surface mount carriers. Problem we are seeing is when building the densely populated side on the second pass we are seeing lifting of the qfp's on the corners,but this problem is so intermittent I just can't put my finger on anything at the moment.The qfp's in question are 0.020 thou in pitch with standard stencil reductions. The printer is correctly set-up using a Indium paste,the reflow profile has been verified and we are using Nitrogen. On inspection the solder paste seems to have flowed and wetted really well ,but the corners still look lifted and a couple of leads just sit on top of the reflowed joints. I'm not ruling anything out at the moment, but am I looking at some kind of a thermal issue or a set-up issue? Also how critcal is it after the first pass that the boards are layed dowm flat and allowed to cool after reflow to ensure they don't bow or warp? could this have an impact ?

Any advice would be gratefully appreciated.

reply »


Re: PCMCIA QFP Process Issue | 22 May, 2000

Hi Sal,

I don�t have the solution to your problems but my first thought is warpage. I�ve seen similar things caused by exessive warpage of PCBs 1,6mm thick sometimes and with 0,8mm and T-SOP 50% showed this symptom. Prebaking and a fixture to counteract warpage during reflow might help in that case. Finding the real cause will lead to solution.

Just one idea


reply »


Re: PCMCIA QFP Process Issue | 23 May, 2000

Sal: You and Wolfgang are spot on, but let's see if we can blow this out a bit.

Before proceding to rattle on ... � PC Card. A credit card-size computer peripheral that add memory, mass storage, and I/O capabilities to computers in a rugged, compact form factor. � PCMCIA (Personal Computer Memory Card International Association). A non-profit trade association and standards body that promotes PC Card technology, along with Miniature Card and SmartMedia cards by defining technical standards and educating the market.

In the past, cards were known as "PCMCIA Cards", but the industry now refers to products based on the technology as "PC Cards," "PC Card Hosts," and "PC Card Software," and refers to the association as PCMCIA.

Now, on with the rattling ...

Sounds like, during reflow one part, either the board or the QFP, becomes a maxed-out potato(e) chip and the other is a minnied-out potato(e) chip. Since this appears to be putting strain on the solder connections at the corners of the QFP, I�m not real excited about the potential impact on long term reliability of your product from the sounds of things.

I�m all farklempt. Discuss among yourselves (as Linda Richman used to say): � Be wary of using components with Alloy 42 leads on thin substrates (ie, PCMCIA, flex), because of the lack of compliance of the leads and different Tce. Some people believe Alloy 42 leads are a reliability problem regardless of the substrate. It�s easy to determine if leads are Alloy 42 (or Kovar), because they�re magnetic. (Copper leads are not magnetic, as you know.) An additional issue is poor solderability of Alloy 42, but this doesn�t seem to be an issue for you � knock on wood. � You didn�t mention carriers (pallets). We use carriers with all thin substrates. Many people, like the disk drive industry, don�t use carriers. OK, we�re wimps. But, we get plugged designs. Izit we get plugged designs and that makes us wimps �cause we�re forced to put-up that bull funky? ... Orzit we�re wimps, so they give us plugged designs? ... What ever. Although the SMTnet archives list carrier suppliers, the transition from fabricating FR-4 wave solder carriers to PC Card and flex carriers is not trivial. � Try panelizing the boards "side-by-side" in a 4-up panel. This improves machine throughput and reduces the scrap material. Additionally, the twisting of each board tends to cancel-out the twisting of the other. � Consider adding outside tabs along the length of the panel (~0.16" - 0.20"). Most pick/place equipment need at least 0.16" for conveyors to move the panel from one site to another. These should be routed off after assembly. Also, global fiducials can be placed on the tabs for the screen printer without taking up precious PCB real estate. This will improve panel to panel rigidity. � Leave at least 3 "break-away" tabs between each board to provide additional support and minimize panel warping. � Consider using the mesh-belt on the reflow oven, rather than the chain-rail system. This reduces panel drooping and warping during reflow. With the mesh-belt, a support carrier is unnecessary, assuming minimal board warping and similar Tce of components. � Think about slowing-down the drive on your reflow to provide more time for temperature change to affect all components equally.

Ta Dave F

reply »

Chris McDonald


Re: PCMCIA QFP Process Issue | 24 May, 2000

I have found that if you have a custom carrier with support between pcbs on both SMT reflows (If you palletize) it will eliminate warping. Also Prebaking PCBS with wieght on the top of them helps too. Also look at the cooldown of your oven profile, you want the pcb cool when comming out of the oven. Your cooldown could be warping the PCB.

Chris Mc

reply »

Equipment Auction Automotive Electronics Supplier - Closure of Tier-One SMT Dvision: (10) ASM & Universal SMT Lines & Feeders Equipment as-new-as 2019! Dek | Koh Young | Speedline | Vitronics | Viscom