Bryan Ingram is Senior Vice President and General Manager of the Wireless Semiconductor Division (WSD) for Avago Technologies. He joined Hewlett-Packard in 1990 after holding a number of engineering and marketing positions at Westinghouse Electric Corporation. Prior to the spin-off of Agilent Technologies from HP in 1999, Ingram headed up the strategic marketing effort and the marketing department for communications components. In 2000, he accepted the post of Vice President and General Manager of WSD. Ingram holds a bachelor’s of science degree in electrical engineering and communications systems from the University of Illinois and an MSEE from Johns Hopkins University.
For over three decades, Avago was part of HP and the Semiconductor Division of Agilent Technologies before becoming an independent company. Today, Avago is one of the world’s largest privately-owned semiconductor businesses, supplying RF, analog and mixed-signal semiconductors to four end-markets including industrial/automotive electronics, wired infrastructure, wireless communications, and consumer/computer peripherals. Avago RF/microwave components include power and low noise amplifiers, BAW and FBAR filters and multiplexers, control products (switches, attenuators, mixers) and discrete devices (transistors, diodes) targeting wireless infrastructure, wireless data, (GPS/WLAN/WiMax), handset, digital TV, VSAT and datalink applications.
MWJ: Agilent is such a strong brand in the microwave industry. Do most of your customers know of your long history or has it been a challenge to establish the Avago brand after the split?
BI: Our customers do know of our history. However industry analysts and people not directly involved in the industry may not recognize our name. Sometimes, an introduction that we were formerly the semiconductor arm of Hewlett Packard, then Agilent Technologies is what it takes to place things in context.
MWJ: What have been some of the benefits and challenges of parting ways with a large and established company?
BI: On the benefits side, Avago is a pure-play semiconductor company and we have aligned the processes and procedures of the company accordingly. When we were HP’s Semiconductor Product Group (SPG), we were a bit of a hidden “billion dollar secret” to those who thought of HP as a computer company. In Agilent, we were the odd fellow among three other product lines that all focused on test systems. Therefore, being in the Avago spinout is very refreshing in this sense. About the only downside I can think of this that you do lose some of the infrastructure support available only to the very largest corporations.
MWJ: Given the company’s Agilent background, I would think Avago would sell heavily into the test and measurement equipment market. Is that the case?
BI: We do have success in the test and measurement market, some of it ascribable to our roots. However our focus is on other, much larger, segments: handsets, data systems, infrastructure.
MWJ: I read that 60% of Avago’s employees were overseas. Does this reflect where your customer base is located?
BI: Certainly Avago is an international company with an international customer base. For instance, WSD has design centers in San Jose, Seoul, Penang, and Munich, sales and technical support worldwide, and manufacturing facilities in Malaysia and Colorado. We have significant sales into Asia Pacific as well as Europe and the Americas. In many cases our customers are also distributed, with design in one region of the world and manufacturing in another. We do place our sales and support teams near our customers and make sure our customers have appropriate input into our design process.
MWJ: The company has such a diverse portfolio with approximately 7000 products. Among the RF components, which product lines have the greatest advantage in the market and why?
BI: On the power side, our line of CoolPAM PAs holds pride of place, having captured significant market share for linear standards (CDMA, W-CDMA) due to their ability to offer longer battery life. Our FBAR duplexers have also been extremely successful, especially in bands such as PCS where narrow duplex spacing necessitates the use of high Q filtering. We also do well with LNAs in data systems (including WiFi and GPS), due to superior performance, usually in terms of linearity and noise figure.
MWJ: The company has sold an impressive number of FBAR filters and duplexers. Was Avago among the first to develop this technology when it was part of Agilent? Is Avago a leader in this technology?
BI: FBAR is a type of Bulk Acoustic Wave (BAW) Filter, a technology that has deep roots. Under the Agilent name, we were the first to successfully commercialize miniature Bulk Wave filtering to a broad customer base, starting with our PCS duplexer production ramp in 2001. We remain the clear leader in FBAR technology, and we can also claim overall leadership in bulk-wave technology. Additionally, we now have well over 200 patents on FBAR and BAW filter technology including several fundamental patents that go well back into the 1990s. We have excellent worldwide coverage with these patents and are continuing to bolster our existing patents with new filings.
MWJ: Avago bought the Infineon’s Bulk Acoustic Wave (BAW) filter business. How do FBAR and BAW filters complement each other from an application perspective?
BI: The Infineon process uses solidly mounted resonator (SMR) technology; film bulk acoustic resonator (FBAR) technology uses a suspended resonator membrane. Both are acoustic bulk wave filtering technologies, though unfortunately the industry seems to have ascribed the “BAW” label to SMR variant. The FBAR process generally offers higher Q, hence steeper filters. The SMR process has some advantages in terms of complex circuit topologies and temperature compensation. Like most engineering comparisons, there are positives and negatives to both sides.
MWJ: What was the strategic thinking behind this acquisition?
BI: Avago is coming off an extremely successful year, and saw a need to grow R&D investment especially in the filter product space. Infineon on the other hand was moving away from a focus on BAW technology. This gave us an excellent opportunity to get an intact design team with real world experience in bulk wave filter design – something not common in our industry. This team is now starting to do designs in FBAR, with the initial products having a European flavor to their applications. We are the only one in the industry capable of benchmarking these two technologies, we ship both in volume, and we discovered that FBAR has overall better performance, lower cost, and better manufacturability as compared to the SMR technology. Therefore, our efforts are focused upon using FBAR for future designs.
MWJ: Earlier this year Avago Technologies released a new family of power amplifiers based on the fifth generation of your CoolPAM technology. What is CoolPAM technology?
BI: CoolPAM is a special technique that allows power amplifiers to operate more efficiently at lower powers. Most linear 3G PAs spend most of their time operating well below the peak power rating of the phone, so efficient operation at lower power levels while maintaining adequate linearity is key to long battery life. The basis of the technology is in intelligently rerouting signals in the power amplifier as a function of output power, establishing much more efficient load lines at lower powers. We apply this technology to both narrow CDMA and UMTS; the latest generation of parts operates at 4 mA of current at lower power ranges.
MWJ: What are some of the biggest challenges for handset PAs these days and what does a company like Avago need to do to compete and win?
Avago focuses mainly on PAs for linear 3G and 4G standards – CDMA, W-CDMA, and, in the future, LTE. In this arena the challenge is to increase the number of bands and modes supported while continuously improving efficiency and lowering costs. Our business model is to employ a mix of proprietary internal technology and outsourced, more commoditized technologies to create cost effective products that allow our customers to differentiate, usually through performance advantages. This approach has allowed us to succeed in CDMA and W-CDMA while competitors fought over the larger but less differentiated GSM PA market.
MWJ: Avago offers power amplifiers with low quiescent current levels, offering greater efficiency and longer battery life. Is this unique to the PA market or is it becoming a pre-requisite for all RF power amps to be competitive?
BI: Highly efficient PAs are valued in virtually all portable applications. Exactly what “highly efficient” means depends on the modulation scheme employed and the details of the application. For mobile (hence battery operated) linear applications where large dynamic range is part of the need, low quiescent bias is a must. These systems include CDMA and W-CDMA, and Avago does well in this segment due to our CoolPAM technology. For saturated systems quiescent current doesn’t matter, but efficiency at high power does. This is more the case in GSM systems. For fixed systems (infrastructure, embedded modules, some data systems) quiescent bias might not be an issue per se, but more efficient operation tends to mean cooler operation, so less cost in fans and heatsinking, plus lower energy costs and “Greener” operation. So I would generally agree that efficient power amplifier operation has become a pervasive industry requirement.
MWJ: Are these devices based on GaAs HBT or E-pHEMT technology?
BI: CoolPAM PAs have been based on outsourced HBT technology. We also have a successful line of PAs that are based on EpHEMT that we fabricate ourselves. I suppose that makes the answer “yes” to both options.
MWJ: A lot of RF semiconductor companies are making a play for the wireless backhaul market. I see Avago has a number of millimeter-wave devices. Is this a market you are heavily involved in?
BI: Avago does make a fairly broad range of products that are applicable to millimeter wave wireless backhaul. This is a segment we have a fairly long history in, as the technology base has a lot in common with the instrumentation market you asked about earlier.
MWJ: What’s your outlook for this market?
BI: It’s common to find analyst reports pointing to the need for more backhaul as the uptake of mobile data is leading to exponential growth in bytes transported. Wireless backhaul based on radios operating at millimeter wave frequencies gets cited as a part of the solution. Exactly how much, when, at what price point is of course something we would all like to know.
MWJ: It seems that a lot of design houses are going fabless. Avago has a foundry in Fort Collins, Colorado, but also outsources chip fabrication to Win Semiconductor and Triquint. What’s the advantage of maintaining your own fab and do you see Avago going fabless someday?
BI: Avago’s business model is to internally source differentiating technologies, while using foundries for more “commoditized” needs. Some of our other competitors similarly run a mix of their own internally fabricated wafers and outsourced wafers. This allows some flexibility in response to peaks in market demand, while avoiding excessive overhead during any “down” market cycles. I don’t see Avago dispensing with the internal fabs that provide us with much of our differentiation, but neither do I see us returning to an entirely in-house-sourced model. Control and flexibility are both important to success.
MWJ: Avago’s near obsession with micro-miniaturization has led to some impressive engineering especially with your new WaferCap chip scale packaging. How dramatic is the space saving and which applications are likely to benefit most?
BI: The WaferCap packaging effort is actually a logical extension of the MicroCap packaging technique we have been using in our FBAR product line for the last 5 years or so. In our filtering product we go on to combine multiple die on a PC board to create the final product, and the primary advantages of the technology are flexibility of design and cost. In the recent VMMK line of products, we have applied the technology to GaAs and shrunk all the circuitry onto the die to eliminate the need for any carrier. While the footprint reduction can be impressive, the lower product height and the simplicity of assembly – the devices can be treated like SMT passives as far as assembly is concerned – are also very strong selling points. Virtually any market area that values size, cost, ease of assembly, and improved performance (a result of reduced package parasitics) can benefit from this technology, so to this end we will be offering a broad range of general purpose RF building blocks in this packaging format.
MWJ: Your RF semiconductor devices serve many different fields, is there good synergy between your various R&D groups or do they operate autonomously?
BI: Our geographically dispersed RF technology centers do have different areas of market focus, and have different technology preferences and different areas of deep expertise as well. That said, we make a point of sharing common tools, and communicate best practices regularly to gain synergies. We also have identical NPI processes employed at each site so that product development and communication is efficient. Additionally, we have scheduled and frequent face-to-face meetings between different sections as a normal part of our business practice. So while most centers are capable of operating autonomously, we have a holistic approach to our execution.
MWJ: The Agilent culture (doing things the HP way) produced a lot of innovation over the years and Avago certainly has produced its share of first-to-market and unique products. How does Avago approach innovation?
BI: With gusto! Innovation is certainly a big part of our cultural roots, it’s a major component of the success of Avago and WSD today, and it is enmeshed with our self image and way of doing business. Our business model depends on our ability to innovate and thereby provide differentiation to customers. That’s something we foster and keep vibrant.
MWJ: What do you consider to be Avago’s greatest strengths?
BI: Our employees. From them comes our innovation, our execution and our success.