WiMax.com: When was PicoChip founded?  Has the company's mission and technical approach changed over the years?

PicoChip was founded in Sept 2000.  The company was funded in June of 2001.  Our technical approach was consistent from the beginning:  to solve the "tough" problems of next generation wireless with a programmable platform.

Our CTO had worked for Vodafone, and then developed wireless base stations with Lucent.  In doing so he realized that the traditional way of designing application specific integrated circuits (ASIC) or using legacy digital signal processing (DSP) just wouldn't work as we moved to broadband - 3G and the like are just too hard.

What has changed is that our initial focus was 3G - specifically wideband code division multiple access (WCDMA). We started work on WiMAX in 2004. And that quickly became the biggest and fastest growing part of the business. Although watch for news on WCDMA and UMTS-LTE (universal mobile telecommunications system-long term evolution) which is how 3G starts to become 4G!

We are the arms dealer that sells bullets to both sides.


WiMax.com: Your product offering appears to rest heavily on the delivery of reference designs.  Is PicoChip primarily a design house that produces designs only for other vendors?  Or does PicoChip actually produce silicon solutions?

We are a fabless chip company (i.e.: the same business model as Broadcom, Xilinx or Sequans). As such, our sales are of silicon.

The distinction is that we are a programmable chip (or, more flamboyantly "software defined radio"), so the same chip can be used for WiMAX, WCDMA, time division synchronous code division multiple access (TD-SCDMA) or others.  The reference designs we offer are the software to implement these systems.


WiMax.com: At a high level (conceptually) what elements do you include in your reference designs? Also, how is your approach innovative relative to past practices or other firms?

The primary deliverable is the chip ("picoArray") - an extremely powerful DSP with a ten-times better performance than any other chip from TI, Xilinx or others.  On top of that we also deliver the software for the physical layer (PHY) for 802.16d, 802.16e and WiBRO.

Critically, the system is software upgradeable. The same hardware can support 802.16d and then seamlessly upgrade to 802.16e in the field. We are the only company to offer such an upgrade.

We are also the only company to offer all four options (16d / 16e, SS and BS) as products.

Included in the reference design is MAC software and a radio from our partners. We manage this, and guarantee that the system as a whole is fully integrated and tested. We take responsibility for interoperability ("PLUGFEST") and getting it through certification at CETECOM.

This model is common in some markets (e.g. TI in asynchronous digital subscriber line (ADSL) products, Broadcom in cable modems). But it is unusual in wireless. I believe we are the only company to offer "carrier class" wireless reference designs for WiMAX and 3G.


WiMax.com: Your product page touts your ability to deliver application-specific integrated circuit (ASIC) price points with the processing power of a field programmable gate-array (FPGA) approach with the ease of use of digital signal processing (DSP).  What exactly does all of this mean in practical terms that a broadband wireless service provider could or for that matter radio vendor could parse as a benefit?

That is a very good question. There are four main benefits to our system:

Because of the performance/dollar advantages of our product, the cost of the system is much, much lower. On a benchmark basis our performance is ten times better. In a real world system we are one-third the cost and one-third the power of DSP / field programmable gate array (FPGA) options.  

The reference design saves an OEM from having to develop the code itself. Thereby saving about a year in development (creating a consequent time to market advantage) and millions of dollars in development costs.  For a manufacturer, the ability to support different standards and updates imparts a big return on investment (ROI) advantage. A manufacturer can invest in one hardware buy. And then enjoy multiple opportunities to sell WiMAX / WCDMA / TD-SCDMA products.  For a carrier there is no need to worry about interoperability, changing standards or hardware updates ever again. For carriers these concerns are very worrying and our ability to defray their concerns is very popular.


WiMax.com: There are numerous chip vendors producing designs and chips for WiMAX applications.  Some are RF chipset vendors. Others are the "backband" or media access controller (MAC) and physical layer (PHY) companies such as Intel and Fujitsu.  Where does PicoChip fit into this ecosystem and why is your approach better?

We are a physical layer (PHY) company. We work with radio frequency chipset (RF) partners (e.g. Filtronic, SiGe, Maxim) and media access controller (MAC) partners (such as Intel, Airspan, Wintegra and others).


WiMax.com: If for example, PicoChip focuses on the PHY layer portion of the design how is that integrated into the MAC design with other third party design houses - in lay terms please?

We deliver a standard interface from the PHY to the MAC. Then third party design houses can interface to this. That is how Intel, Wintegra and others all work.

We can offer that capability at "arms length." For example, some customers write their own MAC to our interface. In that case that portion is their responsibility.

Otherwise, with people like Intel & Wintegra who offer their MAC as a discrete product, we then invest time to develop the interface together. We then jointly guarantee it is integrated and fully compliant - then we (once again together) offer this to customers.


WiMax.com: PicoChip is a successful player with universal mobile telecommunications system (UMTS) designs for broadband wireless systems. How much influence has your design experience with UMTS influenced your WiMAX designs?  Also, how do you compare the relative trade offs between the two types of technology?

That is a very interesting question.

UMTS and WiMAX are really quite different. Other than general experience in programming, there have not been many synergies between the two programs.  There are trade-offs. While 3G technology is very good at mobility, low-power operation and voice delivery, it isn't great for data transport (high speed downlink packet access [HSDPA] is "3G done properly" and does deliver better results, although still not as good as WiMAX).

WiMAX is all IP and great for data.  It can do higher peak data rates - especially for fixed users.  It is also more "forward looking" with capability to support adaptive antenna systems (AAS) and multiple input multiple output (MIMO) antennas.

WCDMA is a lot (a lot, a lot!) more complex for DSP.

That said, I'm very cynical (and have upset a few people with my comments!) about the relative merits of either system being better.  I dont think that the differences are as big as people say.  I also don't think that, in reality, data rates will be that different between the two.  Similarly, it is hard to see a dramatic reason why WiMAX will be cheaper (the biggest costs for a network are radios, cell sites, customer acquisition, etc - and they are technology agnostic).

That said, the advantage WiMAX has it is more open. The business model is more datacom like.  There will be more competition, and the pace of innovation weill be faster.  It isn't the technology - it is the business environment that matters.


WiMax.com:  PicoChip builds designs for WCDMA, testing and TDD solutions as well.  Which is the biggest current segment of your business and how much of your business do you anticipate WiMAX to become in the next three years?

Right now, WiMAX is the biggest. It is a lot easier market to 'crack' than 3G, which is a more conservative market. On the other hand, 3G is a very much bigger market.

I would expect in three years the two markets will be roughly equal for us.

TD-SCDMA is the next opportunity coming along. But it is hard to predict the opportunity just now; it totally depends on the Chinese government.

Test is strategically important for us - being involved with the tester companies is good for credibility and a great way of learning what will happen next, but the revenues are small.


WiMax.com: Who uses your products? What types of manufacturers are they?

PicoChip's PC6520 product is the industry standard basestation reference design for "fixed" WiMAX---with more than 15 major licensees. Products using the PC6520 are available from Airspan, Intel, Ericsson, Nortel, Marconi and a number of other manufacturers. It is being used in a number of deployments, including publicly declared ones in Japan (Yozan), the UK (Pipex) and Italy (Marconi and Italian Ministries).

Other key people are partners like ETRI (who invented WiBRO) and the Chinese Academy of Science.


WiMax.com: Who is the closest similar company to PicoChip and how is PicoChip unique in comparison? Also, what are the three most important technical advantages that PicoChip possesses?

The closest comparison is probably what Broadcom has done in DSL - delivering programmable devices with complete reference designs.

Our three primary advantages:

1.  Deep system expertise. We really understand wireless and offer complete wireless solutions.
2.  Our multi-core DSP delivers a ten times better price/performance than any other solution.
3.  Product flexibility - ease of programming and updating.


WiMax.com: How easily can your designs shift from 802.16-2004 Fixed WiMAX products to the 802.16e mobile product lines?  Are there significant differences and will economies of scale cross that barrier - if so at what level?

The processor can do either 802.16d or 802.16e - but it is different software. You simply download the other code and there you are. There are no problems or barriers at all.  This is a major part of our value; it is a big reason that manufacturers use us.

Indeed, the same chip does both 3G and WiMAX.  We have a very cool demo where one board can do both.  It makes a 3G call to a Nokia handset, then we reboot with new software and have a WiMAX data session via an Intel Rosedale chipset)


WiMax.com: Will you need to deliver separate designs for Fixed WiMAX and Mobile WiMAX? If so, how much difference is there in the two designs and in the actual chips produced from them?

As above: Same hardware, different code.


WiMax.com: Is spectrum range much of an issue for a design house in your niche?

No, we are totally agnostic.  He have customers developing systems everywhere between 450MHz and 5.8GHz.  Obviously, 2.5GHz and 3.5GHz are the most mainstream, but manufacturers like the flexibility we offer and are using that to develop a whole range of different systems.


WiMax.com: Do you see WiMAX being disruptive to 3G technologies or supportive - and why?

It is competitive.  If SKT for example, announces they are cutting 3G capex by $1bn to spend that money on WiMAX or if a customer uses WiMAX to send a MB of data instead of 3G, then they are competitive.


WiMax.com: In practical terms how does your technology reinforce the WiMAX business case of your customers and ultimately wireless providers, both fixed and mobile? Is it just cost savings or are there technical details that really matter?

Yes. As mentioned above, we reduce costs dramatically. The flexibility then has a huge benefit too.


WiMax.com: What question are we not asking about PicoChip's approach to its position in the WiMAX ecosystem that we should be? Essentially, please tell us what we are missing about PicoChip?

One obvious specific thing is funding: We have received over $40 million in investment dollars from investors such as POND, Atlas, Scottish Equity (who invested in CSR & Wolfson),  Intel and Rothschild. There is also a new strategic investor who is very significant.

More broadly, I think we ought to emphasize that although we are leaders in WiMAX, our ambitions are much bigger than that.  "Software Defined Radio" is a word that has some negative connotations, but it is a good description of what we are doing.  As wireless gets more complex (moving to 802.16e, HSUPA, 4G, MIMO etc) then our value proposition gets stronger and stronger.  

Next year, we ship our new chips. They are based on a 90nm manufacturing process (a generation ahead of most people. Indeed, some are still at .18!!)  These will include client side devices but will be software compatible with existing products & leverage the existing robust code across our customer base.

Wireless is changing: as we move to all IP and start thinking of 4G, and we hope that PicoChip is positioned to deliver that future.