October 2015
Overcoming Today’s
TEST
Challenges Embedded Serial Bus Debugging Analyzing High Sampling Rates SMUs for FET Characterization
Join Today CLICK HERE
eeweb.com/register
CONTENTS
Modern Test & Measure
EDITORIAL STAFF Content Editor Alex Maddalena amaddalena@aspencore.com Digital Content Manager Heather Hamilton hhamilton@aspencore.com Tel | 208-639-6485 Global Creative Director Nicolas Perner nperner@aspencore.com Graphic Designer Carol Smiley csmiley@aspencore.com Audience Development Claire Hellar chellar@aspencore.com Register at EEWeb http://www.eeweb.com/register/
Published by AspenCore 950 West Bannock Suite 450 Boise, Idaho 83702 Tel | 208-639-6464
4 8
TECH SERIES Why High Samplng Rates Matter Test Challenges for PAM4 Signals
TECH REPORTS
12 26
Embedded Serial Bus Debugging for Oscilloscopes Torque Monitoring Improves Predictive Maintenance Strategies
EEWEB FEATURE
34
SMU Instruments Offer Alternatives for FET Characterization
8
12
Victor Alejandro Gao General Manager Executive Publisher Cody Miller Global Media Director Group Publisher
24 EEWeb
34
26
Modern Test & Measure
4
TECH SERIES
Why High
SAMPLING Rates Matter By David Maliniak, Teledyne LeCroy
A key to accurate measurements with an oscilloscope is to ensure that the instrument maintains a high sampling rate. This applies to most measurements; conversely, for many measurements, accuracy may suffer as sample rate decreases. In the worst case, some signal components may be “aliased,� meaning that the true signal shape is corrupted by the addition of bogus signal components that arise from undersampling of real signal components.
5
Modern Test & Measure
According to Nyquist’s Theorem, the sampling rate of a digital oscilloscope must be at least twice the speed of the signal’s highest-frequency content. Any signal energy at frequencies higher than Nyquist will be undersampled and aliased (Figure 1). In general, an oscilloscope’s sampling rate may change with the number of channels in use and the time/div setting. At long signal capture times, the sampling rate is reduced so that the acquisition
memory can cover the elapsed signal time. At lower sampling rates, a highbandwidth front-end amplifier turns out to be disadvantageous, because it will send lots of high-frequency content in the signal to the ADC, where those frequencies will suffer aliasing. Say you have three oscilloscopes in your lab. All three may have the same specified bandwidth and sampling rate. But when capturing longer signals, the sampling rates could be very different depending on how much acquisition
Figure 1.
According to Nyquist’s Theorem, the sampling rate of a digital oscilloscope must be at least twice the speed of the signal’s highest-frequency content.
6
TECH SERIES When capturing signals of long duration, be sure that you’re aware of your actual sampling rate. memory each instrument contains. When capturing signals of long duration, such as those from power devices or serial data streams, be sure that you’re aware of your actual sampling rate (which may not be the instrument’s maximum specified rate). Also, make sure that you enter the oscilloscope’s menu of horizontal settings and select the maximum
amount of acquisition memory. Some oscilloscopes also allow you to specify whether all channels are in use, or only one or two. If only two are in use, some instruments can add the memory from unused channels to the active channel(s), thereby increasing the amount of time at which a signal can be captured at that maximum sampling rate.
www.siglentamerica.com
Modern Test & Measure
8
TECH SERIES
TEST
Challenges
for PAM4 Signals By David Maliniak, Teledyne LeCroy
PAM4 encoding offers the advantage of doubling the bit rate in a serial data channel, doing so by increasing the number of voltage levels from two to four. It’s a fairly complex modulation scheme, so it should be no surprise that it presents some test and measurement challenges.
9
Modern Test & Measure
Imagine, if you will, a plain-vanilla NRZ eye diagram, and an ideal one at that. It would have no jitter, just a finite rise and fall time. The first thing that often comes up in discussions of PAM4 signals is: How does one recover a clock? In the case of an NRZ signal, the purpose of clock recovery is to pinpoint the crossover point, or the place at which the signal crosses the threshold (see the green circle at center left of the figure). With a PAM4 signal, it’s a similar situation except for the fact that there are four voltage levels instead of two. The two intermediary levels mean a lot more signal transitions: there’s the one from the bottom level to the next one higher, and from that level to the third, and from the third to the top (blue traces c, b, and a, respectively, in the figure). The blue trace b crosses at the same place as the black traces that cover the full amplitude range (at the green circle). Some transitions do not cross any threshold at all (blue traces in top and bottom thirds), but there are likely not as many of those as there are transitions that do cross thresholds. So considering these transitions, one may assume that traditional clock recovery would be adequate.
10
But then there are the transitions from the bottom level to the second level up, and from the top level to the second level down (yellow traces). Even in a perfect world, these transitions do not cross the threshold at the same place as the others (green circle). Thus, the question is whether these transitions would affect clock recovery? Will they result in clock jitter? The answer is “not really.” If we assume roughly equal numbers of transitions to the left and right of the green circle, and many more transitions than either at the green circle, then the two yellow crossings to the left and right cancel each other out. Traditional clock recovery works fine. Speaking of jitter, the proximity of those yellow crossings to the idealized crossing location at the green circle does raise questions. Referring to the small red arrows at the right side of the figure, if one is accustomed to analyzing NRZ signals, those gaps in time do indeed look like jitter. How does one analyze jitter on a PAM4 signal? First, let’s not lose sight of the ultimate purpose of analyzing jitter at all, which is to determine our predicted bit-error ratio. We want to know where we need to sample the signal to minimize bit errors (ideally, at the red crosses at
TECH SERIES
the center of each of the three eyes). Thus, what we really care about is the eye width and the position of the sampler in the receiver. The number of trajectories that build up the eye aren’t the issue, but rather the eye width at a given bit-error ratio as opposed to an in-depth analysis of what’s going on at the crossings. Another challenge in the PAM4 realm is that of noise tolerance. Instead of having the full amplitude range, we have 33% of that amplitude to work with. Thus, noise analysis becomes much more important. Just as we want to understand the eye width at a given BER to know where best to situate the sampler horizontally, we also need to know where to place it vertically. Eye height at a given BER is another critical parameter. Lastly, somewhat new for PAM4 in contrast to NRZ is the notion of linearity. In NRZ, there are only the two amplitude levels to worry about while PAM4 gives us these two intermediary amplitude levels. Are the distances a, b, and c the same so that the levels are evenly spaced and we have maximum openings for all three eyes? Does a = b = c? That’s what is meant by “linearity” in the context of PAM4. That’s a quick rundown of the test and measurement challenges raised by the advent of PAM4 signaling.
11
MYLINK
MYLINK
Modern Test & Measure
Embedded Serial Bus
DEBUGGING using Oscilloscopes
By Alex Klimaj Keysight Technologies, Inc.
14
TECH REPORT
I
n the electronics industry, the drive to reduce size, power, and cost of products is leading to an increased use of serial buses in embedded devices. While serial buses reduce pin counts, power consumption, and space required for data transmission, they also lead to greater design complexity. Engineers working with serial buses need a quick way to debug and validate their designs. The asynchronous nature of many serial standards presents a challenge to capture and decode. While protocol analyzers are great tools for finding functional and timing issues on serial buses, oscilloscopes also allow engineers to isolate sources of noise, capture transients, perform physical layer compliance tests, and measure power use. With built in serial trigger and decode, oscilloscopes become powerful all-in-one tools for embedded hardware engineers.
15
Modern Test & Measure
Capture and Decode Before serial decoders were built into oscilloscopes, engineers would have to manually decode the serial frames by hand. This required a deep understanding of the serial bus specification and some luck when trying to capture a specific frame. In Figure 1, I have captured one I2C frame on an oscilloscope with SCL (Serial Clock Line) on channel 1 and SDA (Serial Data Line) channel 2. To decode this by hand I had to use the following procedure. 1. Capture one frame on the oscilloscope. 2. Take a screenshot. 3. Open it on a computer and add lines on each rising edge of the clock. 4. Convert the SDA signal at each rising SCL edge to a 1 or 0 corresponding to the level of the signal. 5. Based on the I C specification, convert the binary values to hexadecimal. By counting the clock edges and knowing that the I2C specification uses a start pulse, 7 address bits, a read or write bit, acknowledge bit, 8 data bits, another acknowledge, and finally a stop bit, I can figure out the hexadecimal values. In this case, this is a write frame with address 0x29, and 0x04 data. 2
16
This method works, but is time consuming and there is a possibility that it’s not even the frame of interest. If an engineer needs to decode multiple serial frames, the required time becomes costly. This is where the benefit having serial decode and trigger built into oscilloscopes becomes obvious. In Figure 2, I’ve turned on the built-in I2C decoder in the oscilloscope. Additionally, I’ve used the serial trigger capability to quickly find this exact packet of data. You can see the decoded frame in blue across the bottom of the oscilloscope graticule and the orange trigger point on top. Engineers can use this capability to quickly find the exact data of interest on the bus without needing an intimate knowledge of the serial specification.
With built in serial trigger and decode, oscilloscopes become powerful all-in-one tools for embedded hardware engineers.
TECH REPORT
Figure 1. Manual I2C Decode
Figure 2. Automatic I2C Trigger and Decode
17
Modern Test & Measure
Using Segmented Memory to Capture Infrequent Serial Events Let’s say a design has multiple ICs (integrated circuits) that communicate with a microcontroller via I2C and I want to see all the data that gets written to one specific IC. This can be done easily by setting up a trigger on the IC’s write address. This will set the scope to trigger anytime it sees that address, but it’s still difficult to get an overall view of the transmitted data from the single waveform display that oscilloscopes default to. This is a situation where segmented memory is very useful. Segmented memory will tell the oscilloscope to wait until the trigger condition is met, save the waveform to memory, time tag it, then rearm the trigger and wait for the next occurrence of the trigger. In Figure 3, I’ve set the oscilloscope to trigger every time data is written to the IC with address hex 64 and turned on segmented memory to capture 1000 occurrences. Using this tool, an engineer can see what data is being written to the IC, how often it is being written to, and view the physical signal for every frame. If one or more of the frames had unexpected data, the engineer could select that frame in the lister and check the integrity of the physical signal. Noise or a transient could have caused the error.
Isolating Coupled Transients Infrequent transients coupled onto a serial bus could cause errors or randomly reset an embedded processor. Using traditional oscilloscope triggering methods, they can be very hard to isolate. The presence of a digital serial signal, already with infrequent events, makes it even harder to trigger on the even more infrequent transient. One way to identify if a transient signal is present is to use infinite persistence. The infinite persistence setting will not clear the display between signal captures. The display will keep updating and overlaying the new signal on top of the previous signals. Using a serial trigger to show all transmitted frames and infinite persistence should result in a display with every bit position filled—if a transient is present, you will quickly notice it stand out. Once you’ve identified that a transient is present, you’ll want to trigger on it. Like I’ve said previously, it can be very tricky to capture using traditional triggering methods. With the arrival of touchscreen oscilloscopes, a new triggering method became possible; using a feature called Zone Trigger, capturing transients becomes trivial. Figure 4 shows a USB (Universal Serial Bus) signal with an infrequent transient.
Segmented memory will tell the oscilloscope to wait until the trigger condition is met, save the waveform to memory, time tag it, then rearm the trigger and wait for the next occurrence of the trigger.
18
TECH REPORT
Figure 3. Serial Decode with Segmented Memory
Figure 4. USB Signal with Coupled Transient
19
Modern Test & Measure Figure 5. Isolating a Transient using Zone Trigger
Figure 6. USB with Low Frequency Noise
Figure 7. FFT on USB Signal with Coupled Low Frequency Noise
20
TECH REPORT To trigger on this transient, all I had to do was draw a box on the touchscreen and select Zone 1 Must Intersect. The oscilloscope was already triggering on any USB start of frame. Turning on Zone Trigger told the scope to only trigger when it saw a USB start of frame and a signal inside the designated area, which can be seen in Figure 5. Once you’ve successfully triggered on the transient, you can perform measurements on it in order to figure out the source.
Finding Noise Sources What if your serial bus is not experiencing random transients, but constant coupled noise instead? To figure out the source of the noise, a different approach must be used since you can’t trigger on noise. This is a great use of the oscilloscope’s FFT (Fast Fourier Transform) math function. The FFT converts the time domain signal to a frequency domain representation. This enables you to identify the frequency components of the coupled noise. Figure 6 shows a USB signal with coupled noise. Just from this screenshot of one USB start of frame packet, I know that the noise has about a 150mV peak-topeak voltage. In order to figure out the source of this noise, I use the FFT function and turn on peak markers. When I first turned on the FFT, I did not see a frequency component that stood out as the noise. This is an indication that the noise is a much lower frequency than the USB signal.
21
Modern Test & Measure
In order to find the main frequency component, I had to zoom out to 10ms per division. At this time base, the FFT is able to detect low frequencies. In Figure 7, you can see that the noise is about 61Hz. This is actually noise coupled onto the USB signal from an AC power line in the United States. Another common source of noise on embedded serial buses can be from clock signals running in the megahertz frequencies. Figure 8 shows an FFT running on a USB signal that has noise coupled from a 16MHz clock. Internet of things devices with built in wireless LAN may also experience noise from the radio. Figure 9 demonstrates a USB signal with a frequency peak at 2.4GHz, a common wireless LAN frequency.
Summary Embedded hardware engineers are constantly challenged to create smaller, more power efficient, and lower cost devices. With this comes a greater need for more powerful debug and measurement solutions. Modern digital storage oscilloscopes with built in serial trigger and decode make excellent tools for the job. Decoding individual frames by hand is a thing of the past with built in decoding. Advanced serial triggering has made capturing exact frames of interest easier than ever before. Segmented memory can be combined with serial trigger and decode to capture many specific data transmissions over a long time period. New triggering techniques made possible by touchscreens make it trivial to isolate infrequent transients. Finally, sources of noise can be identified by using the oscilloscope’s FFT function. Embedded engineers have an entire toolbox of debug tools available when they have a modern oscilloscope on their bench.
22
TECH REPORT
Figure 8. USB Signal with 16MHz Clock Noise
Figure 9. USB Signal with 2.4GHz Wireless LAN Noise
23
Your Circuit Starts Here. Sign up to design, share, and collaborate on your next project—big or small.
Schematics.com Click Here to Sign Up
NEW
The Most Accurate High Power DC Loads 3kW to 24kW power ranges - up to 480kW in parallel 150V, 600V and 1200V voltage ranges
Increased Features + Sine wave loading function + User deened waveforms (UDW) + Auto frequency dynamic sweep to 50kHz + Digitizing function to record transients + Complex operating modes
Modes
Remote Operation Softpanels
CC CR CV CP CZ
USB Ethernet LXI GPIB
New Front Panel Design Highest Power Density
6kW @4U APPS
Front panel tilts for easy access and operation on 7,10,13U models Also includes Hot Keys to recall commands
Designed for testing server power supplies, A/D power supplies, batteries and energy storage systems, EV/EVSE, solar, and other power electronics.
Learn More - chromausa.com click here for link
Modern Test & Measure
TORQUE Monitoring
Improves Process and Predictive Maintenance Strategies
By Alan Lowne – CEO, Saelig Co. Inc.
26
TECH REPORT
Effective non-contact torque monitoring can help production quality as well as identify machine problems before they happen. The importance of torque measurement in manufacturing environments is a new concept to some, but an everyday essential to others. Realizing the enormous cost benefits of measuring torque in rotating systems is sometimes not recognized by those tasked with improving profitability. The challenge is to be able to monitor and measure torque as accurately, unobtrusively, and economically as possible. For continuous-manufacturing processes where machines are driven by rotating shafts, machinery failure and subsequent downtime must be avoided in order to maintain profitability as well as consistency of output. The effective use of precision non-contact torque monitoring instrumentation can preemptively identify problems that might affect machinery reliability—extremely important for situations where a single machine failure can lead to costly production losses.
27
Modern Test & Measure
The level of motor torque can be an indicator of the quantity, speed, or viscosity of process material being mixed.
28
Installing torque sensors in the rotating drive train of operating machinery can track some of smallest changes in a plant’s operating parameters, and diagnose a change in state or viscosity of materials being handled, or the condition of the machinery itself. For instance, measured torque increases on a mixer drive may suggest that contents of a food preparation are thickening exactly as expected for a consistent product result, or that a seal or bearing is sticking and may soon fail. So it can be seen that data collected from a torque sensor can be vital for a control computer to accurately capture and assess manufacturing situations. A process engineer’s job is to be responsible for material mixes being transformed from one condition into another, monitoring both consistent and changing variables at all the various stages of the process. Some parameters can be measured directly and simply, such as temperature, fluid levels, volumes, and others. Other factors, such as chemical bonding reactions or process thickening, can be more challenging to measure. Instead of directly measuring these more difficult parameters, an oftenused technique is to instead measure a related parameter—typically one related to the plant or machinery rather than the process material itself—and to infer from this the desired parameter readings. For instance, viscosity may change when a desired chemical reaction is completed.
Significantly, many types of processing equipment—mixers, pumps, conveyors— are motor driven, and measuring motor output characteristics or current consumption will often yield vital process information. For instance, the level of motor torque can be an indicator of the quantity, speed, or viscosity of process material being mixed. Electric motor torque monitoring can be an effective strategy for detecting worn bearings or over-tight shafts on equipment such as fans or blowers, the malfunction of which can significantly increase a plant’s overall operating costs. Torque measurement provides an important set of data on machinery performance and condition monitoring. Knowledge of what data parameters to evaluate will provide proactive or early warning of breakdowns, allowing plant operators to schedule appropriate pre-emptive or predictive maintenance, allowing critical machinery to reliably and continuously run with minimal downtime. Today’s torque measurement techniques have become increasingly simpler and user-friendly. Formerly, torque sensors required a fairly complicated and fragile array of slip rings connected to the rotating drive shaft of a machine under test. UK-based Sensor Technology has incorporated surface wave technology in their non-contact method of torque monitoring, known as TorqSense™. The product requires adhesion of several
TECH REPORT pads on to the side of the driveshaft, supported by an accompanying electronics unit mounted nearby, which monitors the torque, sending readings as a data signal to the control system. These pads are a series of tiny piezoelectric combs, fully encased in plastic. The combs are designed to open or close under torque effects of drive shaft rotational speed. A low-power radio frequency signal is emitted toward the combs, feeding back signals to unit’s integral internal electronics. The reflected signal is returned at a different frequency, with the change proportional to the distortion of the combs, and hence also proportional to drive shaft torque.
Diverse Non-Contact Torque Monitoring Applications Regenerative Energy/Windmills and Nuclear Industries Precision gearboxes supplied to the regenerative energy and nuclear industries must be 100% guaranteed to reliably operate without premature failure, making off-line testing vital. Certain offline test rigs consist of a motor driving a test unit against a load created by an industrial disc brake. Testing generates a performance profile that can be compared with an ideal performance standard. In the
29
Modern Test & Measure
case of an equipment failure within the nuclear industry, a component or system breakdown could mean longer term shutdown of critical operations, automated/unmanned removal of the faulty parts, sealing into a secure flask and automated replacement installation, potentially costing millions of dollars.
A precision torque sensor actually allows a design engineer to measure torque change at the exact moment when a drive is ideally switched from power to regeneration, to make the most of the falling load potential energy release.
30
Chemical Processing, Plastics and Pharmaceuticals Implementing a precision torque sensor in recipe mixing applications can reduce development costs in the chemical, food and plastics industries, and help nanotechnology advances with in pharmaceuticals. Incorporating instruments, which monitor the properties of materials as they are being mixed, capturing torque output to a PC for analysis and real-time control, allows instantaneous reaction to mixer conditions. Often a key parameter is the mix torque itself, which can settle at a constant level once mixing is complete. During process development, many batches may be required before a “recipe� is finalized, so the cost and time involved can be considerable. It is easy to see how quickly torque equipment
costs can be recouped by enabling reproducible experiment variations to be constructed, as well as minimizing a resultant consistent process time. Appliances: Washing Goes Green High performance non-contact torque measurement can also improve the energy efficiency of industrial and domestic washing machines. Process plant manufacturers are redesigning machines to reduce power consumption. In horizontal-axis, front-loading washers, the load (wet laundry) is lifted onto one side of the axis, and falls on the other side. In this scenario, regenerative energy recovery is desirable, if it can be practically achieved. A critical element in designing such systems is the capability to make continuous, non-invasive, accurate torque measurements. A precision torque sensor actually allows a design engineer to measure torque change at the exact moment when a drive is ideally switched from power to regeneration, to make the most of the falling load potential energy release. Since a drive motor within this application could be rotating at up to 1500rpm, accurate data collection and
TECH REPORT
equally responsive control software are essential, and may be built into next-generation washing machines. For industrial-sized loads, torque sensors can assist appliance manufacturers in achieving energy savings of up to 20-30 percent, providing a significant competitive edge in the marketplace. Aerospace and Defense: Unmanned Aerial Vehicle Design The development of low-cost, flexible search and surveillance unmanned aerial vehicles (UAV) for military, homeland security and environmental monitoring can benefit from torque monitoring. Vectored-thrust propulsion system developers are using torque measurement technology as a key technology for the design and implementation of propulsion systems. Hybrid Vehicle Design Vehicle designers around the world know that torque measurement in invaluable in mapping the characteristics of combined motor-generator developments in hybrid cars designed to reduce vehicle emissions. The benefits provided by non-contact sensing technology include
easy data capture, intelligent calibration, and advanced analysis capabilities. Intelligent Lubrication Car designers are constantly working on competitive improvements in their vehicles. Intelligent lubrication systems analyze engine friction and parasitic losses to provide optimal lubrication. Accurate and repeatable measurement of small changes in drive torque is a critical requirement during the development phase. Controlled oil pumps, each capable of supplying individual engine parts with oil under conditions unique to that part of the engine, react to signals that sense the vehicle operating conditions. For example, the engine head may be fed with oil at pressures different to the engine block, and bearings may need more oil when an engine is under heavy load. Profiling engine performance under various lubrication conditions can derive optimum performance configurations for future intelligent engine systems. Both gas and diesel engines run much cleaner than years ago. The required need for efficient operation under a wide range speeds and loads, and environmental conditions from -40 to +40 째C, poses a huge challenge.
Intelligent lubrication systems analyze engine friction and parasitic losses to provide optimal lubrication.
31
Modern Test & Measure
The benefits provided by non-contact sensing technology include easy data capture, intelligent calibration, and advanced analysis capabilities.
32
Intelligent lubrication has the potential to improve performance enormously, but quantifying the best configuration is a lengthy and complex task. Torque sensing is critical to measuring friction effects of a range of different oil supply and formulation strategies. Research and Development/ Test Lab Applications With regard to research experiments, there is often an inordinate amount of dismantling and reassembling of test equipment. This can be time-consuming and therefore expensive, but choosing a non-contact torque sensing method is ideal for this kind of application since it does not need to be dismantled. A non-contact coupling between the shaft and the controller eliminates any issues of mechanical compliance. Easily embedded within servo drive systems, the best torque transducers can withstand heat, dirt and mechanical vibration. The potential of servo drive trains, for instance, that are ‘intelligently rigid’ and so free from torsional losses could result in commercial servo products that deliver improved performance and vastly superior system dynamics with even the most demanding mechanical loads. Ongoing research around
the world is focused on developing control algorithms that improve the efficiency of servo controllers that maximize the use of torque feedback. Conclusion The application of precision torque sensing presents many opportunities for developing improved efficiencies in performance, production and maintenance in a wide variety of industries. Optimal sensing transducers require minimal shaft length, have low inertia, no physical contact between shaft and housing, while offering wide bandwidth, high resolution, high accuracy and good magnetic/ RF noise immunity. And they pay for themselves rather quickly!
For more information, please contact: Alan Lowne Saelig Co. Inc. USA distributor for Sensor Technology Ltd. (UK) 71 Perinton Parkway Fairport NY 14450 USA www.saelig.com www.saelig.com 585-385-1750 info@saelig.com info@saelig.com
www.keysight.com/find/triggerchallenge
Modern Test & Measure
SourceMeter Unit (SMU) Instruments Offer Cost-Effective Alternatives for FET Characterization By Mary Anne Tupta Senior Staff Applications Engineer Keithley Instruments, a Tektronix Company
34
EEWeb FEATURE
As every first-year electrical engineering student knows, a field-effect transistor (FET) is a majority charge-carrier device in which the current-carrying capability is varied by an applied electric field. It has three main terminals: the gate, the drain, and the source. A voltage applied to the gate terminal (VG) controls the current that flows from the source (IS) to the drain (ID) terminals. There are many types of FETs, including the MOSFET (metal-oxide-semiconductor), MESFET (metal-semiconductor), JFET (junction), OFET (organic), GNRFET (graphene nano-ribbon), and CNTFET (carbon nanotube). These FETs differ in the design of their channels.
35
Modern Test & Measure
Figure 1. Keithley’s Model 2636B Dual-channel System SourceMeter® SMU Instrument (200V, 0.1fA, 1.5A DC / 10A Pulse) is optimized for low current measurement applications.
MOSFETs are the most common variety of FET, serving as the basis for digital integrated circuits. Characterizing their current-voltage (I-V) parameters is crucial to ensuring they work properly in their intended applications and meet specifications. Typical I-V tests performed include gate leakage, breakdown voltage, threshold voltage, transfer characteristics, drain current, on-resistance, and others. Although characterization traditionally required several instrument capabilities— including a sensitive ammeter, multiple voltage sources, and a voltmeter—integrating, programming
and synchronizing the operation of multiple instruments can be extremely time consuming. Turnkey parameter analyzers are sometimes used to simplify this process, but their expense (often tens of thousands of dollars) can be difficult to justify. Source-measure unit (SMU) instruments (Figure 1) offer a more cost-effective alternative for parameter testing of FETs and other semiconductor devices. The number of SMU channels a specific test application requires (and the cost of configuring a system) usually depends on the number of FET terminals that must be biased and/or measured.
Typical I-V tests performed include gate leakage, breakdown voltage, threshold voltage, transfer characteristics, drain current, on-resistance, and others. 36
EEWeb FEATURE This article offers a high-level overview on techniques for making I-V measurements on FETs using oneand two-channel SourceMeter SMU instruments (Figure 2). These types of instruments are well suited for electrical characterization of FETs because they provide the capability to source current or voltage and measure current or voltage at the same time, have current resolution as low as 0.1fA, and can be current limited to prevent damage to the device. Embedded test script processors and software tools simplify performing common I-V tests on FETs and other semiconductor devices, without the need to create or install software separately.
A FET’s I-V characteristics can be used to extract many device parameters, to study the effects of fabrication techniques and process variations, and to determine the quality of the contacts. Figure 3 illustrates a DC I-V test configuration for a MOSFET using a two-channel SMU instrument (CH1 and CH2). Here, the Force HI terminal of SMU CH1 is connected to the gate of the MOSFET, and the Force HI terminal of SMU CH2 is connected to the drain. The source terminal of the MOSFET is connected to the Force LO terminals of both SMU channels or to a third SMU channel if it is necessary to source and measure from all three terminals of the MOSFET.
A FET’s I-V characteristics can be used to extract many device parameters, to study the effects of fabrication techniques and process variations, and to determine the quality of the contacts.
Figure 3. Below, MOSFET test configuration using two SMU channels.
Figure 2. Above, an SMU Instrument with an embedded software tool is used to generate a drain family of curves of the MOSFET.
37
Modern Test & Measure
Once the device is set up and connected to the SMU instrument, the embedded control software tool is configured to automate the measurements. First, the SMU instrument is connected to any computer with a web browser using an Ethernet cable. Entering the IP address of the SMU instrument into the URL line of the web browser will open the instrument’s web page. From that page, the user can launch the embedded software and configure the desired test or tests using a project “wizard.” Tests or projects can be saved and recalled for future use. One I-V test commonly performed on a MOSFET is the drain family of curves (VDS-ID). With this test, SMU CH1 steps the gate voltage (VG) while SMU CH2 sweeps the drain voltage (VD) and measures the resulting drain current (ID). Once the two SMU channels are configured to perform the test, the data can be generated and plotted on the screen in real time. Figure 4
shows a MOSFET drain family of curves created using a two-channel SMU instrument optimized for low current measurements. Once exported to a .csv file, this I-V data can be imported into a spreadsheet for further analysis or displayed in a table. The test setup illustrated in Figure 3 also supports testing drain current (ID) as a function of gate voltage (VG). For this test, the gate voltage is swept and the resulting drain current is measured at a constant drain voltage. Figure 5 shows the results of an ID-VG curve at a constant drain voltage. However, in this case, the generated data was exported to a file and plotted on a semi-log graph. This test can be easily reconfigured to step the drain voltage as the gate voltage is swept. The ID-VG data shows the many decades of drain current that the SMU instrument measured, from 1E-12 to 1E-2 amps.
Once the device is set up and connected to the SMU instrument, the embedded control software tool is configured to automate the measurements. Figure 4. MOSFET drain family of curves generated using a two-channel Model 2636B SourceMeter SMU Instrument running a project created using TSP® Express Software.
38
EEWeb FEATURE Low current FET test applications like these demand instrumentation capable of 0.1fA resolution. Low level measurement techniques like shielding and guarding are also essential to prevent errors. Shielding usually implies the use of a metallic enclosure to prevent electrostatic interference from affecting a high impedance circuit. Guarding implies the use of an added low impedance conductor, maintained at the same potential as the high impedance circuit, which reduces the leakage current in the test circuit. A guard doesn’t necessarily provide shielding.
About the Author Mary Anne Tupta is a Senior Staff Applications Engineer at Keithley Instruments, Cleveland, Ohio, which is part of the Tektronix test and measurement portfolio. She earned a B.S. in physics/electronic engineering and an M.S. in physics from John Carroll University. She has assisted Keithley customers with instrument applications since 1988. She can be reached at Mary.Anne.Tupta@keithley.com.
To learn more about cost-effective approaches to FET testing, consult Keithley’s application note, “Simplifying FET Testing with Series 2600B System SourceMeter® SMU Instruments,” available at http://www. http://www.keithley.com/data?asset=57501 keithley.com/data?asset=57501.
Figure 5. Drain current (ID) as a function of the gate voltage (VG) of a MOSFET.
39
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 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