Wireless & RF Magazine: November 2014

Interview with Preet Virk Senior VP & GM of Networking Business Unit, MACOM

MACOM

Embraces Explosive Growth Next-Gen Technologies for Diverse Wireless Markets Configuring IoT Devices Linux-Less Wi-Fi November 2014

Join Today

eeweb.com/register

CONTENTS

High Efficiency Wireless Power Transfer Amplifier Topologies Two high efficiency amplifier topologies will be analyzed, namely the ZVS Class-D and the single-ended Class E. The schematics and ideal operating waveforms for each of the amplifier topologies is shown in figure 2.

The ZVS Class-D topology makes use of a non-resonant ZVS tank circuit to allow the switch-node to self-commutate between switching transitions, which effectively eliminates the output capacitance (COSS) associated losses of the devices, from a Class-D implementation.

The single device Class-E topology makes use of a resonant circuit Le and Csh, whose resonant frequency differs from the operating frequency, to establish the conditions necessary for ZVS. In this design, the output capacitance (COSS) is effectively connected in parallel with Csh and thus becomes part of the resonant circuit needed to establish ZVS. In some cases, the design of the Class E will be limited to the value of COSS as the value of the external capacitor Csh reduces to zero.

Device Comparison A Wireless Power Transfer figure of merit (FOMWPT), as defined in [5], was used to compare the eGaN FETs with best-in-class MOSFETs and is shown in figure 3. Superior devices will have lower values of FOMWPT. It is clear from the FOMWPT that eGaN FETs inherently exhibit potentially superior performance in both the amplifier topologies.

Experimental Verification Four amplifiers were constructed in this example, two of each type, where one was fitted with an eGaN FET and the other a MOSFET. Each of the amplifier designs was carefully adjusted based on the small differences in COSS of each device, which ensured that operating conditions were the same in each case. A special load was also constructed for testing that could be discretely configured for various real and imaginary load settings. This load was calibrated using a VNA and the discrete load range settings were varied from +30j Ω through -30j Ω and 1 Ω through 56 Ω.

Some of measurements are shown in figure 1 for real impedance variations at various imaginary settings. The measured values were then used to accurately determine the power delivered to the load during power testing.

CONTENTS The ProbMe Solution Let’s assume that the user has a device on any operating system and does not have physical access to wireless router.

Let’s also assume the user is minimally tech-savvy, able to navigate the basic menus required for getting the device configured on the home’s network.

In addition, the user doesn’t want to announce to the world a coffee maker on the Wi-Fi, nor does the user wish to have an unsecured coffeepot, an easy target for neighborhood shenanigans.

TECH REPORT

ProbMe Leverages Wi-Fi Technology ProbMe works by utilizing the discovery functionality of Wi-Fi networks, but instead of using the standard 802.11 management frames to simply interrogate the surrounding networks for their SSID information, the ProbMe method uses these frames as configuration carriers. This enables the Econais WiFi module to know when it is receiving configuration information and then use that information to configure itself to the network. With multiple devices on at once, all of them can be configured at the same time with the same SSID and pass phrase. Imagine configuring 10, 100, or even 1,000 devices all at the same time with one configuration command.

Configuring IoT Devices ProbMe™ Makes It Easy A user would need the coffee pot to power-on in a listen-only mode and await configuration. Based on the particular device, it might be desirable to configure in a particular mode or a specific network.

And it would be ideal that if multiple devices were turned on out-of-the-box, that all of them could be configured identically with one single entry by the user. Then, after configuration, to keep out the drive-by hackers, a device would lock itself down from future reconfiguration, assuring it could only be managed by authorized users. This would likely require a user to have access to the device to push a reset button if the device ever needed to be reconfigured.

TECH SERIES A smartphone transmits configuration information and automatically

smartphone transmits the configuration information and automatically connects all the new ProbMe devices to the network.

Once a device is configured with this information, the module by default then disables the ability to be reconfigured and requires a physical and manual button to be pressed on the device in order for the device to be placed back into configuration

WiGaN eGaN® FETs Yield High Efficiency

The frames that include the configuration information can be adapted to support various levels of security, whether manually by users or automatically by a manufacturer or Econais-supplied application.

With the onset of the Internet of Things and the expectation that the tens of billions of devices will be connected in homes and businesses, it is extremely important for manufacturers to supply the end user with a method to easily secure and configure new devices to networks.

The new ProbMe software is included with the EC19D and EC19W Econais Wi-Fi Modules, Development Kits and Reference Designs and are available through major industry distributors such as Digi-Key. For additional information regarding ProbMe, please visit the Econais weblink at www.econais.com/probme.

4 12

Econais is an embedded module manufacturer and solutions company building ultra low power connectivity solutions to address the expanding Internet of Things (IoT) and Machine to Machine (M2M) markets. WiSmart modules integrate 32bit MCU, Wi-Fi, and cloud connectivity with rich and comprehensive libraries of software and vertical applications. Econais’ easy-to-use wireless modules and integrated software, packed with advanced features, enable customers to leverage the globally installed base of Wi-Fi access points and smartphones to quickly and confidently create connected smart products for audio, video, smart home, wearables, healthcare, smart energy, consumer, control, and monitoring in industrial, commercial, and residential markets. http://www.econais.com

Imagine a newer, improved approach to the ones described above. In this approach, the user brings home the coffee maker.

After plugging it in, the appliance owner looks at the quick-start card, and then by using a smartphone, enters the proper network SSID and pass phrase, and finally clicks, “Join,” “Save,” or “Submit.” The

connects all new ProbMe devices to the network.

mode. Various manufacturers could elect to set up a secure approach to access the reset capability through a physical button or software.

Figure 3. Wireless Power Figure of Merit Comparison for the Class E and ZVS Class D topologies between eGaN FET and MOSFETs.

INDUSTRY INTERVIEW

MACOM Delivers Nex-Gen Technologies Interview with Preet Virk, Sr. VP & GM at MACOM

Linux-Less

Wi-Fi

TECHWLAN REPORT for Embedded Design Linux-Less Wi-Fi WLAN for Embedded Design

E

mbedded system designers face

many challenges when selecting

20

26

the right components to meet

functional, budgetary, and regulatory requirements, especially when it involves wireless technologies. Until recently, adding Wi-Fi meant learning and implementing complex software subsystems just to provide a seemingly simple wireless network connection. This typically would involve a deep dive into embedded operating systems such as Linux, complex development

With over 30 billion connected devices expected by 2030, MACOM is prepared to reinvent the mobile network as we know it...pg. 20

environments to compile Wi-Fi drivers,

Article provided by

rebuilding kernels, and countless hours spent wrestling with tools unrelated to your actual product functionality.

3

Configuring

IoT Devices

By Costas Pipilas, Vice President of Research & Development, Econais Inc.

4

TECH REPORT

U

ser adoption of new wireless technologies presents several

challenges. Three of the biggest that hamper deployment and widespread use of new technologies are installing infrastructure, ease of use, and ease of installation and configuration.

Given the globally installed base and use of Wi-Fi for everything from smartphones to tablets, to laptops and desktop computers, the first two items are taken for granted by most Wi-Fi wireless users.

With the emergence of the global Internet of Things (IoT), however, the issues faced by a user trying to quickly, easily, and securely configure a new device to the home network have only recently been solved. ProbMe™, from Econais addresses

ProbMe

simple and easy configuration of devices for the Internet of things.

Makes It Easy 5

The Basics Most devices that users connect to a Wi-Fi network need a service-set identifier (SSID) and a pass phrase.

Devices now being added to homes or commercial spaces are quite small and don’t have a display or convenient dataentry capability that allows users to easily configure them.

Manufacturers of wireless routers initially supplied routers with a default and generic SSID as well as open settings that required no pass phrase to join the network.

Most router manufacturers are now providing unique SSID’s and initial unique pass phrases that are supplied on a sticker on the Wi-Fi router. This means that when the router is turned on, it comes on in a secure mode with a set SSID and pass phrase. Anyone wanting to join the network then uses the “Join Network” or “Add Network” utility on a smart phone, tablet, laptop, or desktop. A user can easily join a network, and when moving between locations, automatically rejoin the network.

The Problem Today Most of the devices that consumers have connected to Wi-Fi wireless routers have rich computing resources, displays, keyboards, and other capabilities that make entering the information for a secure connection relatively easy.

However, the devices now being added to homes, offices, commercial spaces, and industrial facilities are quite small and don’t have a display or convenient data-entry capability that allows users to easily configure them. Thus, the devices need to be configured by a smarter device that is also on the network.

6

TECH REPORT For instance, consumers can now buy coffee makers that are Wi-Fi enabled. The coffee maker needs to be told the SSID and pass phrase so it can connect to the network as a client device and allow the user to turn the coffee maker on and off remotely, change brewing settings, get a signal that the coffee has finished brewing, and possibly even connect the coffee maker to the cloud for additional advanced features. Today, there are a couple of options for manufacturers who want to put Wi-Fi capability in that coffee maker.

pass phrase. The device is then configured as a client on the user’s network.

This is a very straightforward approach whereby the coffee maker is plugged in, powered on, and the SSID of the coffee maker is broadcast so that it is easily identified.

The user disconnects from their wireless network, connects to the coffee maker, configures it, and the process is complete.

The disadvantage of this method can be the required level of comfort with technology.

Also, the process is quite time consuming, There are two likely approaches. The first which is compounded if there are several is to use a Wi-Fi protected setup (WPS) devices to install, such as light switches, button on the coffee maker and assume light bulbs, outlets, speakers, cameras, the user’s wireless router has a WPS button door locks, thermostats, alarms, and so as well. This requires the user to simply on. Most importantly, this may not be the press the button on the coffee pot and also most secure approach depending on how press the button on the wireless router. the manufacturer decides to handle the start-up scenario. In some cases, consumers The advantage of this approach is ease may not decide to configure the Wi-Fi of installation if both devices and both feature to their wireless network, and thus, buttons are easy to access. However, if it would remain visible to everyone within the wireless router doesn’t have a Wi-Fi broadcast range of the device. Disgruntled protected setup WPS button or if it isn’t next-door neighbors could find and then easily accessible, this system won’t work seize control of each other’s coffee makers. at all. Also, if both buttons are pressed and the coffee maker doesn’t show up on the network, there’s no way to troubleshoot.

ProbMe works by utilizing the Most importantly, the WPS push-button discovery functionality of Wi-Fi networks, method has known security flaws.

Another common approach is to have the coffee maker come up as an access point with its own SSID and pass phrase, which is supplied by the manufacturer’s quick-start card. With a browser, a user logs in to the coffee maker, navigating to an area to enter the SSID and network

but instead of using the standard 802.11 management frames to simply interrogate the surrounding networks for their SSID information, the ProbMe method uses these frames as configuration carriers.

7

The ProbMe Solution Let’s assume that the user has a device on any operating system and does not have physical access to wireless router.

Let’s also assume the user is minimally tech-savvy, able to navigate the basic menus required for getting the device configured on the home’s network.

In addition, the user doesn’t want to announce to the world a coffee maker on the Wi-Fi, nor does the user wish to have an unsecured coffeepot, an easy target for neighborhood shenanigans.

A user would need the coffee pot to power-on in a listen-only mode and await configuration. Based on the particular device, it might be desirable to configure in a particular mode or a specific network.

And it would be ideal that if multiple devices were turned on out-of-the-box, that all of them could be configured identically with one single entry by the user. Then, after configuration, to keep out the drive-by hackers, a device would lock itself down from future reconfiguration, assuring it could only be managed by authorized users. This would likely require a user to have access to the device to push a reset button if the device ever needed to be reconfigured.

A smartphone transmits configuration information and automatically connects all new ProbMe devices to the network.

8

Imagine a newer, improved approach to the ones described above. In this approach, the user brings home the coffee maker.

After plugging it in, the appliance owner looks at the quick-start card, and then by using a smartphone, enters the proper network SSID and pass phrase, and finally clicks, “Join,” “Save,” or “Submit.” The

TECH REPORT smartphone transmits the configuration information and automatically connects all the new ProbMe devices to the network.

ProbMe Leverages Wi-Fi Technology ProbMe works by utilizing the discovery functionality of Wi-Fi networks, but instead of using the standard 802.11 management frames to simply interrogate the surrounding networks for their SSID information, the ProbMe method uses these frames as configuration carriers. This enables the Econais WiFi module to know when it is receiving configuration information and then use that information to configure itself to the network. With multiple devices on at once, all of them can be configured at the same time with the same SSID and pass phrase. Imagine configuring 10, 100, or even 1,000 devices all at the same time with one configuration command.

Once a device is configured with this information, the module by default then disables the ability to be reconfigured and requires a physical and manual button to be pressed on the device in order for the device to be placed back into configuration

mode. Various manufacturers could elect to set up a secure approach to access the reset capability through a physical button or software.

The frames that include the configuration information can be adapted to support various levels of security, whether manually by users or automatically by a manufacturer or Econais-supplied application.

With the onset of the Internet of Things and the expectation that the tens of billions of devices will be connected in homes and businesses, it is extremely important for manufacturers to supply the end user with a method to easily secure and configure new devices to networks.

The new ProbMe software is included with the EC19D and EC19W Econais Wi-Fi Modules, Development Kits and Reference Designs and are available through major industry distributors such as Digi-Key. For additional information regarding ProbMe, please visit the Econais weblink at www.econais.com/probme.

Econais is an embedded module manufacturer and solutions company building ultra low power connectivity solutions to address the expanding Internet of Things (IoT) and Machine to Machine (M2M) markets. WiSmart modules integrate 32bit MCU, Wi-Fi, and cloud connectivity with rich and comprehensive libraries of software and vertical applications. Econais’ easy-to-use wireless modules and integrated software, packed with advanced features, enable customers to leverage the globally installed base of Wi-Fi access points and smartphones to quickly and confidently create connected smart products for audio, video, smart home, wearables, healthcare, smart energy, consumer, control, and monitoring in industrial, commercial, and residential markets. http://www.econais.com

9

Your Circuit Starts Here.

Sign up to design, share, and collaborate on your next project—big or small.

Click Here to Sign Up

By Alex Lidow, CEO Efficient Power Conversion (EPC)

12

TECH SERIES

Wi GaN: eGaN FETs 速

In Wide Load Range High Efficiency Wireless Power

eGaN FETs have previously demonstrated higher efficiency in loosely coupled wireless power transfer solutions. When operating on-resonance using either ZVS Class-D or Class-E amplifiers [1, 2, 3, 4, 5]. However, practical Wireless Power systems need to address the convenience factor of such systems, which results in reflected coil impedances that can significantly deviate from resonance as load and coupling vary. These systems still need to deliver power to the load and hence the amplifier needs to drive the coils over a wide impedance range.

Standards such as the A4WP Class 3 have defined a broad coil impedance range that address the convenience factor and can be used as a starting point to compare the performance of the amplifiers.

In this installment of Wi GaN, both the ZVS Class-D and Class-E amplifiers will be tested at 6.78 MHz to the A4WP Class 3 standard with a reduced impedance range to determine the inherent operating range limits. Factors such as device temperature and voltage limits will determine the bounds of the load impedance range each amplifier is capable of driving.

13

A4WP Class 3 Nominal Operating Range The A4WP Class 3 standard defines a wide impedance range—+10j Ω through -150j Ω imaginary and 1 Ω through 56 Ω real. An amplifier needs to be able to drive this impedance range with 800 mARMS nominal current, which is derated when the power delivered reaches 16 W. The entire impedance range is shown on the Smith Chart in figure 1 by the blue shaded area and is also known as the four corners. Since the range is so wide, it is permissible to rotate the impedance range to improve efficiency and performance of the amplifier driving the coil. This impedance rotation is, under certain conditions, referred to as adaptive matching as the circuit actively seeks to find the most suitable operating impedance of the coil and is shown by the dashed blue arc (for no specific rotation).

Given that the A4WP Class 3 impedance range is so wide, the first step in a wireless power system design is to determine the practical operating impedance range. Once known, this value will determine the number of discrete steps needed

Figure 1. Smith Chart showing the A4WP Class 3 reflected impedance range and the discrete load trajectories used for experimental testing.

for adaptive matching to cover the entire Class 3 range. The practical limits for the amplifier include rated device voltage limits, temperature limits, and in some cases, supply voltage limits.

In this experimental analysis, the device voltage limit of 80% of rated and a device temperature limit of 100°C (as observed by IR camera) in an operating ambient of 28°C will be used.

Figure 2. ZVS Class D (left) and Class E (right) amplifier schematic with ideal waveforms.

14

TECH SERIES

High Efficiency Wireless Power Transfer Amplifier Topologies Two high efficiency amplifier topologies will be analyzed, namely the ZVS Class-D and the single-ended Class E. The schematics and ideal operating waveforms for each of the amplifier topologies is shown in figure 2.

The ZVS Class-D topology makes use of a non-resonant ZVS tank circuit to allow the switch-node to self-commutate between switching transitions, which effectively eliminates the output capacitance (COSS) associated losses of the devices, from a Class-D implementation.

The single device Class-E topology makes use of a resonant circuit Le and Csh, whose resonant frequency differs from the operating frequency, to establish the conditions necessary for ZVS. In this design, the output capacitance (COSS) is effectively connected in parallel with Csh and thus becomes part of the resonant circuit needed to establish ZVS. In some cases, the design of the Class E will be limited to the value of COSS as the value of the external capacitor Csh reduces to zero.

Device Comparison A Wireless Power Transfer figure of merit (FOMWPT), as defined in [5], was used to compare the eGaN FETs with best-in-class MOSFETs and is shown in figure 3. Superior devices will have lower values of FOMWPT. It is clear from the FOMWPT that eGaN FETs inherently exhibit potentially superior performance in both the amplifier topologies.

Experimental Verification Four amplifiers were constructed in this example, two of each type, where one was fitted with an eGaN FET and the other a MOSFET. Each of the amplifier designs was carefully adjusted based on the small differences in COSS of each device, which ensured that operating conditions were the same in each case. A special load was also constructed for testing that could be discretely configured for various real and imaginary load settings. This load was calibrated using a VNA and the discrete load range settings were varied from +30j 立 through -30j 立 and 1 立 through 56 立.

Some of measurements are shown in figure 1 for real impedance variations at various imaginary settings. The measured values were then used to accurately determine the power delivered to the load during power testing.

Figure 3. Wireless Power Figure of Merit Comparison for the Class E and ZVS Class D topologies between eGaN FET and MOSFETs.

15

Figure 4. Measured efficiency comparison between the eGaN FET and MOSFET-based, ZVS Class-D amplifiers at various imaginary impedances as function of real impedance.

The efficiency results, including gate driver power, for the ZVS Class-D amplifiers are shown in figure 4 for various imaginary impedance settings as a function of real impedance variation together with the power delivered to the load.

It can be seen that under the same operating conditions, the eGaN FET-based amplifier is always operating much more efficiently. In this case the amplifier is capable of complying with the A4WP standard over a 50j 立 range (shown in figure 1 red dash arc) while the device voltage and temperature remain within specifications.

The efficiency results, including gate driver power, for the Class-E amplifiers are shown in figure 5 for various imaginary impedance settings as a function of real impedance variation together with the power delivered to the load. It can be seen that under the same operating conditions, the eGaN FET-based amplifier is again always operating much more efficiently. In this case the amplifier is only capable of complying with the A4WP standard over a 30j 立 (shown in figure 1 purple dash arc) range while the device voltage and temperature remain within specifications.

Figure 5. Measured efficiency comparison between the eGaN FET and MOSFET-based, Class-E amplifiers at various imaginary impedances as function of real impedance.

16

TECH SERIES

Also notable for the Class-E amplifier is how the imaginary impedance affects the peak operating efficiency. This is due to the imaginary impedance being connected in series with the Class-E resonant circuit Le and shifts the optimal operating point.

Summary In this column, we presented the efficiency comparison results between a ZVS Class-D and Class-E amplifier. Each of these was fitted with either eGaN FETs or MOSFETs, when tested over a wide impedance variation range. In each case, the eGaN FETs yielded higher amplifier efficiency.

The ZVS Class D demonstrated that it is capable of operating over a 50j Ω range, whereas the Class E could only achieve a 30j Ω range while maintaining compliance with the A4WP Class-3 standard and operating within specified voltage and temperature parameters. The device voltage rating for the ZVS Class-D topology is predictable and under full control of the system, whereas the Class E requires not only a higher voltage rated device than the ZVS class D, but operation under various load conditions can lead to unpredictable high peak drain voltages.

The wider operating range reduces the number of discrete bits required for discrete adaptive matching to 2 bits for full Class-3 compliance as each bit in an adaptive matching circuit will add cost to a wireless power transfer system.

For Class E, 3 bits or more for discrete adaptive matching may be required for full Class-3 compliance.

References [1] M. A. de Rooij, “eGaN® FET based Wireless Energy Transfer Topology Performance Comparisons,” International Exhibition and Conference for Power Electronics, Intelligent Motion, Renewable Energy and Energy Management (PCIM - Europe), May 2014, pg. 610 – 617.

[2] A. Lidow, M. A. de Rooij, “Performance Evaluation of Enhancement-Mode GaN transistors in Class-D and Class-E Wireless Power Transfer Systems,” Bodo Magazine, May 2014, pg. 56 – 60.

[3] A. Lidow, “How to GaN: Stable and Efficient ZVS Class D Wireless Energy Transfer at 6.78 MHz,” EEWeb: Pulse Magazine, Issue 126, pp. 24 – 31, July 2014.

[4] W. Chen, et al., “A 25.6 W 13.56 MHz Wireless Power Transfer System with a 94% Efficiency GaN Class-E Power Amplifier,” IEEE MTT-S International Microwave Symposium Digest (MTT), June 2012, pg. 1 – 3.

[5] M. A. de Rooij, “Performance Evaluation of eGaN® FETs in Low Power High Frequency Class E Wireless Energy Converter,” International Exhibition and Conference for Power Electronics, Intelligent Motion, Renewable Energy and Energy Management (PCIM - Asia), June 2014, pg 19 – 26. eGaN® FET is a registered trademark of Efficient Power Conversion Corporation.

17

MACOM

Delivers Next-Gen

TECHNOLOGIES

for Diverse Wireless Markets Interview with Preet Virk Senior Vice President & GM of Networking Business Unit at MACOM

T

he wireless-carrier market is on the verge of explosive growth, with Cisco predicting around 30 billion connected devices by 2030. This device boom will present serious demands on preexisting wireless networks, prompting wireless technology providers to start on the immense task of ensuring network stability for future markets.

Among these companies is MACOM, a developer of wireless semiconductor devices based in Lowell, Massachusetts.

MACOM is bracing for the projected network growth by acquiring companies that not only bring unique solutions to the table, but have invested in new process technologies that will make these massive new networks possible. EEWeb spoke with Preet Virk, Senior Vice President and General Manager of the Networking Business Unit at MACOM, about the unique ways this new process technology will contribute to shaping the wireless networks of the future.

20

INDUSTRY INTERVIEW

21

What is the current state of the wireless-carrier market?

The wireless-carrier market is definitely full of excitement. There has been explosive traffic growth both in the access network as well as the metro & core network. Recently, we have had many discussions about this explosive traffic growth and the vast number of pervasive connections—whether you refer to Cisco’s numbers or Ericsson’s numbers, they forecast 30 billion connected devices by 2030. We are talking about tens of billions of devices in a world that has a population of seven billion. In terms of the tremendous amount of pressure on the network bandwidth, these incredibly high number of devices are already creating a big disconnect on what it costs operators to service a megabit, or petabyte, versus the revenue generated from this megabit, or petabyte. The service provider CapEx spending cannot keep pace with the traffic growth. So, one answer is to fundamentally change the way e architect the network. When you think about network function virtualization (NFV) and software-defined networks (SDN), or the need to open up higher frequency bands, in mmWave frequency bands, with broader bandwidths, upto 500Mhz for 5G, smart, phase arraybased antenna technology with beam

forming and very high capacity optical networks for front haul and backhaul applications. These same dynamics are what lead the strong interest in GaN based products. These are the kinds of fundamental transformations needed to meet the traffic growth challenges.

MACOM’s strengths lie in these exact areas: mmWave, phased array antenna technology, beam forming, Optical and GaN portfolio of products, technologies and capabilities, that are necessary to enable these next generation networks.

What has MACOM’s motivation been behind its recent acquisition spree?

The semiconductor industry is undergoing consolidation. There are several factors leading to this consolidation, driven by the business environment and the customer base.

Customers want more complete, easy to use solutions from proven suppliers.

MACOM’s acquisition of Mindspeed aligns with MACOM’s strategy in networking and 100G optical markets. Between Mindspeed and MACOM’s previous acquisition of Optomai, we now have a leadership position in 100G. The cost structure and the focus on this segment of the business is characterized by high margins and long life cycles, which is exactly the model that MACOM has. From a channel perspective, Mindspeed was

“When you think about Network function virtualization (NFV) and software-defined networks (SDN), the ability to open up higher frequency bands is imperative, and it is one of MACOM’s key strengths.”

22

INDUSTRY INTERVIEW strong in the U.S. as well as in Asian and Pacific markets. These acquisitions allow us to grow the markets we are serving and bring in technology that is becoming more significant, such as capabilities with silicon-germanium (SiGe), which Mindspeed brings to the table.

Our approach with Mindspeed is similar to our tactic with Nitronex because of their capabilities with gallium nitride (GaN), which is required in the RF and microwave application space. With Nitronex’s capabilities, portfolio, and team, we are in a place to drive a significant market share in MACOM’s core businesses. With the new cost structure of GaN on silicon we can have this as with the new cost structures of GaN on silicon.

We can have a phenomenal increase in our target markets as laterally diffused metal-oxide semiconductor (LDMOS) starts to be replaced in mainstream applications. We increased our capabilities with GaN on silicon and GaN on silicon carbide with Nitronex. Now with Mindspeed, we have silicon-germanium as well. Each of these acquisitions has served the purpose of what MACOM’s customers want us to focus on.

How have you been able to leverage all of these new process technologies to address customer needs?

Customers need solutions for a slew of applications. Some of them lend themselves to gallium arsenide (GaAs) and some to SiGe. Most of the mainstream, high-volume applications are still based on GaAs. GaN, with its highpower, high-efficiency breakdown voltage

is extremely valuable, which is something that LDMOS is not able to achieve.

MACOM has a great track record of developing GaAs-based products going back multiple decades. Depending on the application, different process technologies bring different values to the table. If you are talking about higherfrequency and higher-power solutions, then GaN is very interesting. In terms of voltage-controlled oscillators (VCOs), we have implemented SiGe in our products, which has solved problems much more efficiently for our customers. We have the capabilities to address specific opportunities, with specific suitable technologies, whether its GaAs, GaN or SiGe. Overall the great thing about MACOM is that we collaborate with lead customers and offer multiple options to solve the customer’s problem. This culture of collaborating with industry thought leaders and delivering market leading products goes back over 60 years and our ‘trusted partner’ model still holds true today.

With such a broad portfolio, how do you determine a process technology that is appropriate for a specific application?

This is best answered with an example.

Earlier this year, we announced our E-Band portfolio that will address smartset applications in the wireless space.

E-Band is a frequency band that is not easy to develop for. Given MACOM’s history in working at higher frequencies for both commercial and military applications, E-Band will be a great

“MACOM has a great track record of developing GaAs-based products going back multiple decades.”

23

example of how MACOM can bring together its capabilities and familiar technologies for a new application. The concept of a smart-set appeals to customers who look at the RF signal chain in its entirety to create front-ends that are optimized for a particular application with a particular frequency band. This will impact their time-to-market and also their time-torevenue. Having a portfolio that has the VCOs and voltage variable attenuators (VVAs) all coming from MACOM and all specified and architected to work together is what provides the smart set for the customer. Most of the original equipment manufacturers have the same engineering as well as research and development constraints that their competitors do.

For us to solve the RF and signal chain problems through clever design and appropriate choice of process technologies has been very valuable. The E-Band smart set we have put out has garnered much praise. It has not only created the chips, but it has the architectural knowledge to solve bigger RF chain problems.

Examples of process technologies we typically use are GaAs (LNA, Switch), SiGe (VCO) and both GaAs and GaN for power amplifiers.

“Having a portfolio that has the VCOs and voltage variable attenuators (VVAs) all coming from MACOM and all specified and architected to work together is what provides the smart set for the customer.”

24

What are some of the challenges of opening up the network spectrum for commercial uses?

Backhaul point-to-point line of sight has some inherent physics issues because as the wavelengths go higher, their propagation is not the same as is in the lower frequencies.

These physics issues have been addressed by multipath backhaul as opposed to pointto-point to overcome the path loss issues.

The high-capacity link this provides becomes valuable, especially when you start doing densification of the network. If you are going to have a very dense network where the cell distance gets smaller and smaller, then some of the things that used to matter become nonissues. This densification of the network is going on as we speak.

The capacity on the semiconductor side—providing higher gain and retaining linearity—is an interesting challenge. When using advanced modeling techniques and innovative semiconductor circuit topologies along with a high-power amplifier, efficiency is key. The accuracy and precision of these models becomes very important, and that is where MACOM excels. Linearity of the converter not only depends on the mixer topology, but on the accuracy and coverage

INDUSTRY INTERVIEW of the associated models. We have that experience and knowledge base. The combination of 60 years of history and execution along with our capabilities allows us to solve the problems inherent with higher frequency bands.

Let’s talk about fifth-generation wireless systems. What will be some of the issues with them, and what challenges will MACOM need to overcome?

In addition to comments I made earlier, I would add that the specifications of fifth-generation wireless systems (5G) are still being developed. This will involve a complex set of technology requirements, latency requirements, and always-on requirements. The movement towards the cloud is something that will need to be addressed with 5G.

In the RF domain, there has to be support for massive-scale MIMO (multiple-input, multiple-output).

With 5G, there is talk about the need for an order of magnitude higher than that, so massive arrays need to be built. This creates interesting opportunities for MACOM because there will be an explosion in transmit and receive paths-per-base station.

When you go back to the RF domain, two things stick out: higher frequency bands and an explosion of the number of transmit and receive channels. This tells me that our silicon opportunity is going to be absolutely fantastic in this domain, and GaN will have a significant part and, from a full RF-chain perspective, SiGe will have a role to play as well.

“The specifications of fifth-generation wireless systems (5G) are still being developed.”

25

26

TECH REPORT

Linux-Less

Wi-Fi

WLAN for Embedded Design

E

mbedded system designers face many challenges when selecting the right components to meet

functional, budgetary, and regulatory requirements, especially when it involves wireless technologies. Until recently, adding Wi-Fi meant learning and implementing complex software subsystems just to provide a seemingly simple wireless network connection. This typically would involve a deep dive into embedded operating systems such as Linux, complex development environments to compile Wi-Fi drivers,

Article provided by

rebuilding kernels, and countless hours spent wrestling with tools unrelated to your actual product functionality.

27

In addition, the complexity of these subsystems often times can lead to processing and RAM performance requirements that are far greater than what the product application itself would need. Technology has evolved, though, and the software required to support Wi-Fi has been dramatically reduced by pushing the networking stack, Wi-Fi driver, and connection management further into the RF module itself. The latest WLAN modules simplify integration by providing Wi-Fi connectivity in a single package, with intuitive software that is directly compatible with existing host MCU interfaces.

Why has there historically been a reliance on sizeable operating systems such as Linux to enable Wi-Fi? In the past, embedded system designers had few options other than to integrate

Figure 1. Historical Wi-Fi integration approaches required several components as well as a high-speed ARM core processor.

28

“The software required to support Wi-Fi has been dramatically reduced by pushing the networking stack, Wi-Fi driver, and connection management further into the RF module itself.� high-end microprocessors running complex operating systems in order to utilize software platforms compatible with the WLAN radio drivers provided by the manufacturer (see figure 1).

Those manufacturers often chose Linux due to the large amount of existing support for network interface software. In many cases, though,

TECH REPORT features provided by Linux are “overkill” for simple embedded applications and can lead to unnecessary added costs in terms of software complexity, power consumption, PCB layout, and hardware components.

Wi-Fi Design Freedoms via Fully Integrated Modules Technology for supporting Wi-Fi in embedded product design has progressed dramatically in terms of both affordability and ease of use, thanks to System-on-Chip (SoC) WLAN radio modules with embedded networking stack and applications processors (see figure 2). These fully integrated modules provide enough RAM and processing power to integrate all the software components necessary to provide a full-fledged internet-connected data stream to a microcontroller, without burdening the application software

with running the network stack. This solution brings the additional design benefits of lower power and cost than full-featured processors. The trade-off is in network performance, as Linuxbased Wi-Fi implementations (running on full- featured ~800MHz core processors) typically provide higher throughput than fully embedded module applications (running on sub-100MHz cores). Hence, keep your product application’s network performance requirements in mind when evaluating Wi-Fi design options.

In some cases, systems can also run their application-specific firmware directly on the module’s applications processor, eliminating the reliance on an external host MCU. This translates to improved encapsulation of functionality, bill of materials cost savings, and less design time for adding Wi-Fi connectivity to a product design.

Figure 2. SoC module approach minimizes component count and simplifies software integration.

29

The Benefits of Employing a Wireless Module Making use of an RF module solution to add wireless connectivity such as Wi-Fi provides significant benefits in minimizing your development risk, costs, and timeline. Using an RF module that is certified for the appropriate regulatory entities (such as FCC for the US, IC for Canada, etc.) dramatically simplifies your design and certification efforts.

30

With this type of integrated module solution, the dependence on a large, complex operating system to support WiFi connectivity has become a thing of the past. For example, the new TiWi-C-W™ Wi-Fi module can be easily integrated into an existing product requiring just a small amount of PCB space for the 10.5 x 10.5mm package, a chip antenna or connector (two if diversity is desired) and optionally a serial interface to an existing host MCU. Application-specific firmware can communicate directly with the Wi-Fi driver if developed within the TiWi-C-W applications processor, or by using a simple ASCII-based serial protocol if an external host MCU is used. To extend the convenience of the integrated module design, LSR also provides reference designs for both 2-layer and 4-layer PCBs. In the configuration where a host MCU is used, the TiWi-C-W module offers an innovative, modern approach for radio configuration, establishing a network connection and communicating with a server by using popular communications protocols and data formats. Provisioning requires little interaction from the host MCU when using the built-in ‘Soft AP’ functionality for “Configuration Mode”, allowing a smartphone to connect

TECH REPORT directly with the module and select the Wi-Fi network to join from an interactive web interface served right from the module. Once connected, the module includes features for establishing HTTP/HTTPS connections with native support for RESTful client requests, JSON-RPC, and raw data tunneling.

Bringing it All Together Wi-Fi has become a prevailing networking technology for the Internet, providing an infrastructure to instantly communicate with other networked devices and unlock a whole host of new user experience opportunities for your product. With its pervasiveness across existing networked devices, Wi-Fi brings your product the ability to easily communicate with mobile phones, laptops, and cloud servers, enhancing the capabilities of your product in terms of both user interaction and data collection.

With the latest SoC technologies, the complexity of Wi-Fi has been pushed into the wireless module, allowing developers

to focus on their application data instead of designing a platform to support Wi-Fi. LSR’s TiWi-C-W offers just such an alternative to a full-fledged Linuxbased embedded system with external Wi-Fi radio and other components. For low-to-medium data rate applications, the TiWi-C-W offers a competitive option whether application firmware is integrated directly into the internal ARM Cortex-M3 or used as a communications peripheral via serial port.

In many cases, the effort to embed Wi-Fi capability is a critical first step to creating an Internet-connected product. In those cases, the capabilities and design benefits that TiWi-C-W can provide spans even further. That’s because TiWi-C-W is part of the TiWiConnect™ IoT ecosystem, providing a comprehensive end-to-end solution for connecting a Wi-Fi enabled product to a cloud-server, empowering your product to be remotely monitored and controlled by clients such as mobile apps and web portals.

“Wi-Fi brings your product the ability to easily communicate with mobile phones, laptops and cloud servers.”

31

M o v i n g To w a r d s a

David Elien VP of Marketing & Business Development, Cree, Inc.

Clean Energy

Let There Be

LIGHT

FUTURE

How Cree reinvented the light bulb

— Hugo van Nispen, COO of DNV KEMA

Cutting Edge

SPICE

Modeling

MCU Wars 32-bit MCU Comparison

+

Cutting Edge Flatscreen Technologies

+

New LED Filament Tower

View more EEWeb magazines— Click Here

Power Developer O ct o b er

201 3

From Concept to

Reality

Sierra Circuits:

Designing for

Durability

A Complete PCB Resource

Wolfgang Heinz-Fischer Head of Marketing & PR, TQ-Group

TQ-Group’s Comprehensive Design Process

Freescale and TI Embedded Modules

+

Ken Bahl CEO of Sierra Circuits

PLUS: The “ Ground ” Myth in Printed Circuits

+

+

PCB Resin Reactor

ARM Cortex Programming

Low-Power Design Techniques