Bare Die on Flex Print E-mail
Written by Dale Wesselmann   
Friday, 29 March 2013 22:20

Has the time come for flexible printed circuits integration?

Today’s market is a challenging one for flexible printed circuit fabricators and assemblers. Years of increasing device functionality, smaller form factors and lower price points leave little room to imagine where the next cost reduction will be found. One bright area might be the integration of active components directly on the FPC.

One way of integrating active components is to attach bare die directly on the flex circuit. While companies are chasing embedded component solutions, I think broader use of direct-chip attach technologies offers FPC suppliers value-added advantages with known processes. The cost and value-add advantages of removing the chip packaging can be boiled down to three things: removing the packaging and redistribution layers, signal-performance improvement gains from removing the chip packaging, and reducing the total assembly weight and height, which enables valuable design choices for OEMs.

Let’s take a look at how these three potential cost-saving advantages of flex-circuit integration can possibly increase the value of FPCA.

Remove the package. The most obvious cost savings would come from removing the die packaging. Die packaging consists of the redistribution layer, overmolding compounds, and solder balls or tinned leads, depending on the package style. Figure 1 shows two basic package constructions, one with wire bonding and the other with flip-chip attach. Both packages use a redistribution layer to fan out the fine pitch I/O of the die into a configuration that is more SMT-friendly.

[Ed.: To enlarge the figure, right-click on it, then click View Image, then left-click on the figure.]



For the purposes of this discussion, I’ve selected the cost breakdown of a range of IC devices. As Table 1 shows, the die itself comprises only 38% to 45% of the total component cost. That gives us over 50% of the component cost to eliminate or reduce materials, then integrate the cost of design, assembly, and test into the fabricators’ cost structure.

[Ed.: To enlarge the figure, right-click on it, then click View Image, then left-click on the figure.]



Improve signal performance. A second cost savings opportunity is actually related to improved performance. Removing the redistribution layer and a set of interconnects from the packaging results in an improvement in signal quality. As the data-rate increases within the device, there is a perceived need to move toward more expensive, high-performance dielectric materials for FPC that have better high-speed performance characteristics. By leveraging the performance gain from removing the redistribution layer, high-cost materials like LCP or Teflon are no longer needed to achieve signal-quality requirements.

Depending on part size, the potential material cost savings varies, but viewed from an annual bulk-material spend, the differences can be drastic. As a reference, if we assume a scale to purchase and process base substrate, and polyimide is base cost of 1, then LCP would be 3 and Teflon would be 5. Over the length of a production run, the cost-premium of these materials is tremendous. The cost-benefit of continuing to use polyimide without overall performance impacts is potentially substantial.

Reduce size and weight. By removing the packaging and working with the bare die, a reduction in both weight and area can be achieved. A bare silicon die can account for as low as 5% of the total package weight. When it comes to overall length and width of a component, a bare die consumes down to 15% of a packaged device.

What’s the hold up? After having spouted paragraphs of publicly available reasons why attaching silicon die directly can potentially save money and increase value, why do we not see more integration of bare die on flex? The first reason is only recently has a high premium been placed on “thin and light.” Today, value is measured in tenths of millimeters when it comes to overall device-thickness reduction. Another reason is that traditional chip packaging has been very specialized in processing of the substrates. Attaching directly to die requires very fine lines, stacked microvias, and other tight feature tolerances, which normally would not fall into the high-volume capabilities of the average FPC manufacturer. As the capability gap between IC packaging and the FPC/A companies narrows, these density challenges will become less of an obstacle.

The last, most pragmatic reason is cost of quality. The overall cost of the die can be substantially bigger than the price of bare flex, making the potential costs of scrap at industry-standard yields less trivial. As the fabrication process performance continues to increase, this challenge may be overcome as well.

Dale Wesselmann is a product marketing manager at MFLEX (mflex.com); This e-mail address is being protected from spambots. You need JavaScript enabled to view it . His column runs bimonthly.

Last Updated on Tuesday, 02 April 2013 04:20
 

Search

Search

Login

CB Login

Language

Language

English French German Italian Portuguese Russian Spanish
 


Printed Circuit Design & Fab Magazine on Facebook