Category Archives: Guides

What are Wearable PCBs Made of?

The Internet of Things market is growing at a tremendous speed. Among them, wearables represent a sizeable portion. However, there are no standards governing the small size PCBs or Printed Circuit Boards for these wearables. The unique challenges emerging in these areas require newer board level development and manufacturing experiences. Of these, three areas demand specific attention – surface material of the boards, RF or microwave design and RF transmission lines.

Surface material of the boards

PCB materials are typically composed of laminates. These can be made of FR4, which is actually fiber-reinforced epoxy, of polyamide, Rogers’s materials of laminates, with pre-preg as the insulation between different layers.

It is usual for wearables to demand a high degree of reliability. Although FR4 is the most cost-effective material for fabricating PCBs, reliability is one issue the PCB designer must confront when going for a more expensive or advanced material.

For example, with applications requiring high-speed and high frequency operation, FR4 may not be the best answer. While FR4 has a Dk or dielectric constant of 4.5, the more advanced Rogers series materials can have a Dk of 3.55-3.66. The designer may opt for a stack of multilayer board with FR4 material making up the inner cores and Rogers material on the outer periphery.

You can think of the Dk of a laminate as the capacitance between a pair of conductors on the laminate, as against the same pair of conductors in a vacuum. Since there must be very little loss at high frequencies, the lower Dk of 3.66 for a Rogers’s material is more desirable for high frequency circuits, when compared to FR4, which has a Dk of 4.5.

Typical wearable devices have a layer count between four and eight. With eight layer PCBs, the layer structuring offers enough ground and power planes to sandwich the routing layers. That reduces the ripple effect in crosstalk to a minimum, while significantly lowering the EMI or electromagnetic interference. For RF subsystems, the solid ground plane is necessarily placed right next to the power distribution layer. This arrangement reduces crosstalk and system noise generation to a minimum.

Issues related to fabrication

Tighter impedance control is an important factor for wearable PCBs. This results in cleaner signal propagation. With today’s high frequency, high-speed circuitry, the older standard of +/-10% tolerance no longer holds good and signal-carrying traces are now built to tolerances of +/-7%, +/-5% or even lower. This influences the fabrication of wearable PCBs negatively, as only a limited number of fabrication shops can build such PCBs.

High-frequency material such as Rogers require to have a +/-2% of Dk tolerance and +/-1% is also a common figure. In contrast, for FR4 laminates it is customary to have Dk tolerances of +/-10%. Therefore, Rogers’s material presents far lower insertion losses when compared to FR4 laminates.
In most cases, low cost is an essential factor. Although Rogers’s material offers low-losses with high-frequency performance at reasonable costs, commercial applications commonly use hybrid PCBs with FR4 layers sandwiched between Rogers’s material. For RF/microwave circuits, designers tend to favor the Rogers’s material over FR4 laminates, because of their better high-frequency performance.

What Are IP Markings and IP Ratings?

With so many IP ratings, it is easy to be confused about their actual meaning. However, published by IEC or the International Electro technical Commission, their standard 60529 details all the ratings. IP ratings are also known as Ingress Protection Ratings or International Protection markings. They classify and rate the degree of protection that mechanical casings and electrical enclosures offer against dust, water, accidental contact and intrusion by body parts such as hands and fingers.

Typically, IP ratings indicate the protection level offered by the enclosure of a device. Two or three numbers in the rating indicate the protection level. Within the IP marking, the number in the first position indicates protection from solid objects or materials. The number in the second position indicates protection from liquids, including water. The number in the third position indicates protection from mechanical impacts. However, the third number is commonly omitted as not being a part of IEC 60529. For example, IP65 denotes protection from solid objects to the level of 6 and from liquids to the level of 5. The levels, ranging from 0-9, represent increasing amounts of protection from different solids and liquids, with the level 0 representing no protection at all from either contact or ingress.

For protection from solids, level 1 denotes protection from any object larger than 50mm. This could be any large surface of the body such as the back of a hand, but not offering any protection against a deliberate contact with a body part. Level 2 represents protection against objects of size greater than 12.5mm, but less than 50mm, such as fingers or similar objects. Level 3 represents protection against intrusion from objects of size between 2.5-12.5mm, such as tools, thick wires, etc. Level 4 represents protection against objects of size between 1mm and 2.5mm, such as most wires, screws and so on. A level 5 enclosure will protect against ingress of dust, but not entirely prevent it from entering. That means although the dust protected enclosure will offer complete protection against contact, it will not allow dust to enter in quantities that may interfere with the satisfactory operation of the equipment. If your equipment needs to be dust tight, only an enclosure with a level of 6 will ensure there is no ingress of dust including complete protection against contact.

The second number in the IP rating denotes protection from liquids. For instance, level 1 represents enclosures offering protection from dripping water. Level 2 denotes protection from dripping water when the enclosure is tilted by 15 degrees. An enclosure with a rating of level 3 will not allow the ingress of water sprayed on it, while a level 4 enclosure protects against water being splashed on it. If you want the equipment to remain protected against water from jets, you must go for a level 5 enclosure. If the water jets are really powerful, only a level 6 or above enclosure will help. However, if you expect equipment to work even when immersed in water up to a depth of 1m, you need to go for an enclosure with a rating of level 7. For equipment expected to work even beyond an immersion depth between 1-3m, only an enclosure with a rating of 8 is to be used.

Pi-DAC+ — An Audiophile’s HAT for the Raspberry Pi

Earlier, you may have faced problems with sound cards for your single board computer, the Raspberry Pi (RBPi). It is time to look for a DAC or Digital to Analog Converter that is simple to use and easy to set up to work with your RBPi. The IOAudio HAT fits the bill very well and you can use it to learn your way around the audio capabilities of the RBPi.

The earlier cards for the RBPi had a long series of compiling issues that left their users yearning for a simpler card. The Pi-DAC+ HAT from IOAudio is compatible to RBPi models A+, B+ and RBPi 2. It brings to the RBPi the ability of playing back full-HD audio up to 24-bits/192KHz. Additionally, the HAT is compatible with RuneAudio, Volumio, Moode and many others.

The Pi-DAC+ HAT from IOAudio is fully HAT compliant. It meets all the requirements for the Hardware Added on Top board specifications including the auto-detection by the RBPi. The Pi-DAC+ takes the digital audio signals from the RBPi and passes them through the onboard PCM5122 DAC from Texas Instruments. The output from the DAC is an analog audio signal that can be picked up from the phono connectors onboard the Pi-DAC+. The DAC also consists of a built-in electronic volume control. This eliminates the need for a physical potentiometer based volume control, which is likely to introduce noise in the audio path.

You do not need any soldering to use the Pi-DAC+ HAT with your RBPi. Simply plug it on and you are ready to go. When used for the first time, the Pi-DAC+ requires setting some configuration with the existing setup of the RBPi. If you mess up or are unable to get through, a visit to the manufacturer’s website will give you different pre-configured operating systems for your RBPi. Use them and you will find excellent sound quality from the HAT. The resulting audio output is certainly louder than and clearer than the default audio from the RBPI.

The Pi-DAC+ offers leading audio with a signal to noise ratio of 112dB and a total harmonic distortion of -93dB. The PCM5122 is a 32-bit/384KHz DAC from Texas Instruments. The board has advanced ESD protection to prevent it from handling damages. It requires no external power supply, taking all it wants from the RBPi.

If you do not have an amplifier at present, you can listen to the audio output using a headphone through the 3.5mm audio jack on the board. The board has a built-in high quality audio headphone amplifier, the TPA6133A, also from Texas Instruments. For volume control, RBPi can use ALSA, which gives a full range of control.

If you are an audiophile and an RBPi enthusiast too, the Pi-DAC+ will certainly combine both the worlds for you. You can use raw Linux, RuneAudio, Volumio, SqueezePlug, MDP, AirplaySync or similar on your RBPi and Pi-DAC+ combination for listening to internet radio, streaming music services such as Spotify or your own digital music library, in magnificent audio quality.

Why do Speakers use Ferro-fluids?

Speakers reproduce sound by moving a diaphragm to displace air. The mechanism resembles a permanent magnet electric motor. The major difference is the voice coil in a speaker moves linearly instead of in a circular motion. As the coil moves back and forth in step with the electrical signals fed to it, it moves the attached diaphragm. To prevent spurious movements and unwanted oscillations of the diaphragm, conventional speakers generally use a damper. To produce sound from such speakers, extra energy is necessary to overcome the resistance of the damper.

Additionally, the damper has its own natural frequency of vibration that restricts the speaker from reproducing sound accurately at all frequencies. A new technique using a magnetic fluid to replace the damper claims to correct this anomaly by reducing energy consumption and allowing louder and clearer sound across the entire range of frequencies the speaker is capable of reproducing. To quantify the advantages, the new speaker reduces energy consumption by 35% for reproducing the same loudness of sound as from conventional speakers and the improvement in sound quality is nearly 3dB.

NASA originally developed the magnetic fluid in the 1960’s, using it for space exploration and called it Ferro-fluid. It responds to applied magnetic fields because the fluid is infused with Nano-sized magnetic particles. They do not agglomerate or cluster together because of a coating of suitable surfactants. The unique characteristic of ferro-fluids makes them useful in a range of applications. Using applied magnetic fields to control flow or movement, ferro-fluids can replace mechanical parts such as vehicle suspensions, flow of fuel in a reactor and more.

In a conventional speaker, the damper holds several components such as the diaphragm and spring in place, even when the speaker is vibrating. However, the damper causes friction while moving, thereby distorting the original sound waves with secondary vibrations, which are manifest as noise. To overcome the friction requires additional energy while driving and that reduces the speaker’s total volume output by a few decibels.

When replacing the damper in a speaker, the ferro-fluid used has a thickness of only a few microns. The magnets of the speaker create a permanent magnetic field to which the ferro-fluid responds by holding the diaphragm and the coil in place while allowing them to move linearly without any friction. As there are no secondary vibrations from the ferro-fluid, the sound is clearer. The lack of friction allows the speaker to save about 35% of the energy as compared to conventional speakers with dampers.

Ferro-fluids used for the audio field are usually based on two classes of carrier liquids – synthetic enters and hydrocarbons. Both oils are low in volatility and high on thermal stability. The environmental considerations dictate the choice of the fluid used, along with the best balance of viscosity values and magnetization for optimizing the acoustical performance.

Using different carrier liquids and by varying the quantity of magnetic material in the ferro-fluid, it can be tailored to meet different needs. The saturation magnetization depends on the nature of the suspended magnetic material and its volumetric loading. Care is taken to use material whose density and viscosity correspond closely to that of the carrier fluid.

ARDUINO 101: The Curie-Powered Sensor-Packed Arduino

Intel and Arduino have teamed up to generate a new single board computer, the Arduino 101. Scheduled for market availability in the first quarter of 2016, the Arduino 101 is powered by the Curie module from Intel. Aimed at educating youngsters in the emerging technologies, the SBC is packed with sensors, yet affordably priced.

Arduino 101 has the input and output capabilities of the classic Arduino UNO, but also includes hardware for Bluetooth wireless communication. In addition, Arduino 101 comes with a gyroscope and a 6-axis accelerometer.

Intel and Arduino are promoting their cobranded board for furthering their initiative, Arduino 101 in the Classroom. This is a computer science and design curriculum meant for educating students in the age group 11-14 years in emerging technologies. The Arduino 101 will also be following the hardware configuration of the Curie module. Contestants will be using this board during the upcoming reality television show, America’s Greatest Makers, by the Intel and Turner Broadcasting System.

Those familiar with the Arduino UNO will find Arduino 101 has the same form factor of 70x55x20mm. Differences are an on-board antenna on the bottom right-hand corner of the circuit board and a new main processor. This is the Intel Quark, a low-power 32-bit micro-controller also known as the Curie module. The specialty of this particular Quark is the Bluetooth communication hardware, the gyroscope and the 6-axis accelerometer are on its die.

Users can program the Arduino 101 in the same process they followed for the Arduino UNO. You write your code and compile it with the Arduino IDE, before uploading it to your board. To allow programmers utilize the unique features of the Curie module, Intel is expected to offer special libraries. Initially, Intel had packaged the Curie module in the size of a tiny button and it was supposedly meant for wearable projects. Later, they changed direction towards the Curie-powered Arduino.

Intel is following this go-to-market strategy for its system-on-chips. Intel also packaged an earlier SOC, the Edison. Intel also designed accessory boards for the Edison and Sparkfun produced these boards for Intel. Intel and Arduino had teamed up earlier for the Intel Galileo – the micro-controller board certified by Arduino had Arduino-compatible headers.

The specifications of the Curie indicate it is powered by 1.8V, the popular voltage of a coin-cell battery. However, to power the IO on the Arduino 101 properly, the voltage requirements as dictated by the Arduino ecosystem are at least 3.3V. Limitations imposed by the Arduino 101 design rule out the possibility of a coin-cell battery powering the Curie.

The Curie module also has a 128-node neural network built into it, which users could use for machine-learning applications. However, Intel will not be providing software support for the technology at the time of Arduino 101 launch. They may support it later.

David Cuartielles, the co-founder of Intel’s marketing of Arduino, will be using Arduino 101 in their Creative Technologies in the Classroom or CTC. Earlier, the curriculum used the Arduino UNO for teaching students in a playful way. Now, they will be using the Arduino 101 for teaching basic programming skills in electronics and mechanical design.

What You Need To Know About EMI Antennas

Any electronic device, system or subsystem generates EMI or ElectroMagnetic Interference and is susceptible to EMI generated by others. To allow them to coexist and cooperate, all such electronic devices, systems or subsystems must confirm to specific standards, which limit the amplitude and frequency range of EMI generated and tolerated by each of them.

Testing for such radiated emissions and immunity involves EMI chambers and OATS or Open Area Test Sites. To check for EMI generated, these chambers or OATS will have several types of antenna that can handle a wide range of frequencies. As visits to a full-compliance lab are expensive and time intensive, you may want to do pre-compliance tests, for which, it is a simple matter to set up a temporary antenna in a conference room or basement. This helps in troubleshooting and correcting EMI problems beforehand.

Several factors decide the nature of the antenna you should be using for your tests. The choice for the tests mostly ranges among radiated emissions, radiated immunity, pre-compliance, full compliance, frequency range, power and size of the antennas. The most common EMI test people perform is for checking radiated emissions. Here too, the antenna you use will depend on frequency, size, gain and your budget.

For pre-compliance tests, the most popular antenna is the hybrid. This is also called by names such as Combilog, Biconilog, Bi-log and others. Hybrids are so favored because of their wide frequency range, which easily covers different ranges from 30 MHz to 7 GHz, depending on the model. This is a very big advantage, as you do not need to switch antennas in between the tests, which you have to do if you were using log-periodic or biconical.

For a lab, where precision is more important, using multiple antennas gives an advantage in the performance. Typically, a lab might use a horn antenna for frequencies above 1 GHz, a log-antenna from 1 GHz to 200 MHz and a biconical antenna for frequencies below 200 MHz. However, for pre-compliance tests, hybrids or Bi-log antennas are adequate for makeshift labs.

The size of the antenna you can use depends on the space you have in your makeshift lab. Larger antennas cover a wider frequency range along with better sensitivities as compared to those offered by smaller antennas. Some designs of hybrid antennas come with bent elements, which help to fit them in limited spaces. In general, hybrid antennas are larger than most dedicated antennas.

Antennas are available that allow you to use them for both radiated immunity as well as radiated emission tests. However, for immunity tests, it is important to limit the power you drive into an antenna to get the required field strength. Typically, immunity testing requires larger antenna sizes as compared to those necessary for measurements of emissions alone.

Hybrid antennas usually combine a log-periodic element with a biconical element. This extends the frequency range the antenna covers as compared with that covered by single-type antenna. For example, one of the newest hybrid antennas covers the entire range of 26 MHz to 3 GHz, while being able to handle signal power up to 300 W for immunity tests.

What is Digital Signal Processing?

Initially, when DSP or Digital Signal Processing was introduced over thirty years ago, it involved standalone processing. A single micro-controller handled all the parameters for processing the analog signal and transforming it to its digital value. Evolution in this area has introduced multicore processing elements that now extend the DSP’s range of applications.

Simultaneously, evolvement of software development tools for the DSP now allows expansion for accommodating diverse programmers. Therefore, on one hand you can have voice and image recognition with small, low power, but smart devices, while on the other, it is possible to have real-time data analytics with the multiple core high-performance compute platforms. This way, DSPs offer nearly endless opportunities for achieving low-power processing efficiencies.

Although initial DSPs processed only audio, engineers quickly adapted DSP technology for a wide variety of applications. Today, DSPs are available as standalone or as part of an SoC or System-on-Chip offering full software programmability including all the benefits of software-based products.

DSPs take already digitized signals from the real world, such as audio, video, pressure, temperature or position for further mathematical manipulations. Engineers design DSPs for performing quick mathematical operations such as add, subtract, multiply and divide.

This processing of the signals enables displaying, analyzing or converting information to a signal of another type to be useful. In the real world, several analog products are available to detect and manipulate signals such as pressure, temperature, light or sound. These signals are then passed on to converters such as ADCs or Analog to Digital Converters, which transform the analog signals into a digital format of 1’s and 0’s.

The DSP takes over this stream of digitized information and processes it further. The processed digital information goes back for use in the real world. The DSP does this in one of two ways. It feeds the information in the digital format to instruments capable of handling it. Where that is not possible, the digital signal passes through a second converter or DAC, the Digital to Analog Converter and this converts the digital signal to analog. All this happens at very high speeds.

An MP3 player is a very simple illustration of the concept of DSP. The analog audio, during the recording phase, passes through a receiver containing a microphone and an amplifier. An ADC then converts this analog signal into digital information, before passing it over to a DSP. The DSP processes the digital signal further as defined by its internal algorithm and encodes it as MP3, before saving the file to memory.

While playing back the recorded information, the DSP decodes the file from memory and a DAC converts the digital signal to an analog form. That makes it suitable to output the signal through an amplifier and speaker system. If necessary, the DSP handles other functions such as level control and equalization including user interfacing.

A computer can also use information from a DSP. The computer can use this information to control security, home theater systems, telephones and for compressing video. Compressed signals are more efficient when transmitting. Additionally, the computer can easily manipulate or enhance the signals to improve their quality.

Predict Solar Eclipses with Wolfram on Raspberry Pi

Wolfram Research shows how the Wolfram language, used on a Raspberry Pi or RBPi, can help visualize solar eclipses. With this combination, you can view past and present solar eclipses. The most astounding aspect is the solar eclipses you visualize can be not only total or partial, but also as if seen from Earth, Mars or Jupiter.

Depending on your present geographical location, you may or may not be able to witness a solar eclipse. To recapitulate, solar eclipses are events where the Moon blots out the Sun to observers on the Earth. The Moon may be so positioned it blocks out the entire Sun or a part of it. If the Moon blocks out a part of the Sun, the incident is termed a partial eclipse. In a total eclipse, an observer on the Earth will only see the corona of the Sun as a halo around the Moon as it covers the Sun entirely.

By mathematically tracking heavenly bodies, it is possible to predict when a solar eclipse is likely, if it will be visible from a specific location and whether it will be partial or total. Usually, the media drums up a small hype of the event, predicting local weather conditions, telling people how and when to observe the eclipse while including other relevant details. However, this is only if the eclipse is visible in your area.

For people on the Wolfram Community, geographical hurdles do not exist. Novices, experienced users and developers from all over the world share data and knowledge. The Community discusses the latest solar eclipse with anticipation, observation and data analysis. They also participate in the computations for future and extraterrestrial eclipses.

For example, consider the total solar eclipse that occurred on March 20, 2015. Before the event, Jeff Bryant and Fransisco Rodriguez from Wolfram explained how the community could compute the geographical locations from where the eclipse would be totally or partially visible. Fransisco used GeoEntities to highlight with green those countries that would witness at least partial solar eclipse on the date.

Although they predicted the visibility of the solar eclipse, neither Jeff or Fransisco was able to see even the partial solar eclipse, as the former is in the US and the latter in Peru. In their prediction, the intense red area shows the regions from where the total eclipse would be visible, while the lighter red areas depict regions of visibility of the partial eclipse. Another total solar eclipse is predicted in the next decade, of which, at least a partial phase will be visible from almost all countries of the world.

Wolfram now has a new language function, the TimeLinePlot. This is a great way to visualize a chronological event such as a solar eclipse. With TimeLinePlot, you can specify the last few years and the next few years to plot territories and countries from where a total solar eclipse will be visible. TimeLinePlot complies with ISO 3166-1 when depicting territories and countries. Using the incredible powers of computational info-graphics, Wolfram predicts a spectacular total solar eclipse spanning the US from coast to coast on August 21, 2017.

How Gesture Sensors are Revolutionizing User Interface

Imagine a scenario where you control almost everything by simply waving your arms and not by punch any buttons or touching a screen. Welcome to the complicated world of gesture controls. Mechanical buttons and switches are subject to the risk of reliability – they also need protection from the environment. When replaced with electrical controls, such as resistive or capacitive displays and buttons, these do bypass the problems faced by mechanical switches. However, to operate, they still need the physical touch of the operator.

By using optical sensors, it is easy to avoid the reliability risk, mechanical complexity and the requirement for physical touch. You can find optical sensors being used as proximity detectors in many applications such as in water and soap dispensers. Apart from the ease of operation, optical sensors provide the primary potential in recognizing user gestures, thereby reducing system complexity and enhancing user functionality. Today, gesture sensors have evolved to revolutionize user interface controls. They offer the ideal combination of functionality, performance and ease of implementation.

For instance, gesture sensor TMG3992 and others offer simple digital interfaces and do not demand significant processing or memory bandwidth to operate. Being interrupt driven, such sensors interact with the system only when they encounter a recognized event. Simple electrical and software designs are enough to implement two and four direction gesture sensing applications. The sensors work easily from behind plastic or glass transparent to infrared light. That means there is no added complexity or reliability risk in incorporating gesture sensors in electronic devices, as most use plastic housings transparent to infrared.

Gesture sensors help the industry in myriad ways. For example, heavy industries use gloves that limit options for user interface. Operators need specialty gloves to operate most capacitive touchscreens, as they do not respond to commonly used gloves. On the other hand, there are no restrictions for gesture sensors to operate with any type of gloves.

Gesture sensors are eminently suitable for recreational applications such as cold weather or aerial sports and industrial applications such as clean room manufacturing, chemical industries and construction. For example, a skier may keep his or her hands warm within gloves and yet operate a smartphone or manipulate a self-mounted camera with ease.

Touchscreens do not work in environments under water. However, divers can make full use of gesture sensors. It is true water attenuates infrared light and restricts the working distance, so you need additional power. However, multiple benefits overcome this minor restriction. Using gesture sensors such as the TMG3992 and similar greatly simplifies the user interface as underwater cameras can do the job, while the TMG3992 replaces several mechanical buttons and switches for a smaller and more reliable interface.

Smartphone designers and manufacturers already include several user interface options offering multiple solutions for different tasks. However, in many situations – such as exercising or cooking – it is inconvenient to touch the phone while performing the tasks. Gesture controls provide the user different ways of interacting with the phone – such as when checking notifications and scrolling through them. For example, the user can identify a caller and select from a variety of options – answer the call with the speaker enabled, ignore the call without a response or ignore the call with a pre-defined text message.

Synthetic Diamond Manages Power

Power delivery using semiconductor devices is increasing at a rapid pace. This is evident from different forms of power delivered, whether it is controlled power through power inverters, or RF power through amplifiers. Power is necessary to operate nearly everything, such as for alternative forms of energy generation, electric vehicles, radar systems, cellular base stations and even smartphones. However, semiconductor devices need to dissipate the heat they generate, and this poses a stringent challenge for power and thermal management.

Such high-power semiconductor devices are now using a new technology in the form of GaN-on-diamond wafers and synthetic diamond heat-spreaders. The reason behind this is the excellent thermal conductivity of diamond, the highest of any material. At room temperatures, diamond conducts heat about five times better than copper does.

Any semiconductor material can use diamond heat spreaders and these lower the temperature of the semiconductor gate junction by almost 30 percent. In addition, the use of GaN-on-diamond wafers helps lower the temperatures of GaN devices further. With the gate-junction temperature going down by almost 50 percent, GaN-on-diamond devices can handle more than three times the power density than similar GaN-on-SiC can.

Manufacturers use the technique of plasma-assisted microwave CVD or Chemical Vapor Deposition for synthesizing diamond heat spreaders. With this method of growing the synthetic diamond, manufacturers make freestanding diamond wafers up to 140 mm in diameter and nearly 1 mm thick. The wafers have thermal conductivities higher than 2000W/mK, which is five times that of copper. By using microwave CVD for growing diamonds, manufacturers can engineer the properties of the diamond wafers precisely, giving them a range of thermal conductivities. This allows them to offer different cost to performance ratios for matching the specific needs of any application.

When using metalized diamond heat spreaders, manufacturers attach them to the bottom of the semiconductor die. Since they use as thin a layer of solder as is possible for attaching the heat spreader, the diamond lies within 100 to 300 microns of the gate junctions of the device. The diamond heat spreader distributes the heat equally and effectively in both lateral and vertical directions. Heat spreading in the lateral direction is particularly important for RF power amplifiers, as they typically form hot spots of up to 1 micron in diameter with intense heat density.

Manufacturers need to keep the metallization of the die and the heat spreader thin – to the extent of a few hundreds of nanometers. Metallization of the diamond has to be done carefully using a carbide-forming metal as the first layer. The solder layer used to attach the heat spreader must also be thin, preferably lower than 10 microns. With optimal integration into a package, diamond heat spreaders typically help to reduce the gate junction temperatures by nearly 30 percent, when compared to what ceramic packages do that are not using diamond heat spreaders.

The GaN-on-diamond substrates now offer new thermal management tools for GaN semiconductor devices. The reduced thermal resistance of GaN-on-diamond and diamond heat spreaders allows simpler, less expensive thermal management systems. This has a favorable impact on cooling complexity and expenses involved, also leading to better lifetimes of the entire system.