April 2011 - Executive Interview
MJ: Please give us a short history of Scintera for those not familiar with your company?
DL: Scintera was founded in 2002 to deliver high performance and very low power signal processing solutions in CMOS. Like many startups, identifying the best market to address with our technology took time and we are now focused on the wireless markets, specifically wireless infrastructure. We are now in full production with our first generation of RF domain power amplifier linearizers and getting ready to expand the markets we serve with the upcoming launch of our second generation products.
MJ: What are your first priorities?
DL: We are enjoying market acceptance and accelerating adoption and that brings with it a unique set of challenges. Customer focus, development of internal processes, product roadmap initiatives, demand creation, and corporate branding are all key priorities.
I just finished a tour of our customers around the world and those visits have sensitized me to the areas where they need our help and the processes we need to put in place to smoothly grow our business.
MJ: What prompted Scintera to address the PA linearization market – how did that evolve?
DL: About 18 months ago, Scintera saw the beginnings of a trend that’s been widely accepted only recently – that the broadband wireless data explosion would require a new way of deploying cellular networks. Known today as Heterogeneous Networks, these networks require an increasing number of low cost and highly linear transmitters. Other pre-distortion solutions were available but none that offered the integration levels, power consumption and size benefits possible with Scintera’s technology. At that early time, Scintera’s technology was directed at the optical infrastructure market but the realization of this trend in cellular caused the company to retool for this nascent and higher growth market.
The demand for wireless data services caught even the most prescient by surprise, and that demand continues to outpace the available infrastructure put in place by many operators. The trend widely accepted within the cellular industry is that in order to meet the demand for data, operators will have to move away from the voice-driven macro base stations to the small cell, distributed antenna, and MIMO approaches needed for improved data coverage and density.
Scintera’s products solve some of the key challenges faced by operators trying to deploy data services and will continue to play a key role in the transition to 4G and beyond.
MJ: What are the existing methods used for pre-distortion and linearization?
DL: First, it is important to recognize that Scintera’s RF PA linearization (RFPAL) technology shares similarities with the other common form of field proven pre-distortion called digital pre-distortion (DPD). Both implementations compensate for AM-AM and AM-PM distortion, PA memory effects, etc. and both provide adaptive feedback to compensate for impairments like PA gain changes due to temperature and PA aging for example.
Secondly, the most popular method for PA linearization is not active pre-distortion, but in fact, simply operating the PA at a reduced power level (operation in backoff) to achieve the desired distortion performance.
MJ: What are the differences between RF PA linearization and digital pre-distortion solutions?
DL: Though Scintera’s RFPAL shares underlying similarities in theory with digital pre-distortion, the similarities end in their implementation. Scintera takes a new approach to RF power amplifier linearization by re-partitioning portions of the pre-distortion algorithm from the digital to the analog/RF domain. The result is a single chip and highly integrated solution that maintains the flexibility of digital approaches while offering the simplicity and low power consumption of analog approaches.
Looking closer at the differences, digital pre-distortion expands the bandwidth (adds the pre-distortion signal to the desired signal) at the earliest point in the signal chain, the digital baseband. That bandwidth expansion is propagated throughout the entire transmitter chain and back again to the digital baseband thus increasing system complexity and power consumption. Scintera’s approach moves this expansion to just before the PA input resulting in a greatly simplified and lower cost transmitter and baseband architecture. Analog signal processing with digital control enables a new way of thinking about power amplifier pre-distortion – more importantly, it enables new applications where traditional digital-only solutions cannot compete due to size, cost or power consumption limitations.
MJ: What applications/markets can benefit most from this technology?
DL: We focus primarily on three markets, including cellular infrastructure, microwave radio and broadcast infrastructure. However, any application requiring linear amplification of RF signals can benefit from Scintera’s RFPAL. These markets include white space transmitters, military and public safety, etc.
Scintera’s RFPAL can be operated in a standalone fashion enabling its use in applications where previously the only option was using a Class A or Class A/B amplifier operated in backoff.
MJ: What kind of improvements in power, efficiency and ACLR can be expected by implementing this Scintera’s RFPAL?
DL: Many factors determine the system performance of any linearization solution. Comparing RFPAL against solutions operating in backoff, customers can realize the greatest improvements. For a given average output power level at the antenna, customers can expect up to 4X improvement in efficiency, thus enabling dramatic decreases in the system power consumption and power supply costs, and the ability to cut power transistor size and cost in half due to the ability of the linearizer to deliver 3dB of additional output power from the existing device. At the same time, the RFPAL will deliver similar or better ACLR (distortion) performance than when the PA was operated in backoff.
It should be noted that there is an entire class of applications that due to cost, complexity and system power consumption considerations never considered using DPD. This is especially relevant in the small cell market and at output power levels below 5W. Scintera is the only solution that can address that market and effectively compete against a PA operating in backoff.
In applications requiring average output power levels greater than 5W at the antenna, digital pre-distortion has some market share and that share increases as the power level increases. Against DPD, RFPAL delivers a more elegant solution that reduces the solution cost, reduces footprint size and provides comparable levels of correction performance. In some niche applications, where all considerations are put aside for maximum correction performance, DPD is perhaps a better option.
As a secondary note, there are again entire classes of applications that DPD cannot address due to its costly implementation and, by its very nature, requirement for a microprocessor. In these markets which include remote radio heads, power amplifier modules, repeaters and microwave backhaul systems, RFPAL is again the only available solution.
MJ: What protocols, signal characteristics (power levels & PAR) and frequency range is this technology effective over?
DL: The design philosophy behind Scintera’s RFPAL was to create a solution agnostic to output power level and frequency, communication protocol and PA configuration and device technology.
Realistically speaking, RFPAL is best used in systems down to 500mW in average output power at the antenna – below that the RF power device cost is low enough that oversizing it for operation in backoff makes the best economic sense. In the cellular market where at higher output power levels the key specification is the spectral emissions mask (SEM), RFPAL works best in systems operating up to 60W of average power at the antenna. In the broadcast market there is no SEM limitation, and we have been integrated into systems operating up to 600W of average output power and even higher power may be possible.
Scintera’s SC1887 is ideally suited to applications requiring operation from 470 to 2800MHz, PAR of up to 10dB, static average power waveforms like WCDMA or DVB-T and in systems using LDMOS class A/B or Doherty amplifiers. Scintera’s next generation products make further improvements to more closely meet its design philosophy goals.
MJ: What are the main challenges you have in getting companies to use your technology?
DL: Of course, there are the usual suspects of time and resources and, as you might expect, some of the barriers are market dependent.
We should first distinguish between evaluation and adoption. Speaking to evaluation, we have seen customers quickly and easily agree to evaluate Scintera’s RFPAL. Due to our unique RFin/RFout architecture and our use of directional couplers to interface with the power amplifier system, our customers have been able to get results in their labs typically within ½ day of opening our evaluation kit. That low time and resource investment has been instrumental in getting many customers to evaluate our RFPAL and see the benefits of our solution.
Speaking to adoption next, we see a customer base which is naturally conservative and skeptical of new technology based on the customers that they serve. However, we also see an industry looking for solutions that are more cost-effective, simpler to implement, lower power and meet their performance requirements. RFPAL meets all those requirements and now has the field-proven reliability that is demanded by many of our infrastructure customers.
So, a fast evaluation process, compelling benefits and a proven record of in-field robustness and reliability has led many customers to actively design our RFPAL into their products because as a standalone linearizer solution there is no need for large teams of DPD experts, no customized settings and no complex system designs. Our RFPAL is as close as you can get to being plug-and-play for an active RF component.
The missing piece needed to improve customer adoption is more information regarding our pre-distortion linearizer technology (how it works) and raising the corporate visibility of Scintera (who we are). We are moving on that front by writing more contributed articles like the one in April’s issue of MW Journal, presenting on our technology at trade-shows and industry events and, of course, participating in press interviews.
MJ: Are there any plans to partner up with complementary component or distribution companies?
DL: Success rarely happens in isolation. Our distribution partners like Richardson Electronics (now part of Arrow and named Richardson RFPD) has and will continue to play a significant role in our success. Richardson RFPD has a synergistic line card enabling customers one-stop shopping for the transmitter electronics and the deep customer relationships needed for Scintera to gain even broader market acceptance.
Furthermore, we are engaging with the entire ecosystem to better provide guidance and reference designs for power amplifiers, full radio solutions and complete single board base stations including all digital and RF functionality.
MJ: What is the next type of product that will be released?
DL: There is always a need for continued integration, improved levels of performance, more features and lowered solution cost. The industry trend toward small cell data-centric networks, low power transmitters, distributed antenna systems, etc. all play a role in addressing the vectors above and in our roadmap planning. Importantly, we are seeing a need to provide more measurement and monitoring capabilities on the input and output signals of the PA. This requirement ties in nicely with our ability to provide digital and analog signal processing right where it’s needed - next to the PA.
Our next generation products will address all four of these vectors.
MJ: What are your future plans for the company and this technology?
DL: We first applied our analog signal processing technology to the pre-distortion market. That was partially due to the company’s past domain expertise. There was a time when RF problems were partitioned into the digital domain whenever possible because it was thought that that approach delivered the best trade-offs. What Scintera has shown is that for specific classes of problems analog domain signal processing and a more thoughtful partitioning approach can yield superior results to one solely dependent on digital domain processing.
Our customers are pointing the way for us as to the type of products they want to see beyond linearization solutions and we are listening.