One of the hottest and most competitive segments in electronics design and therefore test these days is power. That’s because of the strong demand for energy efficiency to make the most of batteries, help lower energy bills, or to support space-sensitive or heat-sensitive applications.
After 30 years, silicon MOSFET development has approached its theoretical limits. Progress in silicon has slowed to the point where small gains involve significant development cost. Alternative semiconductor materials such as Silicon Carbide (SiC) and Gallium Nitride (GaN) are emerging as the materials of choice. GaN in particular is gaining favor in many areas due to its ability to utilize silicon as a substrate, therefore bringing prices in line with silicon MOSFET. And since it is still young in its life cycle, it will see significant improvement in the years to come.
These new materials provide efficiency not only through faster switching speeds but also lower turn on voltages (Rds On). Of course, any new technology introduces not only a unique set of design challenges, but test and measurement challenges as well. From a test perspective, these materials require test equipment that not only has higher bandwidth, but is also more sensitive. Gone are the days of using a voltage probe out of the box as-is and expecting little signal distortion and loading. In this article, we’ll provide an overview of GaN and then focus on the test challenges.
The market for GaN power semiconductors is forecast to grow from almost zero in 2011 to over $1 billion in 2021, according to a new report from IMS Research(1). The research firm analyzed all of the key end markets for the products and found that power supplies, solar inverters and industrial motor drives would be three main drivers of growth.
The report notes that the speed of GaN transistor developments has accelerated in the last two years. The launch of International Rectifier's "GaNpowIR" and EPC's "eGaN FET" devices started the low voltage market in 2010. And Transphorm's 600V GaN transistors opened the possibility of GaN competing with high voltage MOSFETs and IGBTs.
A key reason for the bullish forecast for GaN is new processes that leverage existing production infrastructure. These fab processes bring down the cost of GaN semiconductors from around 10X that of traditional silicon to a competitive level, especially for applications that require the performance boost. The basic approach is to grow GaN on top of a silicon substrate with an aluminum nitride buffer layer.
EPC’s process (2), for example, begins with inexpensive silicon wafers. A thin layer of aluminum nitride (AlN) is grown on the silicon to isolate the device structure from the substrate. The isolation layer for 200 V and below devices is 300 V. On top of this, a thick layer of resistive GaN is grown. This provides a foundation on which to build the GaN transistor. An electron-generating material is applied to the GaN. This layer creates a quantum strain field with an abundance of free electrons. Further processing forms a depletion region under the gate. To enhance the transistor, a positive voltage is applied to the gate in the same manner as turning on an n-channel, enhancement-mode power MOSFET as shown in Figure 1. This structure is repeated many times to form a power device. The end result is an elegant, cost-effective solution for power switching.
Figure 1. EPC’s GaN leverages existing production infrastructure to bring cost efficiencies.
In terms of applications, the IMS report predicts that GaN will gain traction at first in power supplies where the total system cost savings outweigh the unit price penalty of the device. These include PC and notebook adaptors, servers, etc., and domestic appliances like room air-conditioners, PV microinverters, electric vehicle battery charging and other new applications are likely to adopt GaN power devices in the future.
With their wide bandgap, GaN devices are very attractive for high-temperature applications. For instance, automobile manufacturers are interested in GaN devices for power conversion in hybrid vehicles. In the past, engine designers have used silicon power MOSFETs in these applications, but typically needed to locate electronics far from the engine block due to temperature concerns. Ideally, the power semiconductors would be located nearby for shorter wiring runs, less weight and lower IR losses. GaN devices reportedly can withstand temperatures of up to 300°C and continue to operate efficiently.
In information processing and storage systems, the whole power architecture can be reevaluated to take advantage of the outstanding switching capabilities of GaN materials. As output voltage increases for AC/DC converters, efficiency goes up. As bus voltage increases, transmission efficiency goes up. As frequency increases, size goes down. EPC claims that GaN enables the last stage which enables the first two while increasing AC/DC efficiency when used as synchronous rectifiers. They also allow for intermediate stage converters to be removed for single step conversion, saving the size and cost of the intermediate stage converter.GaN Measurement Challenges
GaN comes a lot closer than silicon ever could to the attributes of the perfect power switch, which would block voltages of infinite magnitude, carry infinite current, switch instantaneously, and require zero drive power. Of course, it doesn’t achieve perfection, but it gets closer than silicon. Overall, GaN delivers higher voltage blocking, a lower on-resistance, and more speed.
In comparison tests, it has been shown that a GaN FET channel can switch in nanoseconds, even while carrying as much as 10 A, with switching frequencies of around 80 MHz. The downside – of course, there is a tradeoff – is that in a GaN switch, current peaks faster and voltage drops correspondingly faster. Not only do GaN devices switch faster, but their turn-on threshold is lower as is the drain-to-source resistance.
In order to take full advantage of the new materials, an oscilloscope will be required to characterize GaN device behavior and to measure losses during transitions. With voltage swings of 600 V or more, GaN devices will require speedy equipment to keep pace. The oscilloscope needs to have sufficient bandwidth to keep pace with the transition and also provide high enough resolution to capture transitions at low voltage levels.
One limiting factor is probes. Currently, the best high-voltage differential probes provide about 200 MHz of bandwidth while measuring signals up to 1.5 kV. Some single-end high-voltage probes can provide 800 MHz of bandwidth (3), enabling them to be use to measure a 600-V swing. Looking down the road, there will be a need for probes with kilovolt-range measurements with bandwidth of more than a gigahertz. Such probes are not currently available.
Another challenge with high voltage probes is ensuring sufficient insulation and clearance between tips without impacting performance. For instance, long leads can result in inductance from circuit loading and ringing. This can make it difficult to determine the real source of a problem. Test equipment manufacturers are using a number of techniques to improve fidelity, such as adding damping resistors to probes.More resolution
High bandwidth and high voltage are usually mutually exclusive. This requires oscilloscopes with high resolutions in order to measure 600-V source-drain voltages along with millivolt-level gate voltages.
The vast majority of oscilloscopes are based on 8-bit resolution ADCs, but this can be significantly increased through used the averaging and high resolution modes.
For signals that are of a repetitive nature, averaging provides an effective way to substantially increase the vertical resolution of the signal (4). This enhancement, measured in bits, is a function of the total number of averages:
Enhanced resolution = 0.5 log2(N)
where: N represents the total number of averages requested
In many oscilloscopes, the averaging algorithm is implemented with fixed-point math. This means that the maximum number of averages is 10,000, which limits the total bits of resolution to an ideal maximum value of 14.64, as shown in Table 1. This approach to averaging maintains the full analog bandwidth of the signal.
Table 1. Enhanced oscilloscope vertical resolution due to averaging.
While averaging is a useful technique for many applications, it won’t work for single-shot acquisitions. In this case, the solution is to use a boxcar averaging technique that calculates and displays the average of all sequential sample values in each sample interval. This mode provides a method for trading off over-sampling for additional information about the waveform. In this case, the additional horizontal sampling information is traded off to provide greater vertical resolution and a reduction of bandwidth and noise.
The bandwidth limiting and the increase in vertical resolution with this averaging technique vary with the maximum sample rate and the selected sample rate of the instrument. The performance increase is shown in Table 2 for an oscilloscope with a maximum sample rate of 10 GS/s. The increase in bits of vertical resolution is:
0.5 log2 * (D)
where: D is the decimation ratio, or the maximum sample rate / actual sample rate
The resulting -3 dB bandwidth (unless further limited by the measurement system’s analog bandwidth) is:
0.44 * SR
where: SR is the actual sample rate
Table 2. Enhanced vertical resolution due to use of boxcar averaging.Summary
With the ongoing push for increased power efficiency, 30-year-old silicon MOSFET has reached its practical performance limits. Now, industry experts are predicting rapid growth for alternatives, most notable GaN due to new fab processes to grow GaN on a silicon substrate using a standard, low-cost CMOS process. This is opening up a range of new commercial and industrial applications for GaN power devices.
With a potent combination of high bandwidth and high voltage, GaN will present significant challenges on the test and measurement front, both in terms of high voltage probes and high-resolution on oscilloscopes. Currently available probes with 2.5 kV and 800 MHz are sufficient for 600V devices while averaging techniques are available to boost resolution. Stay tuned for continued advancements in this important segment.References
(1) "The World Market for Silicon Carbide & Gallium Nitride Power Semiconductors - 2012 edition." http://www.prnewswire.com/news-releases/gan-power-semiconductor-market-to-exceed-1-billion-by-2021-142601706.html
(2) “Master the Fundamentals of Your Gallium-Nitride Power Transistors” http://electronicdesign.com/article/power/master_the_fundamentals_of_your_gallium_nitride_power_transistors
(3) “Passive High-voltage Probes” http://www.tek.com/datasheet/passive-high-voltage-probes-0
(4) “Improving Vertical Resolution in Tektronix Digital Phosphor Oscilloscopes” http://www.tek.com/technical-brief/improving-vertical-resolution-tektronix-dposAbout the Author
Randy White is an embedded applications technical marketing manager at Tektronix. Randy has worked with various aspects of test and measurement solutions at Tektronix over the past few years. He has given seminars on high-speed serial and embedded systems and is actively involved in many working groups for high-speed serial and embedded systems. He holds a BSEE from Oregon State University.
This posting is part of the EDA Designline
power series and is archived and updated. The root is accessible here
. Please send me any updates, additions, references, white papers or other
materials that should be associated with this posting. Thank you for making this a success - Brian Bailey.