Ironwood Electronics: It's all about Choice

CSR: Since we're talking about the evolution of semiconductor devices, let's start by talking about Ironwood's own evolution. Can you provide some background?

Pal: Ironwood Electronics got its start 25 years ago as an adapter manufacturing company that provided prototype adapters to test engineers. In the beginning, they bought sockets from other vendors. But with the emergence of high speed applications requiring shorter lead times than was possible from outsourced vendors, the company decided to start making its own sockets. We began with elastomer sockets put on an adapter, and slowly ventured into manufacturing pogo pin systems. Our portfolio has over six different connector technologies including gold plated wire elastomer, standard pogo pin, diamond particle interconnect, stamped or etched, pogo pins, and silver particle interconnect. We're focusing on solutions for high-speed applications. As engineers are developing newer generation devices, we work with them on both the front-end engineering/characterization test as well as on the back end production/ATE test.

CSR: There are a lot of burn-in test socket companies out there. What would say differentiates Ironwood from its competitors?

Pal: There are a few key points that separate us from our competitors. Now-a-days, burn-in test process is not straight forward as AC and DC parameter tests conducted while burning-in. A major difference is that we're able to provide the right technology for the right application, whether we're talking about the pitch, the solution integration, or even different types of contactors. Other companies may only offer pogo pin, for example, but that doesn't work for all applications. Even if the contactor is mechanically robust and good, if it doesn't meet electrical requirements, it's not the right solution. We believe we should be able to accommodate different needs of customers, specifically on front-end development and characterization process.

Another differentiator for us is lead time. Because all our manufacturing is done in-house, we can custom design a socket with only a three week lead time.

Our focus on front-end development also sets us apart from other socket companies. Thanks to wafer level packaging technologies, the lines are blurred between what is front- and what is back-end. With WLP when the package is done, it's still on the wafer in the front-end environment. We're dealing with a new class of chips that feature next generation frequency and communication and faster processors. There's a need for higher end socket technology to verify these chips that are still in the front-end.

CSR: Ironwood Electronics seems to come out with a lot of new technologies fairly regularly. What do you see as the latest challenges for burn-in test sockets?

Pal: Our biggest challenge lately has been with force vs. pitch. In a typical compression socket, you need to have a certain amount of force to make contact. But with shrinking pitches, the solder balls are smaller and you can't use high force because it deforms the solder balls and causes sticking issues at high temperature. However, if you reduce the force, you risk not making adequate contact. Ironwood is working to develop a product that produces low force for best electrical contact. Additionally, with WLP high temperature testing, you don't want a lot of force because the packages are so delicate. In this case, we're looking for best electrical contact without fracturing the wafer. We're working on a compression mount contact technology that applies combing mechanism. The socket structure and material plays major role to address these limitations.

CSR: In addition to a fairly robust product line, IE also offers custom work. Is there a big call for custom sockets?

Pal: Every day; we design almost 5 custom sockets per week. We also offer about 1000 standard sockets. Applications are so different, that's why we have multiple contact technologies in our product portfolio. Every day, people are coming up with a new application or gadget. The requirements for all of them are completely different and needs a custom solution. One request was for a socket with only three walls to protect critical optical components. They wanted to know if we could make a 3 sided socket that works. Another big thing is 3D stacks. Each one is a different combination of technologies - a two level, or 3 or 4 level stack up. Each requirement turns into a custom socket design. About 50% of our work is custom.

CSR: Since BiTS Workshop is underway the day this interview blasts, maybe you could give readers who will be attending a preview of what Ironwood Electronics will be showcasing?

Pal: We'll be showcasing our ability to integrate the six different technologies into a small unique socket footprint. The main example will be a technology that can help test QFN-type devices, where the biggest issue is matte tin and nickel plated gold contact. Nickel plated contact is so hard that it deforms pin. Tin from the device migrates to the contact tip and is difficult to clean, so the pin needs to be changed. We'll also be addressing how to avoid biggest hurdles customers see in the custom area to help to satisfy production yield with no false failures. We'll also be showcasing our low-cost molded sockets using stamped pogo pins specifically addressing burn-in requirements. If a customer needs a custom configuration, these low-cost molded socket with stamped spring pins allow us to turn things around in 3 weeks as opposed to 12 weeks from a competitor. A lot of memory devices are exponentially growing, so Ironwood has developed a set of sockets specific to each memory so that we can offer them as standard. We'll also be introducing multi-level sockets with memory probe compatible with Agilent's scope/analyzer.

CSR: In general, how is 2011 shaping up for Ironwood?

Pal: We started out with a big bang (lots of orders) so one of the challenges is to try and meet delivery of those with the shortest lead times. 3D packages have a lot of complications on their own; each application needs a custom socket because of different requirements for the top and bottom packages. Also, a different contact technology is needed for testing TSVs. Some sort of probe will be required to access the signal between two packages. It's an adaptor technology which involves integration of socket and adaptor technology. Since we started as an adaptor company and continually evolved our technology capabilities, Ironwood is ahead of the pack with multiple options. We have existing technology to test stacked multi-die, WLP solder bumps with 400µm pitch on the top and bottom. We're currently working on a project to support test challenges involved in the engineering stage.