Inspecting QFNs

Steve Scheiber, Contributing Technical Editor -- Test & Measurement World, 2/26/2007 11:13:00 AM

The constant striving to reduce the size of printed-circuit boards (PCBs) and the devices they contain has forced device designers to become more creative with how they use space. Quad flat packs (QFPs), once recognized as the epitome of the balance between size and performance, have given way to QFNs—same principle, but without the lead.

Inspecting QFNs (quad flat packs no lead) on boards presents something of a challenge. Solder joint shapes are less consistent than with leaded devices, and the use of lower-wetting lead-free solder, which has a higher melting point, makes the problem worse. Although conventional x-ray inspection hardware can see the device-board connections, system software has difficulty diffferentiating good boards from bad. Jeremy Jessen, technical marketing engineer at Agilent Technologies in Loveland, CO, said that the solution lies with providing new tools in the software.

Jessen put it this way: "QFNs come in several forms. Some have their pads right at the corners where they come in contact with the nodes on the board, with metal along the bottom and up the side. Others confine their connections to the bottom of the device. They have no metal at all on the side. Some 'bottom-only' devices feature longer pads for situations where a manufacturer has replaced a QFP with a QFN without changing the board layout. The type of joint necessary to make a reliable connection between a QFN and the board is different from joints on QFPs and varies in shape from one board to the next.

"Conventional x-ray inspection looks at the structure of the solder joint. It assumes that the joint always includes a heel, a center, and a toe or—for BGAs—a round ball. QFN joints, on the other hand, sometimes have a toe, sometimes they don't. You can't perform the same kinds of heel-toe comparisons to decide whether the joint should pass or fail. Instead, you have to evaluate the joint as a whole. That approach makes the decision at the software level much more complicated. Also, the software has to accommodate the greater board-to-board variation, especially using lead-free solder.

"In examining those variations, we discovered that the amount of 'slop' within a single facility tends to be much lower than when you compare it with what happens somewhere else. So, we created a tool that allows customers to perform their own analysis and 'tune' pass/fail criteria individually for each line. Without that capability, the inspection step will generate a lot of false failures. When we implemented the technique, the number of false failures started out quite high but fell dramatically as manufacturers matched the expectations in the software with the realities on the line."

Manufacturers cannot tolerate defects getting to customers, nor can they afford failing boards that should pass. As device designers find new ways to connect with board substrates in the never-ending quest to save space, test and inspection techniques have to better understand not only what to look for on the board but also how to evaluate what they see.