Testing Wi-Fi 6E performance Page 13
JUNE 2021
The challenge of testing obsolete PCBs Page 20
Sorting out PC-based instrumentation Page 36
A SUPPLEMENT TO DESIGN WORLD
TEST & MEASUREMENT HANDBOOK
Cover FINAL — T&M HB 06-21.indd 1
6/8/21 11:23 AM
210430_AutoThermal_DW_US.indd 1
4/26/21 2:23 PM
EE-DW T&MHdbk K-Blog_SMT-THM TP_Radar_6-21p.indd 1
5/7/21 1:50 PM
TEST & MEASUREMENT HANDBOOK
MICROWAVES AND THE HAVANA SYNDROME THE
National Academy of Sciences recently released its conclusions about what sickened dozens of American
Embassy diplomats in Cuba, a phenomenon dubbed the Havana Syndrome. Though the panel reached no definitive conclusion, it found pulsed RF (a.k.a. directed microwave energy) was the most likely cause. Panel members could not rule out the possibility that the whole episode was a case of mass hysteria but considered the idea unlikely. The directed microwave energy theory rests on what’s called the microwave auditory effect. When a human head absorbs a pulse of RF energy, a rise in temperature causes tissue inside the head to expand slightly. The expansion launches a pressure wave that propagates throughout the skull to the inner ear, potentially causing clicking or buzzing sounds. Fortunately, the temperature rise is tiny (microdegrees) and the pressure wave is far too weak to injure tissue unless the microwave power density is huge. Critics have pooh-poohed the directed energy conclusion, claiming a microwave generator big enough to cause tissue damage would stick out like an NBA center at a jockey convention. Fortunately, there is open research on directed energy effects. So sufficiently interested individuals can do a little research and draw their own conclusions. One paper in this area published in Frontiers of Neurology points out experimenters were able to kill rats by exposing them to 2.45 GHz microwaves with a field intensity of 1,000 W/cm2. There’s no data in the open literature on the threshold of microwave power that causes human brain damage, but the researchers suggest a minimum intensity of 1 W/cm2 impinging on the human head–using 50 µsec pulses on a 7 kHz repetition rate--might be a good place to start. It has also been reported that the wife of a member of the Cuban embassy staff once looked outside her home after hearing
disturbing sounds and had seen a van speeding away. The implication is that the microwave generator was small enough to fit in the van. Also, the incident provides a means for making a ballpark estimate of range; we might say 150 ft from the street to the woman’s home would be a reasonable guess. So here’s how the calculation shakes out: We want to know the minimum output power of a 2.45 GHz microwave generator able to produce at least 1 W/ cm2 inside a house 150 ft away, through at least one wall. (For simplicity, we’ll assume the van was transparent to EM radiation.) We might also assume the transmitter is teamed with a parabolic antenna. A 6-ft-diameter parabolic antenna, which should fit in a van, can add about 30 dB of gain at 2.45 GHz. Using these parameters in a back-of-the-envelop calculation will lead you to conclude that the 2.45-GHz transmitter must put out at least 2 MW to get the job done. Today, a 2-MW transmitter in the 2.45-GHz range is about the size of a laundry basket and weighs about 150 lb. So it can certainly fit in a van, along with a power supply and modulation source, while leaving room for a human operator. If this scenario really did unfold as we theorize, we doubt the frequency used was smack in the middle of the WiFi band as in our example. (Though perhaps that doesn’t matter in Cuba.) But big microwave generators put in place to hassle diplomats seem to be at least theoretically feasible. Of course, there is one question this exercise doesn’t answer: Given all the controversy over pulsed microwaves, why not just stick spectrum analyzers in the homes and offices of U.S. diplomats? If this simple solution was mentioned in the NAS report, it certainly isn’t getting any press coverage.
LELAND TESCHLER • EXECUTIVE EDITOR 2
DESIGN WORLD — EE NETWORK
6 • 2021
EE Classroom is a syndicated content resource for electronic engineers looking for need-to-know information about various electronic components and systems. Curated by EE World’s editorial team, this digital content hub includes valuable technology background and insights, key trends affecting your designs of today and tomorrow, and frequently asked questions relating to a wide range of important electronic engineering topics. Topics include: • • • • •
Power electronics Embedded computing Test & measurement Sensors Connectivity
To view free educational content, go to www.eeworldonline.com/learning-center
DESIGN WORLD FOLLOW THE WHOLE TEAM ON TWITTER @DESIGNWORLD
EDITORIAL VP, Editorial Director Paul J. Heney pheney@wtwhmedia.com @wtwh_paulheney Senior Contributing Editor Leslie Langnau llangnau@wtwhmedia.com @dw_3Dprinting Executive Editor Leland Teschler lteschler@wtwhmedia.com @dw_LeeTeschler Senior Editor Aimee Kalnoskas akalnoskas@wtwhmedia.com @eeworld_aimee Editor Martin Rowe mrowe@wtwhmedia.com @measurementblue Executive Editor Lisa Eitel leitel@wtwhmedia.com @dw_LisaEitel Senior Editor Miles Budimir mbudimir@wtwhmedia.com @dw_Motion Senior Editor Mary Gannon mgannon@wtwhmedia.com @dw_MaryGannon Associate Editor Mike Santora msantora@wtwhmedia.com @dw_MikeSantora
CREATIVE SERVICES & PRINT PRODUCTION VP, Creative Services Mark Rook mrook@wtwhmedia.com @wtwh_graphics Art Director Matthew Claney mclaney@wtwhmedia.com @wtwh_designer Graphic Designer Allison Washko awashko@wtwhmedia.com @wtwh_allison
Graphic Designer Mariel Evans mevans@wtwhmedia.com @wtwh_mariel Director, Audience Development Bruce Sprague bsprague@wtwhmedia.com
IN-PERSON EVENTS Events Manager Jen Osborne jkolasky@wtwhmedia.com @wtwh_Jen Event Marketing Specialist Olivia Zemanek ozemanek@wtwhmedia.com
VIDEOGRAPHY SERVICES Video Manager Bradley Voyten bvoyten@wtwhmedia.com @bv10wtwh Videographer Garrett McCafferty gmccafferty@wtwhmedia.com
MARKETING
ONLINE DEVELOPMENT & PRODUCTION
VP, Digital Marketing Virginia Goulding vgoulding@wtwhmedia.com @wtwh_virginia
Web Development Manager B. David Miyares dmiyares@wtwhmedia.com @wtwh_WebDave
Digital Marketing Specialist Sean Kwiatkowski skwiatkowski@wtwhmedia.com
Senior Digital Media Manager Patrick Curran pcurran@wtwhmedia.com @wtwhseopatrick
Digital Production/ Marketing Designer Samantha King sking@wtwhmedia.com
Front End Developer Melissa Annand mannand@wtwhmedia.com
Graphic Designer Hannah Bragg hbragg@wtwhmedia.com
Software Engineer David Bozentka dbozentka@wtwhmedia.com
Senior Manager Webinars/ Virtual Events Lisa Rosen lrosen@wtwhmedia.com
Digital Production Manager Reggie Hall rhall@wtwhmedia.com
Webinar Coordinator Halle Kirsh hkirsh@wtwhmedia.com
Digital Production Specialist Elise Ondak eondak@wtwhmedia.com
Webinar Coordinator Kim Dorsey kdorsey@wtwhmedia.com
VP, Strategic Initiatives Jay Hopper jhopper@wtwhmedia.com
PRODUCTION SERVICES
FINANCE
Customer Service Manager Stephanie Hulett shulett@wtwhmedia.com
Controller Brian Korsberg bkorsberg@wtwhmedia.com
Customer Service Representative Tracy Powers tpowers@wtwhmedia.com
Accounts Receivable Specialist Jamila Milton jmilton@wtwhmedia.com
Customer Service Representative JoAnn Martin jmartin@wtwhmedia.com Customer Service Representative Renee Massey-Linston renee@wtwhmedia.com
2011- 2019
2014 Winner
2014 - 2016
Customer Service Representative Cathy Anthony-Gudel canthonygudel@wtwhmedia.com
4
DESIGN WORLD — EE NETWORK
6 • 2021
eeworldonline.com
|
designworldonline.com
CONTENTS TEST
02 06
&
MEASUREMENT HANDBOOK • JUNE 2021
MICROWAVES AND THE HAVANA SYNDROME GROUNDING CONSIDERATIONS FOR RENEWABLE POWER GENERATION
IT PAYS TO KNOW THE DIFFICULTIES THAT CAN ARISE WHEN TESTING AND EVALUATING GROUND SYSTEMS FOR SOLAR PANEL ARRAYS AND WIND FARMS.
10
13
TEST AND MEASUREMENT CONNECTIONS IN THE WORLD OF 5G
CONNECTORS AND CABLING CAN BE AS IMPORTANT AS THE TEST INSTRUMENTS USED TO CHECK OUT SYSTEMS OPERATING IN THE 5G REALM.
TESTING WI-FI 6E PERFORMANCE
THE 6 GHZ BAND BRINGS 1,200 MHZ OF ADDITIONAL SPECTRUM TO WI-FI WHICH COMPLICATES THE TASK OF TESTING AND VERIFYING FEATURES.
26
A DAY IN THE LIFE: FIVE RF MEASUREMENTS FOR FIELD ENGINEERS
30
KEY CONSIDERATIONS FOR SPECTRUM ANALYZERS
36
SORTING OUT PC-BASED INSTRUMENTATION
40
MEASURING PICOSECONDS WITHOUT BREAKING THE BANK
USING A SCOPE TO OPTIMIZE EMI INPUT FILTERS A SIMPLE OSCILLOSCOPE SETUP CAN SIZE COMMON-MODE AND DIFFERENTIAL-MODE NOISE FILTERING COMPONENTS 45 SEPARATELY AND MORE ACCURATELY. THE CHALLENGE OF TESTING OBSOLETE PCBs HOW TO TEST AND MAINTAIN OBSOLETE SYSTEMS WHEN THE RELEVANT DOCUMENTATION IS ABSENT.
23
CHOOSING A TEST SYSTEM FOR EMERGING PCIE 6.0 DESIGNS
16
20
48
A HANDFUL OF BASIC TECHNIQUES GO A LONG WAY TOWARD CHARACTERIZING MODERN COMMUNICATION SYSTEMS. A BEVY OF SPECIFICATIONS DEFINE HOW SPECTRUM ANALYZERS PERFORM. BUT JUST A HANDFUL OF QUALITIES CAN BE PIVOTAL WHEN IT COMES TO SELECTING INSTRUMENTATION. IT USED TO BE THE CASE THAT ONLY A FEW KINDS OF RELATIVELY SIMPLE INSTRUMENTS WERE BASED ON PCS. NOW SEVERAL CATEGORIES OF LAB-GRADE INSTRUMENTATION TAKE THE PC APPROACH.
MODERN ANALOG/DIGITAL CONVERTERS CAN HELP IMPLEMENT EQUIVALENT TIME SAMPLING TO PROVIDE PICOSECOND TIMING RESOLUTION.
RASPBERRY PI FOR DAQ
THOUGH THEY ARE MORE CLOSELY ASSOCIATED WITH THE MAKER SPACE, RASPBERRY PI BOARDS CAN BE EFFECTIVE AT HANDLING STRINGENT INDUSTRIAL MEASUREMENTS.
ATTENUATION NETWORKS AND THEIR MEASUREMENT HERE’S A BASIC REVIEW OF PADS AND HOW TO CHARACTERIZE THEM.
NEW PCIE 6.0 STANDARDS DEMAND INSTRUMENTS WITH UPWARDS OF 59 GHZ BANDWIDTH AND WHICH ARE CAPABLE OF RUNNING SPECIAL SIGNAL INTEGRITY TESTS.
eeworldonline.com | designworldonline.com
6 • 2021
DESIGN WORLD — EE NETWORK
5
TEST & MEASUREMENT HANDBOOK
GROUNDING CONSIDERATIONS FOR RENEWABLE POWER GENERATION IT PAYS TO KNOW THE DIFFICULTIES THAT CAN ARISE WHEN TESTING AND JEFF JOWETT, MEGGER EVALUATING GROUND SYSTEMS FOR SOLAR PANEL ARRAYS AND WIND FARMS.
INDUSTRIAL
wind turbines are
Lightning strike density varies enormously
typically 280 ft high
across the U.S. Much of the far west experiences less than only one stroke per year per square kilometer. Florida is the most hard-hit area and averages 32.2/yr/km2. The U.S. wind farm industry is now expanding into lightning-prone environments that may call for more extensive protective measures. A thorough protection design must consider not only the exposed blades but also the nacelle, structure, high-voltage power system, low-voltage control, and drive train. During a lightning strike, blades can experience temperature rising to 30,000° C, causing delamination, melting, surface damage, and cracking along their edges.
and getting taller. Their height makes them prime targets for lightning strokes that peak in less than 10 µsec while producing tens of thousands of amperes at one-million volts. A lightning strike can cause expensive damage or force a total replacement of the structure. Estimates are that up to 80% of wind turbine damage is caused by lightning. Yet adequate protection runs only about 1% of the total cost of a new turbine.
6
DESIGN WORLD — EE NETWORK
6 • 2021
eeworldonline.com | designworldonline.com
RENEWABLE ENERGY GROUNDING Much of this damage can go undetected but will shorten the blade life. Switchgear, frequency converters and transformer stations are also vulnerable. Sensors, actuators and steering motors in the control system can be damaged, and batteries can even explode. Lightning damage in the control system is the most common result of a strike while blade damage is most expensive. The single best protection against all such damage is a lowimpedance path to earth. Surge protection devices will fail if ground resistance is high. Towers in wind farms are typically daisychained or connected in a star configuration, making the grounding system potentially enormous and complicating the testing requirements. Phase cables have grounding sheaths connected to common buses. Communication cables have their own grounded conductors. As farms grow larger, it can become impossible to isolate the ground of a single turbine. If a field is under construction, it is a good idea to test the grounding of individual turbines before they become interconnected, as the distances involved in testing the entire completed field as a unit may be prohibitive. Also, lightning protection invokes a different set of criteria than power grounding because of the enormous currents and high frequencies of lightning strokes. Fault clearance and lightning protection have different criteria and may even be cross-connected inadvertently. Fault-protection grounding tends to be shallow buried. The current need only seek its source, typically a similarly grounded transformer. In contrast, safe dissipation of a lightning stroke’s enormous energy involves much greater depths. Another point to note: Sometimes wind turbine installers take shortcuts. For example, they may have grounded the tower to a nearby power conductor. The mere physical act of stretching test leads to extreme distances can be daunting. Testing may not be difficult for only one turbine. But a good practice is to keep precise records of exactly where test probes are placed to provide a frame of reference for subsequent tests as the power grid expands. Testing often becomes inadequate and unrepeatable as turbine sites expand. But exact records of test probe locations and maintenance tests over years can still be useful for comparative purposes, even if prohibitive distances to “remote earth” make it impractical to get true resistance readings. It is beyond the scope of this article to cover the complicated phenomenon of eeworldonline.com
|
designworldonline.com
ground coupling, but it is a good idea to become familiar with it. There is a great deal of metal in the ground at a wind farm. Thus it is easy for the lightning ground to couple with a nearby grounded cable. The coupling provides a low-resistance path that shorts out test current and prevents it from evenly distributing throughout the soil. Consequently, ground testing is inaccurate. Though not as dramatically exposed as wind turbines, solar paneling is often a lightning casualty as well. Rooftop solar structures have the same height problem as tower blades when it comes to lightning susceptibility, though not to
the same degree. And ground mounts on solar farms may be the only raised structures for acres. There is a high-frequency impedance between the grid frame and grounding electrode such that the rate-of-rise of the impulse is much higher than it would be at dc. In the absence of surge protection, lightning energy will enter the dc side of the system
AN EXAMPLE OF A GROUND TESTING INSTRUMENT IS THE DET2/3, DESIGNED TO MEASURE EARTH ELECTRODE RESISTANCE AND SOIL RESISTIVITY. THE DET2/3 EMPLOYS A FLEXIBLE AND USER-FRIENDLY APPROACH TO EARTH TESTS VIA ERROR DETECTION CAPABILITIES AND FULL TEST INFORMATION SHOWN ON A LARGE COLOR DISPLAY. THE DET2/3 CAN ALSO PROVIDE A LIVE TRACE OF ITS MEASUREMENTS ONSCREEN, WHICH GRAPHICALLY SHOWS THE AMOUNT OF CHANGES AND/OR NOISE FROM THE SYSTEM UNDER TEST.
6 • 2021
DESIGN WORLD — EE NETWORK
7
TEST & MEASUREMENT HANDBOOK
and couple into the ac side at the inverters, destroying much in its path. If panels reside on a building, lightning energy enters the house wiring and connected devices. Home surge protection power strips often lack the capacity to deal with so much energy and are generally only effective against much lower voltages and currents. The grounding conductors at the panels can develop significant voltage against the grounding electrode. Another point to note: Adding a separate grounding electrode for the solar panel frames can only make the situation worse. In this case, there is no longer an equipotential ground. Rather, the building electrical system may develop a high voltage compared to line or neutral conductors. Proper grounding and bonding is essential. NFPA 780 specifies requirements that can minimize damage to a PV system and attendant structures. National Electric Code Articles 250 and 690 specify compliant wire sizes, materials and techniques. Underwriters Laboratories, in UL 96A, specifies lightning protection components including size, location and installation. For safety, all electrically conductive components should be maintained at ground potential. With the frames and mounting structures of a solar array, this is a lot of equipment. Any extraneous current, as from a lightning strike, is provided with a low impedance path-to-ground that effectively shunts it around personnel while activating protective devices. The NEC states that “exposed noncurrent-carrying metal parts of module frames, equipment, and conductor enclosures shall be grounded.” The equipment-grounding conductor runs alongside other conductors in the array and combines components like inverters, disconnects, combiner boxes, battery boxes and any metallic-box-holding electrical equipment. The system must have
8
DESIGN WORLD — EE NETWORK
a properly sized conductor connected to the ground lug and to the grounding system. Every module must be connected to an equipment grounding conductor (EGC). There is frequently a connection point residing at about the middle of the long edge of the frame. The EGC must be sized properly. The requirements are set forth in the NEC based on the overcurrent protective device (OCPD). A ground fault protection device should also be installed. Not all PV systems require an OCPD. In such cases, the grounding conductor can be sized based on the short-circuit current. No gauge smaller than 14 AWG should ever be used. Installation in conduit can be useful as protection from the environment, corrosives and vermin. The EGC will be ineffective if not terminated in a grounding electrode (ground rod) of low resistance. Different organizations set their own standards, but generally, a ground of 5Ω or less should be installed. Deepdriven rods may be the most effective but can also be expensive. Instead, a buried grid or array of interconnected rods can suffice. Rods should always be placed farther apart than they are deep, so their fields don’t coalesce and act as one. Roof mounts can use the steel reinforcement in the concrete foundation
6 • 2021
(generally called a Ufer ground). However, it is a good idea to have a supplementary rod as fault clearance can damage the foundation. A remaining element is the grounding electrode conductor (GEC). It goes from the electrode to a point where all other grounded conductors can be connected. This point is commonly the ground busbar inside the main distribution panel. If the PV system is utilityinteractive, the grounding is then paralleled with the existing utility system ground. This practice prevents development of a voltage gradient between them that will behave counter to the purpose of the grounding system. One method is to connect an additional ground rod at the inverter and parallel it to the existing utility electrode. The inverter ground rod must have the same gauge as the utility GEC. Another method is to run a GEC from the inverter to the existing utility electrode, sized according to NEC. Standalone, battery-based systems where the solar array represents the sole power source to the ac load require a new grounding electrode. Here two 6 AWG or 4 AWG GECs are connected, one from the dc wiring enclosure where the inverter connects to the battery bank, the other from the ac MDP (main distribution panel). The NEC recommends separate GEC
eeworldonline.com | designworldonline.com
RENEWABLE ENERGY GROUNDING
sizes for ac and dc sides. In stand-alone systems, the grounding electrode is installed as close to the MDP as possible. Without conduit, 6 AWG is the recommended size for the GEC which should be robust enough for physical protection. If the GEC is not exposed to physical damage, 8 AWG on the dc side is sufficient. The minimization of lightning damage depends fundamentally on just two common factors: bonding and grounding. Bonding provides a low-impedance path via a grounding conductor from the point of lightning contact to the earth, where the grounding electrode facilitates transfer into the soil. In wind power facilities, tests of bonding effectiveness are electrically simple but physically difficult. Tests typically use a low- resistance ohmmeter having a four-terminal Kelvin bridge to establish a test current from 1 to 10 A through the grounding conductor and measure the total resistance with micro-ohm accuracy. A mere continuity test is insufficient because it only establishes that an electrical path exists, not that it is adequate. For wind towers, the difficult part is the long distances involved. But special leads can mitigate the distance problem while the bridge configuration automatically eliminates lead and contact resistance. The layout of huge solar arrays can make such tests daunting. But a lowresistance ohmmeter is still the correct tool. There may be many parallel paths around the interconnected panels, and these will affect the measurements. So the operator must focus on the direct path between the probes while allowing for other, circuitous routes. Finally, the grounding electrode resistance must be minimized. This can be a tough job in poorly conductive soils, but the more metal in the earth, the lower the resistance. A dedicated three- or four-terminal ground-resistance tester with a square wave output is necessary, not an ohmmeter. Here procedure is all-important. Merely stretching out leads and driving in probes will provide a correct measurement only by pure luck. The operator must understand the step-by-step procedure. Fall of Potential is the most accurate and reliable technique, but can experience problems related to distance and labor intensity. An astute operator should be familiar with various test procedures and know the best conditions for applying each one. And this is true for both wind and solar maintenance.
REFERENCES
Bruce Thatcher, Grounding: The Key To Lightning Protection, Wind Systems, https://www.windsystemsmag.com/grounding-the-key-tolightning-protection/#:~:text=Sankosha%20develops%20conductive%20 grounding%20cement%20to%20decrease%20turbines’%20vulnerability%20 to%20storms.&text=By%20their%20very%20nature%2C%20wind,volatile%20 weather%20makes%20them%20vulnerable. Vaisala National Lightning Detection Network, https://www.vaisala.com/en/ products/national-lightning-detection-network-nldn
INNOVATIVE HIGH VIBE / SHOCK CONNECTORS •T&M • INDUSTRIAL • MEDICAL • ROBOTICS • OIL & GAS • AUTONOMOUS
National Fire Protection Association handbook, https://www.nfpa.org/Codesand-Standards/All-Codes-and-Standards/Handbooks
VEHICLES
Megger, Getting Down To Earth, https://megger.com/support/technicallibrary/technical-guides/getting-down-to-earth-a-practical-guide-to-earth Richard P. Bingham, Flash, Then Bang; When Lighting Strikes, Electrical Contractor, https://www.ecmag.com/section/systems/flash-then-bang-whenlightning-strikes
®
www.LEMO.com 6 • 2021
DESIGN WORLD — EE NETWORK
9
®
TEST & MEASUREMENT HANDBOOK
TEST AND MEASUREMENT CONNECTIONS IN THE WORLD OF 5G JOHN MUZZIO, TIMES MICROWAVE SYSTEMS ANDREW DINSDALE, SV MICROWAVE
CONNECTORS AND CABLING CAN BE AS IMPORTANT AS THE TEST INSTRUMENTS USED TO CHECK OUT SYSTEMS OPERATING IN THE 5G REALM.
UNLIKE
previous cellular technology generations that were focused on a specific frequency band, 5G deals with a much larger potential
frequency range. For example, 4G frequency bands are typically below 3 GHz. 5G on the other hand can range from 450 MHz to 3.9 GHz and includes 20-52.6 GHz millimeter-wave bands for high-speed operations. 5G also encompasses unlicensed frequency bands, such as the 6 GHz band. This broad span of frequencies has introduced new challenges for 5G testing including repeatability, reliability, and reproducibility. As is the case with any RF testing, measurements require unique coaxial cable and connector setups— and it is critical to ensure signal integrity for RF interconnects. Measurement repeatability is key as these systems are connected and disconnected often. The connectors and cables must withstand extensive handling, and the materials they employ must be optimized. The testing process typically involves a device-under-test connected to a vector network analyzer (VNA), oscilloscope or spectrum analyzer. The signal path from the instrument to the circuit board is critical, and the user must ensure the test setup-including the test cable assembly, cable, adapters, and board mounted connectors--does not introduce unwanted variables. Many product developers working in 5G are dealing with a new frequency range that is unfamiliar, and the cables and connectors they’ve traditionally used will no longer work. For example, many RF test labs working in 4G LTE are outfitted with sub-10-GHz VNAs and associated test leads and accessories using Type N or 7/16 connector interfaces. To handle the full 5G spectrum, they must invest in not only expensive RF instruments, but also millimeter-wave test cables, adapters and board-level interconnects that can accommodate higher bandwidth without compromising signal integrity.
EXAMPLES OF SMP (SUB-MINIATURE PUSH-ON) SERIES PCB CONNECTORS. 10
DESIGN WORLD — EE NETWORK
6 • 2021
eeworldonline.com
|
designworldonline.com
5G CONNECTORS
When a signal transitions from the PCB to the connector, it is imperative to minimize reflections. At higher frequencies, imperfections in the transition from a coaxial connector to a circuit board structure become more apparent. These imperfections can cause parasitic and spurious signal responses. They manifest as return loss or insertion loss, spikes, and other undesirable effects. If the signal integrity is off and there is noise in the measurement, the test results won’t be accurate. High-fidelity measurements demand a repeatable, low-insertion-loss cable that functions throughout the desired frequency range. Similarly, a board-level product should be matched to the circuit board at the launch to ensure low return loss. For board-level products, millimeter-wave connectors offer a secure, threaded connection and can operate mode free in 5G frequencies (2.92mm DC-40 GHz, 2.4-mm DC-50 GHz, 1.85-mm DC-65 GHz, and 1.0mm DC-110 GHz). Push-on connectors such as SMP, SMPM, and SMPS connectors are candidates for applications involving high signal density. Phase control is another parameter for optimizing system performance in 5G smart antennas. Electronically steered antennas use multiple antenna arrays and vary the phase relationships among them to control the radiation pattern. This technique allows, for example, switching from a search radiation pattern to a tracking radiation pattern, or shifting the direction of radiation quickly. The arrays are fed by transmission lines; beam accuracy depends upon the phase relationships among the signals in those cables. Phase must be accurately controlled in the components within those systems, making phase-stable cable assemblies important. The super-short wavelengths characterizing the 5G frequency range imply that even short lengths of cable handling 5G signals span a lot of wavelengths at those frequencies. Differences in impedance over the cable length, though they are typically minuscule, make it tough to maintain phase accuracy. Cables can be phase-matched in pairs or groupings. But as temperatures warm or cool, the cables may not exactly track together; the signals they carry can go slightly out of phase. That small amount of degradation, known as the cable phase tracking characteristic, reduces antenna performance. To minimize such phase problems, phase-stable cables typically contain a TF4, or microporous PTFE dielectric, coupled with a helically wound metalized interlayer.
eeworldonline.com
|
designworldonline.com
THE MINI-D RF CONNECTION SYSTEM HAS A 0.110IN PORT-TO-PORT SPACING FOR HIGH MATING CYCLE APPLICATIONS WITHOUT DAMAGING THE HOUSING. THE LOW-PROFILE EDGE AND SURFACEMOUNT PCB CONNECTOR OPTIONS HANDLE FREQUENCIES THROUGH 67 GHZ.
EXAMPLES OF MILLIMETER-WAVE THREADED PCB CONNECTORS.
6 • 2021
DESIGN WORLD — EE NETWORK
11
TEST & MEASUREMENT HANDBOOK
HIGH DENSITY
5G specifications are based on the use of MIMO technology requiring multiple arrays of transmit and receive antennas. To accommodate larger bandwidth requirements, more antennas are squeezed into smaller spaces. This trend leads to use of highdensity interconnections, such as multi-pin connectors and minicoaxial cable bundles, between the antennas and the radios. Use of high-density multi-port connectors additionally allows testing to take place next to the DUT, shortening distances involved in testing and reducing the chance for interference. Also important is the use of flexible cable material that can be easily positioned on a test bench, either in R&D or during production. Testing often moves from module to module. At 5G frequencies, every movement of a module or cable could require a recalibration. However, a cable that can bend and flex will greatly reduce the amount of recalibration required while maintaining stability. Frequency, time delay, and physical properties such as length, dielectric constant and propagation velocity all affect electrical length. Coaxial cables are constructed with a consistent dielectric material throughout the length of the cable that creates a constant velocity factor. Though the material is consistent, environmental and handling factors can alter the cable electrical properties--like temperature fluctuations, flexure, twisting, pulling, and crushing--during installation and maintenance. The metals used in cables expand as they heat up and contract as they cool. Similarly, the dielectric constant expands and contracts as well, and its density changes, altering the signal velocity. The dielectric effects of the plastic offset and dominate the metal effects. The materials used to ensure phase stability and amplitude include PTFE dielectric, which allows for good signal transmission.
5G TEST SETUP
AN EXAMPLE OF A COMPLETE 5G RF TEST SET-UP INCLUDING COAXIAL CABLE AND CONNECTORS TO ENSURE SIGNAL INTEGRITY FOR RF INTERCONNECTS.
REFERENCES
Times Microwave Systems, www.timesmicrowave.com SV Microwave, www.svmicrowave.com
APPLICATIONS SUCH AS 5G DEVELOPMENT, BENCH VNAS AND ANALYZERS, AND RF MODULE TESTING ARE CANDIDATES FOR THE CLARITY SERIES 50-GHZ TEST CABLES WHICH BEHAVE PREDICTABLY OVER TEMPERATURE AND FEATURE A LONG FLEX LIFE.
12
DESIGN WORLD — EE NETWORK
6 • 2021
eeworldonline.com | designworldonline.com
WI-FI 6E
TESTING WI-FI 6E PERFORMANCE THE 6 GHZ BAND BRINGS 1,200 MHZ OF ADDITIONAL SPECTRUM TO WI-FI WHICH COMPLICATES THE TASK OF TESTING AND VERIFYING FEATURES.
NOW THAT THE
6 GHz region of the spectrum
is available for unlicensed use, Wi-Fi 6E devices are on their way to store shelves. Manufacturers face the juggling act of comprehensively testing these complicated devices while minimizing test time without compromising product quality. There’s a difference between Wi-Fi design validation testing (DVT) and manufacturing testing. DVT encompasses all aspects of device performance and validates a broad list of items that truly push the design to its operating limits. By contrast, manufacturing testing aims to identify manufacturing defects or component tolerance issues that affect device performance; the calibrations and verifications performed on the manufacturing floor are only those necessary to ensure device quality. At the baseband/modem level, Wi-Fi eeworldonline.com | designworldonline.com
6E and Wi-Fi 6 are similar – both follow the 802.11ax standard. However, the Wi-Fi 6E RF front end (RFFE) requires components that cover a new and wide frequency range from 5.925 to 7.125 GHz. Manufacturing tests of Wi-Fi 6E devices must focus on the weakest links in the RFFE to identify performance issues. In particular, active RF components that make up the front end can degrade performance. The 6 GHz band brings 1,200 MHz of additional spectrum to Wi-Fi (twice that of Wi-Fi 5 2.4 and 5 GHz bands combined). In the transmitter chain, the power amplifier (PA) must cover a wide bandwidth with high linearity. It must provide consistent gain even at the high edge of the band. At the highest channels, PAs with poor linearity will degrade the RF signal and, in turn, deliver poor signal range and coverage. Thus manufacturing tests must cover all channels in the 6-GHz band at maximum power to detect non-linearities. The Unlicensed National Information Infrastructure (U-NII) radio band refers to the part of the spectrum used by 802.11a/n/ac/ 6 • 2021
EVE DANEL, LITEPOINT
ax devices. The 6-GHz band (U-NII 5 to U-NII 8, 5.925 to 7.125 GHz) provides only a 75-MHz guard band for the highest Wi-Fi channel in the 5-GHz band (U-NII 3, 5.725-5.850 GHz), while the U-NII 4 band (5.850–5.925 GHz) reserved for dedicated short-range communications (DSRC) is only separated by a guard band of 25 MHz from the first channels in the 6-GHz band. Stations or access points (APs) operating in the low 6-GHz band must comply with the tight out-of-band (OOB) emissions requirements imposed by regulatory bodies. These narrow guard bands make it challenging to transmit in the low-6-GHz band without leaking into neighboring bands and force the use of advanced filtering techniques. Another approach used to meet these requirements is to back off from maximum transmit power in the first channel. Unfortunately, this approach sacrifices range on the first channels in the low-6-GHz band. The reduced range is especially problematic for the first 160-MHz channel (Channel 15) because some regulatory regions only provide DESIGN WORLD — EE NETWORK
13
TEST & MEASUREMENT HANDBOOK
FRONT-END NOISE SOURCES
enough unlicensed spectrum for three 160 MHz channels. To maximize transmit power, each device should be optimized during manufacturing for output power. This process, combined with spectral mask measurements, ensures a maximum transmit power without garbling emissions in adjacent channels.
POWER LEVEL AND RSSI ACCURACY
RSSI, or Received Signal Strength Indicator, is a measure of how well a device can obtain a signal from an access point or router. It’s basically a gauge of signal strength and determines if there is enough signal to get a good wireless connection. While not specific to Wi-Fi 6E, it poses a challenge tied to a feature introduced in the 802.11ax standard designed to improve capacity: Orthogonal frequency division multiple access (OFDMA) is a multi-user version of OFDM where the Wi-Fi channel is divided among multiple users who simultaneously
BLOCK DIAGRAM OF A TYPICAL RF FRONT-END FOR A WIFI 6 DEVICE AND THE TYPICAL SOURCES OF DISTORTION ENCOUNTERED IN EACH STAGE. exchange data with the access point. When multiple users share the available spectrum, interference from a single bad actor can degrade network performance for everyone else. Client stations located at various distances must adjust their power based on their estimated path loss to the AP. Devices closer to the access point transmit with less power, while devices farther away transmit at higher levels to produce the same power level at the access-point receiver. For uplink OFDMA transmissions, the 802.11ax standard requires stations to precisely measure the received signal strength indicator (RSSI) to evaluate path loss. In turn, stations
COMPARING SPECTRUMS
HOW THE 6-GHZ BAND SPECTRUM COMPARES TO OLDER 5-GHZ AND 2.4-GHZ BANDS.
14
DESIGN WORLD — EE NETWORK
6 • 2021
must precisely adjust their transmit power to participate during the transmission. Class A devices must meet ±3 dB RSSI accuracy and ±3 dB transmit power accuracy. Just as calibration during the manufacturing process ensures accurate transmit power, calibration of RSSI on the receive chain ensures accurate path loss measurement; the meeting of IEEE specs requires both RSSI and transmit power accuracy. Previously, RSSI was only used to determine the number of signal strength bars displayed on the user-interface screen–tight RSSI calibration was unnecessary. Now, RSSI is an integral part of how Wi-Fi 6 and Wi-Fi 6E work. Beyond standard compliance, for the end user RSSI calibration provides the added benefit of a better roaming experience. It allows station devices to determine and choose the signal from the best AP more precisely.
HIGH MODULATION: 1024 QAM OR 4096 QAM
1024 QAM modulation is already part of Wi-Fi 6 in the 2.4 and 5 GHz bands, while the previous- generation 802.11ac (Wi-Fi 5) only supported up to 256-QAM. 4096 QAM is not part of the 802.11ax standard. However, some chipsets have started to include it as part of their Wi-Fi 6E offering. These modulations unlock the highest Wi-Fi data speeds thanks to 10 bits of data encoded per sub-carrier in 1024 QAM and 12 bits of data per sub-carrier in 4096 QAM. With high-order modulation, the constellation points are much more closely spaced, and devices become more sensitive to impairments. Thus the receiving radio requires a higher SNR to properly demodulate signals. Obtaining the highest level of modulation accuracy for 1024 QAM and 4096 QAM forces the chipset and RF front end to have excellent phase noise and linearity performance. This accuracy is measured by the error vector magnitude (EVM). The 802.11ax standard requires < −35 dB EVM for 1024 QAM. 4096 QAM is not defined in the 802.11ax standard, however, its operation would require an EVM <-38 dB to deliver an equivalent signal quality. In manufacturing, EVM is measured because it provides an accurate picture of transmitter quality, and this metric will reflect any distortion in the transmit chain. When measuring EVM, it is important to ensure the test equipment’s own EVM floor is 8 to 10 dB better than the device under test (DUT). This margin eliminates sources of uncertainty in the measurement and improves DUT yield performance. eeworldonline.com | designworldonline.com
WI-FI 6E
ADJUSTING POWER
primary challenge comes from the addition of a new and wide frequency range from 5.925 to 7.125 GHz. The test plan should include the following items to ensure complete coverage: •
•
WI-FI 6 UPLINK OFDMA POWER CONTROL: MORE DISTANT STATIONS PUT OUT MORE POWER TO COMPENSATE FOR LARGER PATH LOSS.
Though 160-MHz channels were defined in the 802.11ac standard (Wi-Fi 5), they were often not supported or deployed because of the reduced spectrum that resulted in the 5-GHz band. The addition of the 6-GHz band opens enough contiguous spectrum to make full use of these wide channels. So expectations are they will be widely used in 6-GHz deployments. Wider channels present design challenges because more bandwidth also means simultaneous transmission of more OFDMA data carriers. The 802.11ax standard boosted the number of subcarriers by four over the previous generation. Sub-carriers are now 78.125 kHz apart in 802.11ax, compared to 312.5 kHz spacing in 802.11ac. A 160-MHz channel is comprised of close to 2,000 sub-
carriers. Distortions arise when the carriers at different frequencies are attenuated or amplified by different factors; the larger the range of frequencies, the more likely they are to exhibit this type of distortion. Thus manufacturing tests must validate transmitter performance for 160 MHz channels. The spectral flatness is an excellent metric for this purpose as it measures the distribution of the power variations for all the sub-carriers of the OFDMA signal. In a nutshell, the main goal for production test is to exercise the device as much as necessary to identify manufacturing defects while minimizing test time. An optimal production test focuses on areas exhibiting the largest degree of variability associated with the manufacturing process. For Wi-Fi 6E the
•
•
•
Low, mid and high 6-GHz band channel test coverage to verify consistent performance over the entire frequency band. Transmit-power calibration and spectrum-mask measurements on 6-GHz channels to ensure meeting of regulatory emissions and optimization of transmit power. RSSI calibration to ensure accuracy of path loss measurement for up-link OFDMA transmission. EVM measurement for 1024 QAM (MCS10, MCS11) and 4096 QAM (MCS12, MCS13) (if supported by the chipset). Test coverage for 160 MHz channels to ensure consistent performance.
For such high-level testing, test speed becomes the most efficient tool for controlling test costs. The test tool must be able to make use of advanced testing techniques combining multi-DUT parallel testing and test-command sequencing to reduce test time. These capabilities together yield greater test throughput for the next generation of Wi-Fi 6E devices.
REFERENCES
LitePoint, www.litepoint.com
A VIEW OF A 4096 QAM CONSTELLATION ILLUSTRATES WHY WIFI 6 CHIPSETS AND RF FRONT ENDS MUST HAVE EXCELLENT PHASE NOISE AND LINEARITY PERFORMANCE TO MEET THE IEEE REQUIREMENTS.
eeworldonline.com | designworldonline.com
6 • 2021
DESIGN WORLD — EE NETWORK
15
TEST & MEASUREMENT HANDBOOK
USING A SCOPE TO OPTIMIZE EMI INPUT FILTERS A SIMPLE OSCILLOSCOPE SETUP CAN SIZE COMMON-MODE AND DIFFERENTIAL-MODE NOISE FILTERING COMPONENTS SEPARATELY AND MORE ACCURATELY. MARCUS SONST, ROHDE & SCHWARZ
ANY
switched-mode power supply (SMPS) needs an EMI (electromagnetic
interference) input filter to avoid disturbing power lines. Input filters generally contain both commonmode and differential-mode filter elements, and the two are rarely optimized separately. In particular for high-power applications, this practice can result in a significantly larger EMI filter than actually necessary. But a simple procedure using a dual-output LISN (Line Impedance Stabilization Network) and an oscilloscope can optimize common-mode and differential-mode noise components separately. The result is more accurate data for designing an optimal input filter. An SMPS generates a fair amount of noise. The choice of SMPS topology is significant and influences filter design. For example, a dualinterleaved boost topology creates less noise than a simple boost converter. For a given topology, there are several design parameters that influence noise levels. The switching frequency of the converter is key. Often, a high switching frequency is chosen in the interest of a compact design but at the cost of more EMI. It is important to understand the correlation between rise and fall time of the switching element and the generated noise. Nowadays, wide-bandgap devices based on SiC or GaN are more widely used in power converter designs to boost efficiency. These technologies are known for their fast switching and potential for generating EMI in the absence of mitigation
16
DESIGN WORLD — EE NETWORK
THE LOCATION OF COMMON-MODE AND DIFFERENTIALMODE NOISE CURRENTS IN A TYPICAL DUT.
AN OSCILLOSCOPE SETUP FOR CHARACTERIZING NOISE CURRENTS. BOTH SCOPE CHANNELS ARE SET TO DISPLAY SIGNALS IN THE FREQUENCY DOMAIN.
6 • 2021
eeworldonline.com | designworldonline.com
INPUT FILTER OPTIMIZATION A TYPICAL SCOPE FFT DISPLAY WITH NO EMI FILTERS APPLIED USING THE TESTING SETUP DESCRIBED. measures. Parasitic elements can also play a role. For example, a metal housing near a high-voltage switching element can create a parasitic capacitance acting as a path for common mode noise to leave the system. The key components for the differentialmode EMI filter in an ac/dc converter are the differential-mode inductors and the X-capacitors. For the common-mode EMI filter, the common-mode choke and the Y-capacitor are key. In some cases, the differential-mode inductors can be omitted as the common-mode choke can also act as differential mode inductor. Electromagnetic compatibility standards require measurement of the conducted emissions on both power lines and dictate that the EMI voltages remain below specified limits at all frequencies in the range. This measurement takes place first on one power line, then the other. Unfortunately, this procedure provides no insights into noise propagation mechanisms because it measures common-mode and differential-mode noise simultaneously. The common mode current Icm flows from the DUT (device under test) on both power lines into the LISN and to the DUT back via the external ground path. Thus the sum of the two currents flow to the external ground path. The amplitude and phase are the same on both conductors. The differential mode current has
a different characteristic; the current on the positive conductor flows into the LISN while the noise return path is the negative conductor. The only difference is the phase between these two currents; they differ by 180° and ideally should cancel out. A little bit of mathematics separates the common-mode and differential-mode noise terms. Using the individual currents:
we can easily calculate the voltages on the two conductors:
IP = ICMa + IDM IN = ICMb – IDM
VP + VN = VCMa + VCMb
Based on the relations between individual voltages and common-mode and differentialmode voltages:
we can calculate the common-mode and differential-mode voltages as follows: VCM = VP + VN VDM = ½ (VP − VN)
HOW A DIFFERENTIAL-MODE FILTER AFFECTS THE DUT SPECTRUM.
eeworldonline.com | designworldonline.com
VP = (ICMa + IDM) × ZLISN VN = ( ICMb − IDM) × ZLISN
The simple subtraction results in a value that is twice the differential mode noise level, or 6 dB extra. Using these simple calculations, we distinguish between common-mode noise and differential-mode noise (including subtraction of 6 dB from the differential result). The simple calculation works well if the setup (cable, components of the LISN, etc.) is as symmetrical as possible; the noise on the two conductors must be measured simultaneously. A simple but effective setup to separate common-mode and differential-mode noise employs a dual-output LISN (or two identical LISNs) to probe both power lines. The two signals are captured on two channels of an oscilloscope. The sum and difference signals are calculated on the scope as is the (Fast
6 • 2021
DESIGN WORLD — EE NETWORK
17
TEST & MEASUREMENT HANDBOOK
INSERTION OF A COMMON-MODE CHOKE SERVING AS A COMMON-MODE FILTER EFFICIENTLY DAMPS NOISE.
Fourier Transformation) FFT. The result is direct access to the commonmode and differential-mode noise. While any non-symmetry between the two LISNs will influence the measurement result, this method provides reasonably accurate results. It is important that the setup use cables of identical length and cables with sufficient quality to avoid a shift in time or loss in amplitude. Such factors degrade the ability to separate the noise components. Furthermore, the oscilloscope needs a sufficiently low-noise front end and a sufficiently fast FFT function. To help illustrate this method, consider the case of a simple stepdown buck converter DUT. Its input filter is a simple PI-LC configuration which is effective for damping differential-mode noise. The setup simplifies the task of either applying or excluding the PI-LC filter. No common-mode filters are included on the PCB, so a common-mode choke is attached externally. The converter has no housing; the PCB simply sits on an isolation block that is on a metal ground plane. The setup deliberately avoids generating excessive common- mode noise. The first measurement is taken to show the highest spectrum in the input power conductors. A reference level measurement establishes the noise level of the system while the DUT is switched off. The extra 6 dB in differential mode is compensated by dividing the sum expression by two, then performing the FFT. For common mode, the sum expression is used directly. The total amount of common-mode noise is represented by the sum of the two measurement channels. The peak at 300 kHz in the reference line is caused by the system, not the converter, and can be ignored at least up to 25 dBµV. The high differential-mode noise (approximately 65 dBµV) during the measurement at 300 kHz arises from the switching frequency of the converter. The harmonic and all higher odd multiples of this frequency stem from the reflected ripple current, which dominates the differential-mode spectrum. Some peaks are also visible in the common mode spectrum; these are not filtered by a differential filter. Calculations show an LC filter will damp the fundamental at 300 kHz. The calculated filter resonance frequency is at 19.3 kHz, which should
18
DESIGN WORLD — EE NETWORK
6 • 2021
suppress the switching frequency by 40 dB. The filter structure is of second order and thus the damping is about 40 dB/decade. The differential-mode noise is efficiently reduced up to 10 MHz, damped up to 30 dB compared to the unfiltered value. The magnitude of the fundamental at 300 kHz and the multiple harmonics are much lower. At higher frequencies the filter is less effective, damping noise only by 10 dB. The filter doesn’t affect common-mode noise significantly because the it was designed to filter differential- mode noise. Another filter is added to damp common mode noise, taking the form of a common-mode choke from Würth Electronic. The common-mode noise is greatly reduced from 2 MHz to 60 MHz. In addition, the differential-mode noise is also damped as the commonmode choke is not ideal. The resulting leakage inductance functions as a differential filter. Furthermore, the differential-mode noise may also be affected if the setup is not optimized (no PCB for the common-mode choke). Thus some asymmetrical components may cause this additional damping effect. All in all, an effective input filter helps fulfill EMI standards for SMPS conducted emissions. Accurate information about the commonmode and differential-mode noise contributions greatly aids the task of designing and optimizing the EMI input filter.
REFERENCES
Rohde & Schwarz, www.rohde-schwarz.com/
eeworldonline.com | designworldonline.com
TEST & MEASUREMENT HANDBOOK
THE CHALLENGE OF TESTING OBSOLETE PCBs HOW TO TEST AND MAINTAIN OBSOLETE SYSTEMS WHEN THE RELEVANT DOCUMENTATION IS ABSENT. ALAN LOWNE, SAELIG CO. INC. 20
OBSOLETE PCBs
or boards lacking circuit
schematic diagrams present a conundrum when there’s a need for repair. Online searches for old datasheets may not be productive. What happens when obsolete boards must be fixed but there is only one “golden board” left?
DESIGN WORLD — EE NETWORK
6 • 2021
The scenario often arises in the military sector where equipment has a long operational life but the original manufacturer may be long gone, and the relevant documentation is unobtainable. The military must maintain its equipment no matter how long ago a product was designed and assembled. The ability to repair or replace legacy components is essential for military readiness, but it can also be costly.
eeworldonline.com | designworldonline.com
OBSOLETE PCBs
When systems have an operational life of upwards of 25 years, the support strategy for spares and repairs is critical. Product support often evaporates when companies are taken over, restructured, or closed. Rapidly changing technology also makes support problematic. Consumer products, for example, often lose support long before the end of their operating life. But obsolescence is unacceptable for a product lacking a physically and functionally identical (form/fit/function) replacement that is also part of a validated system having a long life. A change to a new component in critical equipment such as military gear or medical diagnostic machines may be enormously costly and could require recertifying the whole system. The need is for a system that will create essential product documentation available to all parties. Even better would be one that provides a schematic of the PCB under test and even creates a new PCB layout to enable the manufacture of new boards. The typical way of redocumenting a PCB involves laborious, manual pointto-point circuit detection, observation, and notation. There is a better and easier way. A product from ABI Electronics (UK) called RevEng (REVerse ENGineering ) has been designed as a reverseengineering hardware tool for recreating professional quality circuit schematic diagrams from a sample PCB lacking supportive documentation. The RevEng System consists of a PC-controlled continuity-detecting hardware system, control software called SYSTEM 8 Ultimate, and EdWin, a fully featured CAD package. RevEng ‘learns’ the connectivity of the sample circuit, producing a netlist of the components and connections which can be imported into either the EdWin software or into another schematic software package, producing professional quality circuit diagrams. New PCB layouts can also be created from these netlists using autorouting PCB layout software. Learning in RevEng takes place using a selection of clips, connectors, and probes that are attached to clusters of components on the board DUT. SYSTEM 8 Ultimate software guides the operator to place and move the clips around the board. SYSTEM8 Ultimate generates an efficient sequence of clip combinations and movements to learn all possible connections, but the operator can override the automatic placement of the clips if necessary. To minimize operator errors, the system applies an orientation check and pin check to confirm clip contact and position. RevEng completes its learning process without applying power to the board and limits the applied voltage and current from the probes so semiconductor gates don’t turn on - which also makes
eeworldonline.com | designworldonline.com
PREVIOUS PAGE, THE ABI REVENG SCHEMATIC LEARNING SYSTEM. ABOVE, THE XP CABINET SYSTEM VERSION WITH 512 CHANNELS.
6 • 2021
DESIGN WORLD — EE NETWORK
21
TEST & MEASUREMENT HANDBOOK
A SCHEMATIC GENERATED BY THE REVENG SYSTEM. it safe to use on low-power technologies. The resulting net list can form an input to CAD packages and as data input for ATE programs. SYSTEM 8 Ultimate software also provides links to CAD software. It is also possible to ‘learn’ part of the circuit so just an obsolete or non-working section of a board can be replaced. The use simply defines those components to be included in the drawing and follows the normal RevEng guided procedure. Part or all of the remainder of the circuit can be included later if needed. To ensure complete confidence, RevEng offers rescan and verification procedures.
A drawing package called EdWin comes with RevEng and enables quick generation of drawings by importing the net list and using an automated process to place components and signal routing. The resulting drawings can include bus structures and multi-page schematics. EdWin has features such as importing the netlist to place components, auto-routing signals, multi-page schematics, bus structure support, a library of over 12,000 devices, ‘rubber band’ and ‘rats nest’ functions, block move and rotate, etc. The component library includes discrete parts, analog and digital ICs, microprocessors, memories, and connectors - but new and custom devices can be easily added without necessarily knowing the function of the device or even which pins are inputs or outputs. The CAD software can be enhanced to include artwork for a circuit board layout, design verification, and simulation. Optional capabilities include BOM, PCB layout, mix-mode simulation, arizona autorouter, thermal analyzer, EDSpice, EMC + signal integrity. Edwin’s special features enable sameday drawings of unknown, complex boards. Low-budget or small-to-medium circuits can use an entry-level RevEng System with 256 measurement channels. Larger, more complex circuit needs can use a high-pin-count system equipped with up to 2,048 channel measurement channels.
REFERENCES
RevEng, www.saelig.com/product/TSTEQPCB009.htm
A NEW VIRTUAL EXPERIENCE FROM ROBOTICS BUSINESS REVIEW!
Fostering Innovation, Expanding Opportunities, Building a Community
FEATURING:
Speaker Presentations
Live Discussions
Follow-up Q&As w/ presenters
robobusiness.com
Saeling — TM HB 06-21_converted.indd 22
6/8/21 11:42 AM
PCIE 6.0 TESTING
CHOOSING A TEST SYSTEM FOR EMERGING PCIE 6.0 DESIGNS NEW PCIE 6.0 STANDARDS DEMAND INSTRUMENTS WITH UPWARDS OF 59 GHZ BANDWIDTH AND WHICH ARE CAPABLE OF RUNNING SPECIAL SIGNAL INTEGRITY TESTS. CONTRIBUTED BY ANRITSU CO.
PCI EXPRESS (PCIE)
6.0 addresses the highspeed data transmission
needs of emerging applications ranging from data centers to connected cars. With the doubling of data rates and upgrades of performance specifications, PCIe 6.0 adds complexity to high-speed interconnect designs. Engineers need signal integrity tests and analysis tools able to verify products comply with the new PCIe 6.0 standards.
latency, PCIe technology is a load-store protocol, dividing actions into those associated with memory and registers. Therefore, it must strictly maintain all specifications, especially when it comes to latency, power, and high bandwidth. Using FEC and CRC allows PCIe 6.0 to realize the specified low latency with latency reduction in most cases. It also realizes low complexity and a low bandwidth overhead.
PAM4 EFFECT ON BER
PAM4 signaling alleviates channel loss because it runs at half the frequency with two bits per UI. Because PCIe 6.0 has three eyes in the same UI, however, eye height and width are reduced. As a result, the BER will be several levels of magnitude higher with PAM4, which is why FEC is necessary. For PCIe 6.0, BER is a combination of the FBER, correlation of errors in a lane, and correlation of errors across lanes. There are two primary mechanisms to correct the errors in a lane and those across lanes. The most notable are through FEC and detection of errors by CRC, resulting in the eventual correction through link-layer retry. FEC operates on the principle of sending redundant data that can be deployed to correct some errors at the receiver. CRC is an error
When jumping to 64 GT/sec for PCIe 6.0, the PSI-SIG standards committee only slightly altered the compliance requirements. Tighter channel and connector loss and reflection parameters have been implemented to address signal degradation. There were slight improvements in receiver and transmitter equalization as well. There are no major innovations to address the expected complications associated with steeper rise-fall times, narrower unit intervals (UIs), and greater insertion loss associated with doubling the data rates. PCIe 6.0 utilizes 32 Gbaud PAM4 signaling. Though the underlying frequency is the same as the PCIe 5.0 specification, there is extra circuitry and logic involved for the PAM4 mode to track three eyes, along with the logic changes needed to operate in Flow Control Unit (FLIT) mode. FLIT was selected because it allows error correction on fixed-sized packets. Because error correction happens on FLIT, the cyclic redundancy check (CRC) and retry must take place at PCIE SPECIFICATION the FLIT level. In addition to PAM4, the PCIe 6.0 specification includes error 1.0 assumptions, including correlation between errors on a lane, as well as across lanes. PCIe 6.0 uses a special approach to maintain 2.0 low latency through a combination of relatively lower First Bit 3.0 Error Rate (FBER) combined with a light-weight, low latency 4.0 Forward Error Correction (FEC) for initial correction. A robust CRC then detects any errors that remain after correction. The result is a 5.0 link-level retry, which is also low latency. 6.0 (WIP) Unlike networking standards that have 100+ nsec of FEC
eeworldonline.com | designworldonline.com
6 • 2021
THE DATA RATES OF EACH PCIE GENERATION. DATA RATE(GT/S) (ENCODING) X16 B/W PER DIRN**
YEAR
2.5 (8b/10b)
32 GT/s
2003
5.0 (8b/10b)
64 GT/s
2007
8.0 (128b/130b)
126 GT/s
2010
16.0 (128b/130b)
252 GT/s
2017
32.0 (128b/130b)
504 GT/s
2019
64.0 (PAM-4, Flit)
1024 GT/s (~1Tb/s)
2021*
DESIGN WORLD — EE NETWORK
23
TEST & MEASUREMENT HANDBOOK
RECEIVER LINK EQUALIZATION TEST CONFIGURATION. THE BERT PPG DIFFERENTIAL OUTPUT IS TRANSMITTED TO BOTH THE DUT-RECEIVER AND THE OSCILLOSCOPE. THE DUT-TRANSMITTER OUTPUT GOES BOTH TO THE OSCILLOSCOPE AND THE BERT ED, WHICH ACTS AS A REFERENCE RECEIVER.
detection code used to authenticate packet transmission between the sender and the receiving end. PCI-SIG established a low-latency FEC of below 2 nsec for PCIe 6.0, and that is to be part of the specified overall signal latency of below 10 nsec. FEC is based on a fixed number of symbols, making it simple to transition to FLITs, as they are fixed size as well. Link frequency is 64 GT/sec. FEC logic can be run at any frequency. The expectation is that the logic will operate at 1G (or 500 MHz or 2G) and easily reach a latency exceeding 2 nsec. PCI-SIG recommends a lightweight FEC for correction. The robust CRC for detection, combined with a fast linklevel replay, handles any errors that the FEC cannot correct. As long as the replay probability of a FLIT is approximately 10-6, there is no appreciable performance impact either from the FEC latency or the replay latency in case of an undetected error. A combination of FEC correction and CRC detection results in a replay that effectively corrects nearly all common errors. A recommended approach is to establish a FEC symbol error threshold. By doing so, engineers have broader control over error conditions that affect patterns during capture by ignoring insignificant events that are
24
DESIGN WORLD — EE NETWORK
normally corrected in the FEC environment. To set a threshold, a bit error rate tester (BERT) generates a PAM4 signal to the Device Under Test (DUT) receiver input. The DUT determines the logic state of the input signal and loop decision to transmitter output for the error signal in the BERT for analysis. The BERT’s built-in Error Detector (ED) determines if the DUT’s decision was correct. For relevant results, the BERT’s jitter and noise profiles must comply with standards.
LINK EQUALIZATION
Link training and stressed receiver tolerance are simultaneously evaluated using a stressed signal in the link equalization test. Two tests – for receivers and transmitters – must take place using SigTest software developed by the PCI-SIG. Though similar to a standard stressed-eye receiver BER, receiver link equalization has one notable difference. The DUT must first perform link negotiation to correctly compensate for the test channel. The idea of stressed receiver tolerance testing is to send the DUT-receiver the worst-case signal that still complies with the specification. Prior to conducting the test, the signal transmitted from the BERT must be precisely calibrated to mimic the worst-case signal at 6 • 2021
the end of the test channel. The test signal has jitter and interference impairments that include random jitter (RJ), sinusoidal jitter (SJ), sinusoidal differential mode interference (DMI), and common mode interference (CMI), The BERT PPG differential output is split so the signal goes to both the DUT-receiver and the oscilloscope. The DUT-transmitter output is also divided so its signal goes to the oscilloscope and the BERT ED, which acts as a reference receiver. Transmitter link equalization is a required compliance test that verifies the device correctly changes equalization within the specified time when the link partner requests it. The BERT requests an equalization change from the devices that simultaneously sends a trigger to the oscilloscope so the time delay to the DUT can be measured in the electrical domain. The BERT PPG sends requests to the DUT-transmitter through the PCIe physical layer logic-sub-block protocol. The BERT PPG sequentially sends requests to the DUT-AIC for every FFE preset at each PCIe data rate. The DUT-transmitter modifies its FFE scheme and transmits the signals. The DUT-transmitter output is split so its signal goes to the oscilloscope and to the BERT ED. The oscilloscope observes the high-level eeworldonline.com | designworldonline.com
PCIE 6.0 TESTING
A TRANSMITTER LINK EQUALIZATION TEST SETUP.
equalization change while the ED serves as a reference receiver that confirms the preset change. The BERT uses the PPG auxiliary output to trigger the oscilloscope acquisition of each signal. The oscilloscope captures the waveforms with every FFE preset and every data rate. It then runs SigTest to evaluate each waveform according to the compliance requirements. Reports on the results can also be created through SigTest.
SELECTING THE PROPER TEST SYSTEM
To conduct these measurements accurately, PCIe 6.0 test systems need a feature-rich, protocol-aware BERT and an oscilloscope. The BERT needs a built-in instrument-quality PPG that can apply precise levels of specific signal impairments and a built-in ED capable of verifying compliance with the PCIe specifications. The BERT should have multiple NRZ pattern-generating channels and error detectors that operate at 32 GT/sec and PAM4 channels at 64 GT/sec to support PCIe 6.0 and earlier generations. Low intrinsic jitter of 115 fsec and 12 psec 20 to 80% rise/fall times are also necessary for signal integrity. The BERT must apply every required signal impairment in amplitude ranges that exceed those required by the PCIe 6.0 specifications, Certain BERTs have FEC analysis functions for the built-in ED. The functions leverage high input-sensitivity performance or the ED to detect FEC symbol errors based on the 400 GbE FEC standard. Bit error changes and FEC symbol errors with alterations in input amplitude and jitter conditions can be monitored in real-time to quickly and reproducibly conduct evaluations when symbol error counts exceed the correction ability of FEC. The oscilloscope should have real-time sampling bandwidth exceeding 50 GHz. For both transmitter signal evaluation and calibration of stressed-eye receiver eeworldonline.com | designworldonline.com
tolerance tests, the oscilloscope must also support PCISIG test software analysis tools. Test systems must have this performance level to verify PCIe 6.0 designs in emerging applications. Two notable applications are data centers and automotive. Data centers now must support next-generation, high-speed, large-capacity 5G mobile communications. To do so, they are installing equipment meeting the 400 GbE communications standard. There is also investigation into 800 GbE and 1.6 TbE standards to facilitate faster speeds. PCIe 6.0 will be the connectivity technology used for these emerging high-speed designs. Automotive applications essentially require data center-class computing power and home-theater complexity. PCIe 6.0 technology is enabling the automotive infotainment and connectivity ecosystem, including critical safety applications such as Advanced Driver Assistance Systems (ADAS). It is becoming the connectivity of choice because it features fault tolerance by multiple error correction system architecture, secured interoperability, and high bandwidth. All in all, PCIe 6.0 is creating new design challenges that require sound testing processes. The basic test system required is comprised of a protocol-aware BERT with FEC functionality and a high-speed oscilloscope that produces high-quality eye diagrams and has comprehensive analysis tools.
REFERENCES
Anritsu Co., www.anritsu.com
6 • 2021
DESIGN WORLD — EE NETWORK
25
TEST & MEASUREMENT HANDBOOK
A DAY IN THE LIFE: FIVE RF MEASUREMENTS FOR FIELD ENGINEERS A HANDFUL OF BASIC TECHNIQUES GO A LONG WAY TOWARD CHARACTERIZING MODERN COMMUNICATION SYSTEMS.
FIELD ENGINEERS
often find that every
day presents its own unique challenges. A typical work day may involve measuring a variety of devices or signals — cables, antennas, over-the-air 5G signals, or intermittent spurious signals. Successful completion of these tasks requires a basic understanding of RF measurements and a portable, easy-to-use handheld instrument. There are five common measurements field engineers perform: real-time spectrum analysis (RTSA), noise figure measurements, cable and antenna tests (CAT), over-the-air (OTA) testing, and electromagnetic field (EMF) exposure evaluation. Here’s a rundown on these basic procedures.
26
DESIGN WORLD — EE NETWORK
SARAH GROSS, KEYSIGHT TECHNOLOGIES
RTSA processes signal samples without gaps and generates measurements, such as scalar, power, or magnitude, that correspond to traditional spectrum analysis measurements. Signal interference over wireless networks is on the rise, resulting in poor signal quality that leads to dropped links or choppy audio. Interference profoundly affects wireless devices and communications that range from car radios to mission-critical applications such as public safety. Traditional spectrum analysis techniques have a built-in dead time during which the analyzer processes the data for display. Intermittent interference may arise during this dead time. Additionally, in a highly dynamic signal environment, wider or longerduration signals mask weak signals and cause interference. Gapless RTSA, on the other hand, detects and reveals these transient, overlapping signals so engineers can visualize the interferer.
6 • 2021
Co-channel interference detection and troubleshooting are the most challenging tasks in a communications network because interferers can hide under the serving frequency. Typically, the user must turn off the carrier transmitter to find any other signals in the same frequency channel. The process of turning off the carrier signal tends to be intrusive and can disrupt normal communication services. Besides, under many circumstances, turning off serving transmitters is not viable, depending on the nature of the services such as base station testing. Fortunately, RTSA profiles over-the-air characteristics, detecting hidden interferers under the serving carrier.
AT LEFT, A WI-FI SIGNAL CAPTURED BY KEYSIGHT’S FIELDFOX HANDHELD ANALYZER USING 10-MHZ BANDWIDTH. AT RIGHT, A DISPLAY CAPTURED WITH 100 MHZ BANDWIDTH. THE 100-MHZ BANDWIDTH SETTING ALLOWS VISUALIZATION OF THE WHOLE BAND IN ONE DISPLAY.
eeworldonline.com
|
designworldonline.com
RF FIELD MEASUREMENTS connects a transmitter to its antenna, or between an antenna and its receiver. CAT identifies the location of poor performance in adapters and damaged antennas, as well as where breaks or bends reside in cables. Faulty cables, connectors, and antennas cause many cellular base station problems. Problems include poor coverage and unnecessary handovers. Component failure frequently results A SCREENSHOT OF A NOISE FIGURE MEASUREMENT from harsh weather that damages exposed cable system transmission lines. Sheltered cable INCLUDING UNCERTAINTY CALCULATIONS. installations are also subject to heat, stress, and oils leaking into the system. Additionally, cable faults commonly arise at interfaces between cables and connections where soldered joints and crimps in the cable weaken and break. Transmission lines are often too long to make end-to-end cable measurements that would reveal the location Noise figure measures the degradation of the signal-to-noise ratio as a of a fault. There are two cable troubleshooting techniques to try when signal passes through an active or passive device. end-to-end measurement is impossible and a kink or cut forms in a Noise figure uniquely characterizes entire systems as well as line: Distance-to-fault (DTF) reports the location of each cable fault. their components, including preamplifiers, mixers, and intermediate Time-domain reflectometry (TDR) characterizes the type of fault, such frequency amplifiers. By controlling component noise figure and gain, as a bend in the cable, or cut. A bend in the cable appears capacitive the designer controls the noise figure of the overall system. With the (the return trace reflects downwards) while a cut in the cable appears noise figure identified, you can easily estimate system sensitivity from inductive (the return trace reflects upwards). the system bandwidth. Sophisticated handheld analyzers quickly and accurately One key performance indicator for a receiver is its sensitivity — the characterize an entire cable transmission system, as well as the individual ability to reliably discern small signals close to the noise floor. Closely components in the system. With DTF and TDR capabilities available at related is the system’s signal-to-noise ratio. Lower noise figure values the touch of a button, you can quickly pinpoint the location and type typically mean better device performance. of damage. These capabilities also help verify the performance of a Internally-generated noise, however, can degrade device performance. Internally-generated noise reduces the link budget and THE YELLOW TRACE SHOWS A DTF MEASUREMENT OF A CABLE WITH AN OPEN END. THE BLUE force use of a more powerful transmitter or a more expensive antenna TRACE, STORED TO MEMORY, SHOWS THE SAME CABLE TERMINATED IN A 50-Ω LOAD. the receiver. A complete picture of system performance requires an evaluation of internally-generated noise. Minimization of receiver noise is the most costeffective way to optimize communication systems without reducing quality. The ability to perform noise figure measurements, in addition to network analysis, spectrum analysis, and power sensor capabilities, enables the complete characterization of amplifiers and converters in the field. Handheld analyzers typically make noise figure measurements using the “Y-factor” method. This technique allows you to measure system components such as amplifiers, downconverters, and upconverters. You can easily view the change in uncertainty in real time, with a built-in uncertainty calculator that displays vertical bars over the trace data. The ability to make these measurements quickly for characterization of the noise figure is important for quickly optimizing designs. Cable and antenna measurements (CAT) verify and troubleshoot RF/microwave/ mmWave transmission systems and antennas. These measurements take place along the coaxial cable that
NOISE FIGURE MEASUREMENTS
eeworldonline.com
|
designworldonline.com
6 • 2021
DESIGN WORLD — EE NETWORK
27
TEST & MEASUREMENT HANDBOOK
single antenna at the installation site with signal reflection, return loss, and voltage standing wave ratio functions. When there are multiple antennas at one site, handheld analyzers can verify the antenna-to-antenna isolation, whether the antennas are associated with the same system or with different systems.
OVER-THE-AIR (OTA) TESTING
OTA measurements assess the level of cell coverage needed to ensure continuous connectivity in various mobile communication scenarios, including voice, text messages, and data services. Wireless networks continue to grow increasingly complex, especially with pioneering technologies such as 5G. Network coverage is a significant challenge because today’s wireless networks consist of layers of macrocells, microcells, and picocells. With users shifting between LTE and 5G, operators face difficulties in defining and troubleshooting wireless coverage.
28
DESIGN WORLD — EE NETWORK
OTA antenna testing in the field is the best way to verify that each cell has enough neighbors for successful handovers. OTA measurements allow scanning an area to determine how many cells are available, identify which cells are good neighbors, and troubleshoot handover problems such as missing neighbors. OTA applications enable LTE and 5G New Radio (5G NR) demodulation to give you insights on cell coverage. This information includes physical cell ID and control channel (often referred to as component carrier) metrics on any given frequency for all available cells. OTA measurements also help address the common problem of identifying missing neighbors. Some analyzers provide a useful capability: They display the strongest cell on different component carriers. This capability expedites the process of selecting the best frequencies for any given location to optimize inter-frequency handover. 6 • 2021
THE 5G NR OTA SCAN RESULTS FROM KEYSIGHT’S FIELDFOX HANDHELD ANALYZER SHOWING A TABLE WITH DEMODULATED INFORMATION AND A BAR CHART WITH CELL SIGNAL STRENGTH.
EVALUATING ELECTROMAGNETIC FIELD (EMF) EXPOSURE
Operators must verify EMF exposure levels for compliance and often use EMF-specific measurements and a triaxial antenna to do so in the field. As technology has developed over time, the RF spectrum has become more crowded. Currently, many commercial technologies function in the low-band radio frequencies. As consumers add more and more smart devices to their lives, the push to explore higher frequencies grows larger. The characteristics of 5G signals require more base station antennas than LTE — especially in densely populated areas. Besides needing numerous antennas, 5G mmWave eeworldonline.com
|
designworldonline.com
RF FIELD MEASUREMENTS
A VISUALIZATION OF THE ELECTROMAGNETIC SPECTRUM AND THE CORRESPONDING TECHNOLOGY FREQUENCIES. IMAGE COURTESY, THE INTERNATIONAL TELECOMMUNICATION UNION (ITU).
signals have different EMF properties than previous standards. Consequently, operators must verify EMF exposure levels in the field, and companies implementing 5G must verify their EMF levels during deployment. Exposure limits for EMF radiation differ by country. Many countries base their regulations on findings from organizations such as the International Commission on Non-Ionizing Radiation Protection (ICNIRP), the Institute of Electrical and Electronics Engineers (IEEE), and the US Federal Communications Commission (FCC). The installation and maintenance of modern communications systems often requires in-field verification and adjustment of components such as filters, duplexers, or antennas. As OTA systems become more complex and evolve, field engineers must carry a handheld instrument that performs the tests necessary to keep a network up and running — and they must know how to use it. An understanding of the measurement basics and technologies gives field engineers the tools to handle the challenges that RF networks present.
REFERENCES
Keysight Technologies, www.keysight.com
Proven integrity AND industry know-how Electrocube is one of the most respected design manufacturers of passive electrical component products for a wide range of standard and custom applications – from aerospace and audio to elevators and heavy equipment – as a capacitor supplier, resistor-capacitor distributor, and more.
Bishop Electronics, Seacor, Southern Electronics, F-Dyne
ELECTROCUBE.COM | 800.515.1112 | INFO@ELECTROCUBE.COM
TEST & MEASUREMENT HANDBOOK
KEY CONSIDERATIONS FOR SPECTRUM ANALYZERS A BEVY OF SPECIFICATIONS DEFINE HOW SPECTRUM ANALYZERS PERFORM. BUT JUST A HANDFUL OF QUALITIES CAN BE PIVOTAL WHEN IT COMES TO SELECTING INSTRUMENTATION. SIGNAL HOUND
30
DESIGN WORLD — EE NETWORK
AS
wireless technologies pervade our lives, more engineers find they must make RF measurements. In the past, RF testing required the use of expensive
equipment handled only by engineers with deep RF knowledge. Today, however, engineers and technicians with a wide range of backgrounds must measure RF. Spectrum analyzers are the most common lab instrument for measuring RF signals. Historically, these have been expensive instruments protected by the resident RF guru. But advances in RF integrated circuits have made it possible to push the envelope in spectrum analyzer architecture. Many companies are surprised to learn that modern RF instrumentation needn’t be expensive. Today, the approach is often to buy one physically large and costly spectrum analyzer for a few challenging requirements and then stock every bench with less expensive PC-based analyzers that can handle 95% of the workload.
6 • 2021
eeworldonline.com | designworldonline.com
PC-BASED SPECTRUM ANALYSIS Dynamic range is an important specification for any spectrum analyzer. When looking for low-level signals, it determines your ability to distinguish a signal from the noise floor. An important consideration will be determining the lowest level your test situation demands. For many measurements, the main objective may be to accurately characterize the main signal. In these cases, a 60 dB signal-to-noise floor may be more than enough. In some labs, a few test stations may require a more expensive spectrum analyzer that offers 85 dB or more. Spectrum analyzer dynamic range specifications can be confusing. Often, instruments display a low number for DANL, or Displayed Average Noise Level (e.g., less than -150 dBm/Hz). The DANL numbers can sometimes be misunderstood, as they depend on many factors such as frequency, attenuation, detectors, preamps, etc. For a simple comparison among spectrum analyzers, look at the maximum dynamic range specified as two-thirds of the difference between the third-order intercept point and DANL at 1 GHz. In many applications, sweep speed is of critical importance. For example, spectrummonitoring situations often require broad frequency sweeps in search of a variety of signal situations. For example, Signal Hound’s SM200B offers a 1 THz/sec sweep speed at any of its resolution bandwidth settings ≥30 kHz. Covering 1 GHz to 20 GHz in just 19 msec allows for a constant sweep of the airwaves. Best of all, this can take place automatically, with no operator present, over long periods of time. Simply define a baseline and any signals that violate it will be logged to a CSV file in real-time. This maximizes both efficiency and security, as data is preserved even if the computer shuts down. The Signal Hound spectrum analyzer architecture also allows for an additional technique to further boost speed. In many cases the instrument software can consume computer processor overhead. An application like Signal Hound’s Spike spectrum analyzer software may have a small impact on overall test times. In cases where fractions of a second are critical, Signal Hound allows its users to bypass Spike to allow for direct device API programming and faster measurements. Signal Hound spectrum analyzers can be programmed in C++, LabView, Matlab, Python, C# or any language that has C bindings. eeworldonline.com | designworldonline.com
AN ILLUSTRATION OF NOISE FLOOR FOR TWO DIFFERENT PC-BASED SPECTRUM ANALYZERS, THE SIGNAL HOUND SM200B AND BB60C.
PHASE NOISE MEASUREMENTS
Many devices and systems require accurate phase noise measurements. Spectrum analyzers are commonly used for this measurement. However, the spectrum analyzer itself must have phase noise sufficiently low so as not to contribute to the device measurement. For example, the Signal Hound SM200B uses a low-IF architecture design that enables exceptional phase noise performance, comparable to the performance of a more expensive spectrum analyzer. The Signal Hound SA44 and BB60 use a more traditional superheterodyne architecture to achieve a
more modest phase noise performance but less expensively. The low-IF architecture of the SM200B does come with a trade-off of higher image response spurs. While the residual response spurs are comparable to the MXA, the image response spurs are not as low as those of the MXA (<-74 dBc). The SM200B Spike software provides a Signal ID feature able to be activated and deactivated, allowing the differentiation of low-level mixer spurs from RF Input signals. In many cases the spurs can be identified as coming from the spectrum analyzer and not the device-under-test (DUT), allowing them to be ignored during the actual measurement. The
THE SIGNAL HOUND SPIKE SPECTRUM ANALYSIS SOFTWARE HAS PHASE-NOISE MEASUREMENT CAPABILITY BUILT IN RATHER THAN AS A PREMIUM FEATURE.
6 • 2021
DESIGN WORLD — EE NETWORK
31
TEST & MEASUREMENT HANDBOOK DYNAMIC RANGE MAX DYNAMIC RANGE @1 GHZ
PRICE
Signal Hound SA44B 4.4 GHz, 250 kHz IBW
104 dB
$1,020 USD
Signal Hound BB60C 6 GHz, 27 MHz IBW
95 dB
$3,040 USD
Signal Hound SM200B 20 GHz, 160 MHz IBW
118 dB
$12,990 USD
Keysight N9020B-526-B1X (MXA) 26.5 GHz, 160 MHz IBW
116 dB
$85,389 USD
ANALYZER
A COMPARISON OF KEY SPECS FOR PC-BASED SPECTRUM ANALYZERS FROM SIGNAL HOUND AND A KEYSIGHT MXA. A POINT TO NOTE: IT IS POSSIBLE TO OBTAIN A SPECTRUM ANALYZER ABLE TO VIEW WIFI AND OTHER ISM-BAND EMISSIONS DIRECTLY FOR AROUND $1,000.
SSB PHASE NOISE FOR A 1 GHZ CARRIER KEYSIGHT N9020B 26.5 GHZ MXA
SIGNAL HOUND OFFSET
SA44B (TYP.)
BB60C (SPEC.)
10 Hz
SM200B (SPEC.)
SPEC.
-76
100 Hz
-80
-70
-108
1 kHz
-88
-76
-123
10 kHz
-91
-83
-132
-113
100 kHz
-100
-93
-136
-116
-133
-135
1 MHz
more traditional superheterodyne architecture of the BB60C typically has -70 dBc image rejection. The BB60C spurs are generally not from the image response. For many real-world signals—from complex modulated communications signals, to interference events, to pulsed tactical signals— the signal energy can be sporadic, non-recurring, or even random. With traditional spectrum analysis, it could be nearly impossible to catch
-91
COMPARING PC-BASED SPECTRUM ANALYZERS AND DEDICATED VERSIONS FOR SSB PHASE NOISE. THE SIGNAL HOUND SM200B USES A LOW-IF ARCHITECTURE DESIGN THAT ENABLES EXCEPTIONAL PHASE NOISE PERFORMANCE. THE SIGNAL HOUND SA44 AND BB60 USE A MORE TRADITIONAL SUPERHETERODYNE ARCHITECTURE AND REALIZE A MORE MODEST PHASE NOISE BUT AT A MODEST PRICE. NOTE THAT FOR THE SA44, THE RESULTS LISTED ARE TYPICAL AS IT DOES NOT HAVE A HARD SPEC FOR PHASE NOISE.
these signals in an analysis window and trigger on them. Modern communication modulation schemes are increasing the challenge further with techniques such as frequency hopping, spread spectrum, pulsed, and cognitive radio low probability of intercept techniques.
Real-time spectrum analysis (RTSA) is a digital signal processing method that leverages overlapping FFTs and high-speed memory to have a 100% probability of intercept (POI) in even extremely dense environments. Real-time bandwidth, which is the maximum frequency
DIGITAL MODULATION ANALYSIS CARRIED OUT ON A PC-BASED SIGNAL HOUND SPECTRUM ANALYZER. DIGITAL MODULATION ANALYSIS CAPABILITIES INCLUDES CONSTELLATION DIAGRAMS AND SYMBOL TABLES FOR MODULATION FORMATS SUCH AS QPSK, BPSK, 8PSK, ∏/4DQPSK, DQPSK, AND QAM16/32/64/256. MEASUREMENTS INCLUDE: ERROR VECTOR MAGNITUDE (EVM), SIGNAL-TONOISE RATIO (SNR), MODULATION ERROR RATIO (MER), MODULATION QUALITY METRICS, LINEAR COMPENSATIONS SUCH AS CARRIER OFFSET, I/Q OFFSET, AMPLITUDE DROOP (LINEAR AMPLITUDE CORRECTIONS), SYNC PATTERN TRIGGERING, AND EYE DIAGRAMS. 32
DESIGN WORLD — EE NETWORK
6 • 2021
eeworldonline.com
|
designworldonline.com
PC-BASED SPECTRUM ANALYSIS
THE PERSISTENCE AND WATERFALL DISPLAY IN REAL-TIME ANALYSIS SHOW THE OCCUPANCY OF THE 2.4 GHZ ISM BAND. ON SCREEN ARE THE TRANSMISSIONS OF A BLUETOOTH HEADSET AND A CELL PHONE SEARCHING FOR A WIRELESS NETWORK.
span offering gap-free overlapping FFT processing, is an important parameter of an RTSA that can enable a more detailed analysis of a spectrum. Real-time no longer equals expensive when it comes to PC-based spectrum analysis. For example, Signal Hound offers RTSA capabilities up to 160 MHz with a 100% probability of intercepting signals as fast as 12-µsec. However, in many cases you don’t even need 160 MHz of analysis bandwidth. For example, suppose you are testing a device having a 25-kHz maximum bandwidth signal, such as a key fob. You may just need to push the key and make sure that the center frequency, bandwidth, and modulation are right. Complex signal analysis capabilities include adjacent channel power ratio (ACPR) or adjacent channel leakage ratio (ACLR), occupied bandwidth (OBW), and channel power measurements. For example the 27 MHz of instantaneous bandwidth provided by devices such as the BB60C enables real-time OBW and ACPR measurements of very wide-bandwidth signals, transient or continuous. For applications where you need more instantaneous bandwidth, the SM200B offers 160 MHz. Another consideration is the separation is the separation of the hardware systems. Traditionally, test instrumentation utilizes a built-in PC taking the form of special internal controller boards produced at a much lower volume than commercial PCs and
CONNECT WITH US! CHECK US OUT on ISSUU.COM!
eeworldonline.com Signal Hound — TM HB 06-21_converted.indd 33
6/8/21 11:41 AM
TEST & MEASUREMENT HANDBOOK REAL-TIME SPECTRUM ANALYSIS CAPABILITIES ANALYZER
REAL-TIME BANDWIDTH 100% PROB OF INTERCEPT (POI)
Signal Hound SA44B 4.4 GHz
250 kHz
592 μs @ 10 kHz RBW 4.7 ms @ 1 kHz RBW
Signal Hound BB60C 6 GHz
27 MHz
19.2 μs @ 300 kHz RBW 38.4 μs @ 100 kHz RBW
Signal Hound SM200B 20 GHz
160 MHz
12 μs @ 300 kHz RBW 49 μs @ 100 kHz RBW
Keysight N9020B-526-B1X (MXA) 26.5 GHz
85 / 125 / 160 MHz
17.3 μs (RBW not specified)
COMPARING REAL-TIME ANALYSIS CAPABILITIES OF PC-BASED SIGNAL HOUND AND TRADITIONAL SPECTRUM ANALYZERS.
typically manufactured by a third-party. It is not unusual for these internal controllers to be two to five generations behind their commercial peers because they are embedded. This invariably means that by the time the spectrum analyzer reaches the market, the internal controller is dated. Additionally, these built-in controllers are usually only updated when the instrument is replaced – meaning that by the end of an instrument’s life cycle, the user is dealing with a processor that is several generations slower than current PCs. The ability to use a current-generation PC will keep the overall performance of your test equipment much closer to current standards. Now factor in the cost of the traditional internal controller and mark it up by at least four times the vendor’s cost to arrive at the customer’s cost. USB-based test instruments eliminate this issue with the use of an external PC as the measurement controller. And of course, use of an external PC to control the measurement allows all collected data to reside on the PC.
REFERENCES
Signal Hound real-time USB-powered spectrum analyzers, signalhound.com/learn.
ADJACENT CHANNEL POWER RATIO (ACPR) MEASUREMENTS ON BOTH BB60C AND SM200B PC-BASED SPECTRUM ANALYZERS. NOTE THE SM200B HAS A NOISE FLOOR 22 DB LOWER.
34
DESIGN WORLD — EE NETWORK
6 • 2021
eeworldonline.com | designworldonline.com
SEARCH MILLIONS OF PARTS FROM THOUSANDS OF SUPPLIERS
PRICING & AVAILABILITY
DATA SHEETS & SPECS
SOURCE & PURCHASE
Get real-time pricing and stock info from authorized distributors and manufacturers.
View and download product data sheets and technical specifications.
Compare options from suppliers and buy direct from distributors and manufacturers.
ABOUT DESIGNFAST
DesignFast is a search engine for finding engineering components and products. With DesignFast, engineers and sourcing professionals can quickly search for products, compare prices, check stock, view data sheets and go direct to the supplier for purchase.
HOW DOES IT WORK?
DesignFast aggregates product data from thousands of suppliers and distributors and makes it available for searching. DesignFast provides pricing, availability and product data sheets for free download.
designfast.com
TEST & MEASUREMENT HANDBOOK
SORTING OUT PC-BASED INSTRUMENTATION TREVOR SMITH, PICO TECHNOLOGY
IT USED TO BE THE CASE THAT ONLY A FEW KINDS OF RELATIVELY SIMPLE INSTRUMENTS WERE BASED ON PCs. NOW SEVERAL CATEGORIES OF LAB-GRADE INSTRUMENTATION TAKE THE PC APPROACH.
THE DIVERSITY OF
measurements that must be made by electronics engineers
constitutes a huge challenge. The debugging and validation of components and systems calls for versatile instruments with wide-ranging functions and easy programmability. Engineers working on high-speed logic one day might be called upon to run a long-duration soak test on the next and check for immunity to power harmonics on the day after that. Such chores require versatile tools that span a broad range of tasks with a minimum of fuss.
THE PICOSCOPE 6824E OSCILLOSCOPE, HERE CONNECTED TO HOST PC, AND THE SCOPE’S PCB CARRYING THE FPGA THAT HANDLES MUCH OF THE WORK. THE AVAILABILITY OF HIGH-PERFORMANCE FPGAS IS ONE FACTOR THAT MAKES PC-BASED SCOPES PRACTICAL
36
DESIGN WORLD — EE NETWORK
6 • 2021
eeworldonline.com
|
designworldonline.com
PC-BASED INSTRUMENTS A SCREEN SHOT GENERATED BY PICOSCOPE 6 USER INTERFACE SOFTWARE THAT PROVIDES THE FUNCTIONS OF AN OSCILLOSCOPE, SPECTRUM ANALYZER, SERIAL PROTOCOL ANALYZER, LOGIC ANALYZER, FUNCTION/SWEEP GENERATOR, AND ARBITRARY WAVEFORM GENERATOR. PC-based test and measurement products address those needs. These instruments utilize industry-standard PC hardware for instrument control, analysis and display of measurement results. Instrument acquisition hardware sits in a compact enclosure that connects to the host PC, typically with a USB 2.0 or 3.0 connection. Connectors for the input and output signals to and from the device under test (DUT) are provided as is a separate power connection, where required. Hardware design typically centers around advanced FPGA technology that controls the channel settings, signal conditioning, triggering, on-board memory management and communication with the host PC. The FPGA does the heavy-duty processing of test data acquisition and, if required, signal generation. Data flowing on the USB interface is limited to commands, responses and composite information required for processing and display of graphical or numerical test data. Separating the acquisition hardware from the analysis and display platform provides several advantages. Software typically installs on any Windows 10 machine as well as on Linux and macOS industry-standard platforms, so users have the freedom to choose a desktop machine in the lab or a laptop at home or in the field. Most Pico software packages can take advantage of high-definition and UHD
displays for clear presentation of waveforms, measurements and analysis results. Custom applications are also possible. For example, every Pico PC-based instrument comes with a free software development kit (SDK) that includes a programmer’s guide and code examples in widely used languages such as C#, C++ and Python. The application programming interface exposes the full instrument hardware so programmers can develop their own code to control the instrument. It’s possible to find PC-based oscilloscopes featuring real-time bandwidths from 10 MHz to 1 GHz and a choice of two, four or eight analog channels, plus up to 16 digital channels on MSO models. There are flexible- and high-resolution models too, with up to 16 bits for precision analog
measurements. Many models have an integrated function generator and arbitrary waveform generator. For example, all PicoScope real-time oscilloscopes come with PicoScope 6 user interface software that delivers six instrument functions: an oscilloscope, spectrum analyzer, serial protocol analyzer (21 serial decoder/analyzers included, with more in development), logic analyzer (on MSO models), function/sweep generator, and an arbitrary waveform generator. PC-based scopes such as the PicoScope 2000 Series feature a small enclosure that fits easily in a laptop bag. These scopes are well suited for education, sporting bandwidths from 10 to 100 MHz and two channels or 2+16 channels on MSO models. More generalpurpose scopes typically provide 200 MHz bandwidths. Examples include the PicoScope 3000 Series with two or four channels plus 16 digital channels on MSO models. Specific models of PC-based scopes are optimized for subsets of requirements common to test and measurement scenarios. In applications that need precision measurements ranging from a few millivolts to hundreds of volts, high-precision scopes with fixed 12- and 16-bit resolutions are available with over 70
THE PICOLOG 6 DATA ACQUISITION SOFTWARE HANDLES DATA LOGGING AND IS COMPATIBLE WITH WINDOWS, MACOS, LINUX AND RASPBERRY PI OS. PICOLOG 6 WORKS WITH ALL PICOSCOPE REAL-TIME OSCILLOSCOPES AS WELL AS WITH PICO DATA ACQUISITION PRODUCTS.
eeworldonline.com | designworldonline.com
6 • 2021
DESIGN WORLD — EE NETWORK
37
TEST & MEASUREMENT HANDBOOK dB dynamic range and differential-inputs. Some versatile scopes can make high-speed timing measurements with 1 nsec resolution at 8-bit vertical resolution or 12/14/15/16-bit vertical resolution at lower sampling rates. Also available are high-performance, deep-memory scopes for demanding signal integrity and other scientific applications. These devices can provide up to 1 GHz bandwidth and 4 GS capture memory for big-data analysis of waveforms. Even high bandwidths are now available on PC-based scopes. An example is the PicoScope 9000 Series providing bandwidths to 25 GHz. Clock recovery enables characterization of high-speed NRZ/RZ data signals as eye diagrams to reveal jitter, noise, rise time and other important parameters. TDR/TDT functions help characterize high-speed communication channels through cables, connectors, backplanes and PCBs. Models with optical inputs can perform eye diagram measurements with automatic measurement of parameters including extinction ratio, S/N ratio, eye height and eye width. As with dedicated bench instruments, PC-based instrumentation makers also provide specialized probes for tricky measurements. Examples include the PicoConnect 900 Series low-impedance RF, microwave and gigabit pulse probes allowing fingertip browsing of broadband signals up to 5 GHz (10 GB/sec), A3000 Series active probes to 1.3 GHz bandwidth, and high-voltage differential probes to 7,000 V. Vector network analyzers also come in the form of PC-based instruments. Examples include the PicoVNA 108 8.5 GHz and PicoVNA
A NETWORK METROLOGY TRAINING (NMT) KIT HELPS PERFORM VNA CALIBRATION AND NETWORK MEASUREMENT TASKS. SUITABLE N AND SMA ADAPTERS, CALIBRATION STANDARDS, TEST LEADS AND FIXED WRENCHES ARE ALL INCLUDED.
38
DESIGN WORLD — EE NETWORK
106 6 GHz instruments able measure all four S-parameters at each frequency point in under 190 µsec—thus producing a 201-point two-port .s2p Touchstone file in less than 38 msec. The PicoVNA 108 delivers a dynamic range of 124 dB at 10 Hz (118 dB for the PicoVNA 106) and less than 0.006 dB RMS trace noise at its maximum operating bandwidth of 140 kHz. These instruments can also work as deep-dynamic-range scalar network analyzers or single-port vector reflectometers as well as fullfunction dual-port, dual-path VNAs. Economical check standards can be used to validate the accuracy of a network analysis test setups and calibration before and after measurements. Akin to the Beatty line, each check standard is a short length of mismatched line (75 mm of 25 Ω) with a predictable, smooth and stable mismatch and transmission characteristic that spans the VNA frequency range.
DATA ACQUISITION
Another category of PC-based products is that for data logging. Data loggers require no power supply and simply plug into a USB or an Ethernet port on a PC or network. Data acquisition software handles measurements, recording, and data analysis. Moreover, data acquisition software, such as PicoLog 6, works with all PicoScope real-time oscilloscopes as well as with acquisition products. Examples of these include the TC-08 eight-channel thermocouple data logger. It can measure from –270 to 1,820 °C (–454 to 3,308 °F) with high resolution and accuracy and is expandable to 20 units/ 160 channels. Another device called the PT-104 platinum resistance temperature data logger measures temperatures ranging from –200 to 800 °C with either PT100 or PT1000 sensors. It uses high-precision reference resistors for high stability and realizes up to 0.001 °C resolution and 0.015 °C accuracy. High-resolution data loggers such as the ADC-20 and ADC-24 detect small signal changes precision voltage data loggers deliver 20 and 24-bit resolution respectively. Features such as true differential inputs, galvanic isolation and software-selectable sampling rates all contribute to noise-free resolution. Multi-purpose data loggers are also available, typically with a choice of 10 or 12-bit input resolution. And there are special-purpose data loggers. An example is the PicoLog CM3 which is designed for measuring the current consumption of buildings and machinery and, in that regard, is suitable for single or three-phase ac.
REFERENCES
Pico Technology, www.picotech.com
6 • 2021
eeworldonline.com | designworldonline.com
7x more semis 3x more passives 2x more interconnect Fast & free shipping* Choose from a new and expanded range of board-level components with free shipping on orders over $150 USD & faster delivery. SHOP THE WAY YOU PREFER
newark.com order@newark.com / quote@newark.com
1 800 463 9275 7:00 a.m. – 8:00 p.m. M-F (EST)
newark.com/technical-support 1 877 736 4835
E-PROCUREMENT SYSTEMS We can link with any eProcurement system, and our dedicated e-team will facilitate your implementation. Email: eproc@newark.com
GLOBAL REACH, LOCAL RESOURCES · 10 million products available, 950,000 in stock · Supporting 46 websites in 27 languages · Authorized distributor for 3000+ brands globally
SHIPPING Order
Time
Domestic
M-F by 9:00 p.m. (EST)
Export M-F by 6:00 p.m. (EST) In-stock products normally ship the same day
NO MINIMUM ORDER FEES
NEW! FREE SHIPPING, FASTER DELIVERY, GLOBAL INVENTORY Access our global inventory with free shipping on orders over $150 USD & faster delivery *Some exclusions apply, for more details see: www.newark.com/help-delivery-information
YOU CAN FIND US AT…
TEST & MEASUREMENT HANDBOOK
MEASURING PICOSECONDS WITHOUT BREAKING THE BANK MODERN ANALOG/DIGITAL CONVERTERS CAN HELP IMPLEMENT EQUIVALENT TIME SAMPLING TO PROVIDE PICOSECOND TIMING RESOLUTION.
DAVE GUIDRY, TEXAS INSTRUMENTS INC.
A PICOSECOND IS AN
astonishingly short period of time. Imagine a stopwatch with a trillion divisions between each tick of the second hand. While it sounds fantastical, this resolution of time
TIME INTERVALS DEFINED
measurement is indeed necessary for a diverse set of applications, including quantum computing, particle physics, automatic test equipment (ATE) and phased-array radars. Measurement of time on this scale has traditionally been a quite complex and expensive endeavor requiring specialized instrumentation. Now, features in modern analog-to-digital converters (ADCs) make picosecond-resolution time measurement possible without breaking the bank. Time measurement is often defined as the difference in time between a start and a stop event. In modern electronics, the time interval will likely be between two highspeed digital signals where the event is the moment when the signal transitions through a predefined threshold. Instruments that can precisely measure the time difference between two events are known as time interval counters or time measurement units. These can be quite complex and expensive instruments, especially when picosecond resolutions are required. A generic time interval measurement does not assume a relationship between the start event or stop event. Thus, the instrument must fully resolve the time difference between two asynchronous events. There is no easy way to accomplish this task with
40
DESIGN WORLD — EE NETWORK
discrete circuitry; only specialized – and likely expensive – hardware will work. Many time-interval measurement applications do, however, involve measuring the difference between events that have a known relationship. For instance, the skew between two outputs of a clock distribution network will not only have the same frequency as each other, but also the same frequency as
the input signal. Leveraging this fact enables the use of equivalent time sampling.
EQUIVALENT TIME SAMPLING EXPLAINED
Sampling theory specifies the need for at least two samples per period of the highest frequency signal measured. This reality, defined by Nyquist, is impossible to avoid, but you can sidestep it if the input signal measured is repetitive. Consider an example involving an ADC. The nearby figure shows a fourfold increase in the effective sample rate from that of the ADC sample rate. Each triggered measurement
EQUIVALENT TIME SAMPLING. EACH TRIGGERED MEASUREMENT CAPTURES SIX SAMPLES PER CYCLE OF THE INPUT WAVEFORM.
6 • 2021
eeworldonline.com
|
designworldonline.com
EQUIVALENT TIME SAMPLING
ADC APERTURE DELAY ADJUST FEATURE. APERTURE DELAY IS THE TIME FROM THE CLOCK TRANSITION TO THE ADC’S CAPTURE OF THE INPUT SIGNAL captures six samples per cycle of the input waveform. Between triggers, I delayed the sampling clock for the ADC by one-fourth the sample period. I then took the four sets of six samples and interleaved them to produce a resulting waveform that boosted the sample rate fourfold. This scheme is widely used in digital storage oscilloscopes to increase the effective sample rate for repetitive signal measurement and can achieve sample rates of hundreds or even thousands of gigasamples per second. An alternate to equivalent time sampling is known as coherent undersampling. Instead of using an analog delay to adjust the sampling instant between each triggered capture, undersampling changes the sample rate slightly to create a beat frequency between the input frequency and the clock frequency. This technique has the same effect as equivalent time sampling (in other words, each input period slips a small amount of time), but does not require an adjustable analog delay. The downside to this approach is that it requires a synthesizer with a high-frequency resolution to create the sample clock frequency. Fortunately, features of Texas Instruments high-speed ADCs can help implement equivalent time sampling without any special clock synthesizer or external adjustable analog delay components. TI has released three high-speed ADC families that employ a feature known as aperture delay adjust. Before diving into how this feature can provide picosecond resolution, a brief introduction is in order. Aperture delay is the time from when the clock transitions until the ADC sample-and-hold captures the input signal. Every ADC experiences aperture delay because its circuitry has a finite propagation delay, but the TI family of high-speed ADCs add
DELAY CALIBRATION SETUP 6 • 2021
DESIGN WORLD — EE NETWORK
41
PHASE OFFSET ERRORS PHASE OFFSET (DEGREES)
PHASE ERROR TIMING ERROR (DEGREES) (fs)
-30
0.0375
104.2
-20
0.0120
33.3
-10
0.0062
17.2
10
-0.0015
-4.2
20
0.0022
6.1
30
-0.0384
-106.7
THE TABLE SUMMARIZES THE ADC PHASE ERROR BETWEEN SETTINGS WHEN USING A ROHDE & SCHWARZ SGS100A 12-GHZ RF GENERATOR, WHICH HAS A PHASE OFFSET CAPABILITY WITH 0.1° RESOLUTION, AND AFTER REMOVING THE SYSTEMATIC OFFSET.
TEST & MEASUREMENT HANDBOOK
APERTURE DELAY ADJUST ERROR VS. SETTINGS AFTER CALIBRATION FOR EACH 1-PSEC DELAY SETTING AFTER CALIBRATION.
a feature to make aperture delay adjustable, with an astounding 19-fsec resolution. A femtosecond is 1/1,000th of a picosecond, or a quadrillionth of a second. The aperture delay-adjust feature is implemented in such a way that it is largely unaffected by temperature, process or supplyvoltage variations, and has minimal impact on sampling jitter. This quality is especially useful for time interval measurements because there is a high-resolution delay feature that can be used to implement equivalent time sampling. These ADCs have an inherent typical aperture delay of 360 psec, with the aperture delay adjust feature set at its minimum. Up to about 600 psec of additional aperture delay can be added by setting 8-bit coarse and fine settings, with 1.13-psec and 19-fsec step sizes, respectively. For sample rates below about 1.67 GSPS, the sample clock can be inverted to boost the delay range by half of the sample clock period. This means that any sample rate over about 833 MSPS will have full coverage across the sample period to adjust the sampling point.
CALIBRATING DELAY
SETUP FOR VALIDATING TIME MEASUREMENT
TIMING VALIDATION-RESULTS CHANNEL B SIGNAL GENERATOR SET TO PHASE OFFSET OF -30 TO 30° TO 30 IN 10° STEPS, COMPARED TO THE PHASE OF CHANNEL A. 42
DESIGN WORLD — EE NETWORK
6 • 2021
Though it is stable, the aperture delay-adjust feature is not factorycalibrated. But one can easily employ a calibration routine to select the optimal coarse and fine aperture delay settings for each desired delay value. For example, operating the ADC12DJ5200RF in dualchannel mode at 5 GSPS results in a 200-psec sample period, which the aperture delay-adjust feature then further subdivides into 1-psec time slices, yielding an effective sample rate of 1,000 GSPS. The most straightforward way to calibrate the delay settings is to use a fixed coherent input frequency with respect to the sample clock where you can measure changes in phase (and thus time) across aperture delay settings. The input need not be high fidelity as long as it is phase-stable, an integer frequency ratio, and sufficiently low-jitter with respect to the sample clock. I selected 1-GHz for the purposes of this article, which results in five samples per period of the 5-GSPS clock. The nearby figure illustrates the setup for calibration. I adjusted the input signal amplitude to -1 dBFS to optimize the signal-tonoise ratio – but that adjustment is not a strict requirement. I captured 50,000 samples for each measurement, then divided the measurements into 10,000 arrays of five samples each. Averaging these arrays sample by sample resulted in a five-sample array. This five-sample array represents one cycle of the calibration signal averaged 10,000 times. Taking the unwindowed Fast Fourier transform of this array results in an optimal estimation of the phase of the signal. If you trigger subsequent captures to start at the same place, multiple captures can then be further averaged to mitigate the impact of low-frequency noise (also known as 1/f noise). Remember: You are calibrating the differential delay between what is defined as the “zero” delay setting and each relative delay setting. The absolute delay over any significant period of time will drift because of temperature changes in the interconnect cables, clock and inputsignal generators. While these changes may be seem small, without mitigation, the calibration will skew by as much as several picoseconds. The solution to this problem is thankfully straightforward. For each delay setting, simply measure the delta delay between the zero-delay setting and the coarse and fine settings under inspection. Repeating this measurement a sufficient number of times will mitigate eeworldonline.com
|
designworldonline.com
EQUIVALENT TIME SAMPLING
I set the channel B signal generator to a phase offset of -30° to +30° in 10° steps. At each offset, I captured and compared the signal to the phase of channel A, which was left at its initial setting.
A PRACTICAL EXAMPLE
IMPACT OF ADAPTERS ON TIMING SKEW. THE “WELL-MATCHED” DISTRIBUTION PATH HAS ABOUT 9 PSEC OF SKEW. ADDING AN SMA ADAPTER INCREASES SKEW BY 6 PSEC/ADAPTER. low-frequency noise until you’ve attained the desired precision. Again, the coarse and fine delays have approximately 1.13-psec and 19-fsec step sizes, respectively. If you are searching for 1-psec steps, a number of possible combinations of coarse and fine delay will result in several delay values close to your target. This procedure helps realize a convergent calibration, as there are multiple ways to get to each 1-psec step. The nearby figure shows the resulting residual error for each 1-psec delay setting after calibration. At less than ±60 fsec, it is difficult to appreciate the small size of this error. In a vacuum, light travels a mere 18 µm in 60 fsec – and perhaps only 60% to 70% of this distance – in a typical high-speed circuit board. In the setup used to validate the calibration of the delay settings, each channel of the ADC is driven with an independent 1-GHz signal generator. Both signal generators are locked to the same common frequency reference as the ADC clock generator. This phase locking ensures phase stability between the ADC inputs and enables the use of coherent sampling. The synchronization of captures to the input frequency ensures that the recovered waveform has the same phase each time. Before collecting data, I adjusted the phase and amplitude of the generator driving channel B to minimize the difference between the channels. I used a Rohde & Schwarz SGS100A 12-GHz RF generator, which has a phase offset capability with 0.1° resolution. With this adjustment, it was possible to get both signals aligned within about 0.5 psec.
To relate the practical reality of tightly aligning clock signals in a distribution network, I employed the output of a LMK04832 clock dual-loop jitter cleaner and clock distribution IC to convert the sine wave from the RF signal generator into a low-voltage positive emittercoupled-logic square wave. I used an RF power divider to split this signal into two copies, which were then connected to both ADC channels using six-inch phase-matched cables. The “well-matched” distribution path has about 9 psec of skew, which is attributable to the combination of mismatch in the cables, splitter, PCB interconnect and ADC itself. Adding a subminiature version A (SMA) “connector saver” adapter, which seems innocuous, causes an increase in skew of 62 psec to 6 psec per adapter. This figure is quite close to what you’d expect for an addition of about a half-inch-worth of transmission line with a PTFE dielectric. This skew further enforces the difficulty in trying to tightly match timing across high-speed systems and the need for a high-resolution instrument to measure skew. In the example, it’s clear that a simple RF adapter may be enough to consume a significant portion of the timing budget. All in all, equivalent time sampling is a simple and cost-effective way to realize high-resolution time measurement of repetitive signals. The aperture-adjust feature in TI’s family of high-speed ADCs enables 1-psec time resolution and precision in the order of ±100 fsec. The ADC12DL3200, ADC12DJ3200 and ADC12DJ5200RF families all incorporate this aperture-adjust feature.
REFERENCES
TI ADC data sheets: https://www.ti.com/lit/gpn/adc12dl3200 https://www.ti.com/lit/gpn/adc12dj3200 https://www.ti.com/product/ADC12DJ5200RF Clock dual-loop jitter cleaner, https://www.ti.com/product/ LMK04832?HQS=asc-dc-hsc-null-contrib-pf-null-wwe
ALL ELECTRONIC COMPONENTS HAVE SOME TEMPERATURE DEPENDENCE. DEPICTED HERE IS THE DRIFT IN TIMING MEASUREMENT BETWEEN ADC CHANNELS A AND B ACROSS TEMPERATURE. FOR THIS EXPERIMENT, THE INPUTS WERE HELD AT A CONSTANT 1 GHZ WHILE THE CASE TEMPERATURE OF THE ADC WAS SWEPT FROM 25° TO 85°C. CLEARLY, THIS ADC HAS NEGLIGIBLE DRIFT IN DIFFERENTIAL APERTURE DELAY BETWEEN THE TWO CHANNELS ACROSS A WIDE TEMPERATURE RANGE. JUST THE CHANGE IN THE AMBIENT TEMPERATURE ON THE BENCH DURING THIS MEASUREMENT LIKELY ACCOUNTS FOR A SIGNIFICANT PORTION OF THE DRIFT, AS THE CABLES EXPAND OR CONTRACT DIFFERENTLY BETWEEN CHANNELS A AND B.
eeworldonline.com | designworldonline.com
6 • 2021
DESIGN WORLD — EE NETWORK
43
WEBINAR SERIES
CUSTOM CONTENT IN A LIVE, INTERACTIVE OR ON-DEMAND FORMAT. CHECK OUT OUR WEBINARS TODAY: ■
designworldonline.com/category/webinars
■
fluidpowerworld.com/category/webinars
■
therobotreport.com/category/robotic-webinars
■
eeworldonline.com/category/webinars
■
solarpowerworldonline.com/category/featured/webinars
■
windpowerengineering.com/category/featured/webinars
■
medicaldesignandoutsourcing.com/webinars
WTWH MEDIA’S WEBINARS OFFER: • Coverage of a wide range of topics • Help engineers better understand technology or product related issues and challenges • Present educational material related to specific topics
Medical edical Design & OUTSOURCING
Webinars-FullPgAd.indd 1
4/8/20 4:16 PM
DATA ACQUISITION
RASPBERRY PI FOR DAQ THOUGH THEY ARE MORE CLOSELY ASSOCIATED WITH THE MAKER SPACE, RASPBERRY PI BOARDS CAN BE EFFECTIVE AT HANDLING STRINGENT INDUSTRIAL MEASUREMENTS. STEVE RADECKY, MEASUREMENT COMPUTING CORP.
RASPBERRY PI’S
original aim was far from the typical engineering lab.
Originally designed to be a cheap, almost disposable computer, Raspberry Pi was created to get kids interested in programming. Since its introduction in 2012, Raspberry Pi has become the third-best-selling general-purpose computer platform, trailing only the Microsoft Windows PC and Apple Macintosh. It’s estimated that up to 44% of Raspberry Pi’s are now bought by “industrial” customers. A powerful feature set along with the Raspberry Pi community and $35 price tag have continued to fuel this tremendous growth.
THE DAQ HAT PLUGS INTO THE 40-PIN GPIO HEADER ON A RASPBERRY PI COMPUTER.
eeworldonline.com | designworldonline.com
The current pandemic has also helped grow the Raspberry Pi market. With more people working from home, the demand for small, inexpensive engineering projects has soared. Raspberry Pi has become a good fit for these types of applications and has made programming and engineering even more accessible. Raspberry Pi does not have built-in test and measurement capabilities such as analog-to-digital converters (ADCs), digitalto-analog converters (DACs), or conditioned digital inputs and outputs (DIO). However, these capabilities can be added through the 40-pin GPIO header. A device that connects directly to the 40-pin header and stacks onto the Raspberry Pi is called a HAT (Hardware Attached on Top). Over the years, individuals have published open-source designs and small companies have sold low-cost HATs for a variety of tasks, including support for analog and digital I/O. These designs and products are adequate for the education and hobbyist/maker market but have some serious short comings for professional test and measurement applications. Most of these devices are provided partially assembled, without specifications or programming support, and without performance guarantees that can only be achieved with a thorough device validation process. Professional-quality measurement products are now available on the Raspberry Pi platform. An example is the series of DAQ HATs from Measurement Computing. These HATs offer similar specs and accuracy as traditional USB and
6 • 2021
DESIGN WORLD — EE NETWORK
45
TEST & MEASUREMENT HANDBOOK
DAQ HATS INCLUDE C/C++ AND PYTHON EXAMPLES AND AN OPEN-SOURCE LIBRARY.
Ethernet-based DAQ products from MCC with resolution up to 24-bits and sample rates up to 100 kS/sec. MCC offers five products designed for test and measurement applications that conform to the Raspberry Pi HAT standard. These devices provide data acquisition features like analog and digital I/O in a small, stackable format. MCC 118 is an eight-channel voltage measurement HAT. It allows
THERMOCOUPLE MEASUREMENT VIA RASPBERRY PI
46
DESIGN WORLD — EE NETWORK
6 • 2021
users to measure 12-bit data at an overall throughput of 100 kS/sec. Eight devices can be stacked on a single Raspberry Pi to create a 64-channel device capable of reading data at a combined rate of 320 kS/sec. MCC 128 features 16-bit resolution and includes eight analog inputs with a maximum sample rate of 100 kS/sec. Multiple gains ranges are also included which provides users the ability to make precision measurements. MCC 152 provides two 12-bit analog outputs along with eight 5-V or 3.3-V DIO channels enabling the creation of a full multifunction Raspberry Pi measurement and control system. MCC 134 is designed for temperature measurement applications and features four thermocouple input channels. A 24-bit A/D and cold junction compensation provide professional grade accuracy. Multiple thermocouple types are selectable on a per channel basis. Thermocouples provide a low-cost and flexible way to measure temperature, but measuring thermocouples accurately is difficult. Nevertheless, it is now possible to measure thermocouples accurately in the uncontrolled Raspberry Pi environment. MCC 172 is designed for sound and vibration applications and offers two IEPE input channels capable of measuring IEPE sensors like accelerometers and microphones without any additional signal conditioning. Inputs can be simultaneously sampled at up to 51.2 kS/sec per channel. DAQ HATs can be used in basic applications with just a few voltage input channels or more sophisticated applications with up 64 channels of multiple signal types. Up to eight MCC DAQ HATs can be stacked onto one Raspberry eeworldonline.com | designworldonline.com
DATA ACQUISITION
Pi. HATs are available with voltage inputs, thermocouple inputs, IEPE-based sensor inputs, analog outputs, and digital I/O, allowing users to configure multifunction, Raspberry Pi-based solutions. MCC DAQ HATs come with software libraries that support Python and C/C++ to facilitate quick and easy development. Comprehensive API and hardware documentation are also provided. The DAQ HAT software library was created and is supported by MCC. The development repository is located on GitHub where users can find libraries, examples, firmware updates, and more.
APPLICATIONS FOR RASPBERRY PI DAQ
Raspberry Pi is a good fit for many DAQ applications. MCC DAQ HATs have been used in a variety of applications and industries. These applications can be lab-based, remote and IoT solutions, and OEM/embedded systems. Some customer applications/industry examples include, biomechanics, wind energy, power monitoring, machine condition monitoring, predictive maintenance, and more. One such customer is RO Scientific. It specializes in providing custom DAQ solutions and had a requirement for developing an automated permeameter machine. Accurate rock permeability measurements are critical for the oil and gas industry as the costly decision to drill in a particular spot heavily relies on such permeability measurements performed on prospective core samples. RO Scientific decided on the Raspberry Pi platform because they needed something as small as possible yet capable of providing enough computational power to perform some basic machine vision, do real time data acquisition and processing,
and transfer the outcome via Wi-Fi to a remote interface. The completed machine consists of a high-resolution camera, an X-Y table, controllers, a Raspberry Pi 4, and the MCC 118 and MCC 134 DAQ HATs. The HATs measure four pressure sensors, one high-precision temperature sensor, and two thermocouples for measuring slow changing temperature measurements, like ambient temperature. Using a hi-rez camera, the system scans the measurement table, determining the edges of the rock sample. A pneumatic actuator then lowers a probe to the sample surface and a gas is injected into the rock sample. The setup measures pressure at several points, logs temperature data, and determines gas viscosity. A fast-responding pressure sensor makes measurements at 10 kS/sec, monitoring the pressure decay. The sample permeability is calculated from this pressure decay curve, in conjunction with the other pressure and temperature measurements. This process is repeated in several locations within the sample. The outcome was a 90% drop in the overall measurement time and the measurements proved to be as accurate as when performed by a human expert,
and equally important, made the whole process repeatable. The Probe Permeameter machine makes measurements in core samples using an X-Y control system with under one millimeter precision. A Raspberry Pi 4 board is at the heart of it, providing enough processing power for the machine vision and data acquisition and control functions.
REFERENCES
Measurement Computing Corp., www.mccdaq.com DAQ HAT software library, https://github.com/mccdaq/daqhats
THE RO SCIENTIFIC AUTOMATED PERMEAMETER MACHINE.
eeworldonline.com | designworldonline.com
6 • 2021
DESIGN WORLD — EE NETWORK
47
TEST & MEASUREMENT HANDBOOK
ATTENUATION NETWORKS AND THEIR MEASUREMENT HERE’S A BASIC REVIEW OF PADS AND HOW TO CHARACTERIZE THEM.
AN ATTENUATOR IS A
four-terminal, two-port network that reduces
the amplitude of a signal without distorting its waveform. Attenuators are used in electronic instrumentation to precondition a signal for measurement and display. For example, the commonly-used 10:1 oscilloscope probe attenuates the signal voltage upstream from the analog channel input by a factor of ten, letting the user display signal amplitudes exceeding the instrument’s rating.
All attenuators have an associated insertion loss. In that many attenuators are designed to operate in the RF range, their insertion loss is often characterized by a complex impedance. For accuracy, this insertion loss is generally measured using a network analyzer. Thus the loss is represented by the S-parameter S21 measured via a two-port connection. In a typical measurement, the network analyzer makes a full set of S parameter measurements on the DUT for 0 dB attenuation, saving the value as the reference insertion loss. The S parameter measurements then are repeated for each attenuator setting. The values of S21 are adjusted for reference S21 values measured for 0 dB attenuation. The difference is the attenuation accuracy. Some commercial attenuators include an on-board EEPROM holding correction factors, in which case the corrected S21 values are usually written to EEPROM. Similarly, attenuators operating in the RF range also have a return loss, characterized by S11 or S22. Of course, return loss varies with frequency when the attenuator contains impedances. So for wide frequency ranges, the characterization of return loss can take time. One practice is to develop a frequency offset table consisting of the frequency and attenuation points to determine the attenuator performance across the frequency sweep of interest. Perhaps the most common use for attenuators is in power measurements. In this case the attenuator has to handle the maximum and minimum output power levels of the DUT. In particular, the attenuator must be able to handle max levels for signals having a high crest factor. The amount of attenuation is equal to the output/input ratio, which is always less than one. However, voltage attenuation is more conveniently measured according to the logarithmic decibel scale, given by the equation dBV = 20 log10 Vout/Vin. Of course, all dB measurements must assume a reference point. In rating a passive attenuator, the reference point is 0 dB. There are three principal passive attenuator configurations named after letters that their schematics resemble: L, T, and π. Each of these configurations can be either balanced or unbalanced. The balanced counterpart of the T-pad attenuator is known as an H-pad. Another
48
DESIGN WORLD — EE NETWORK
6 • 2021
DAVID HERRES, CONTRIBUTING EDITOR
variation is the bridge T-pad. The balanced π-pad is an O-pad. In balanced attenuators, additional impedances are connected across the transmission line. The ground is then located at a center point with respect to the balanced parallel impedances. The simplest attenuator is the L-pad, which is merely a four-terminal, two-port voltage divider comprised of two impedances. In an audio system, the L-pad attenuator is often used to reduce a signal that would otherwise overwhelm the speaker, and also to match the source and load impedances for maximum power transfer. If, however, source and load impedances are different, the attenuator can be designed to match one or the other, but not both. For this reason, the L-pad is considered an asymmetrical attenuator. To match two unequal impedances, symmetrical attenuators such as the T-pad or the π-pad are used. The T-pad passive attenuator consists of three resistors, two in series with the input and output, and a third resistor located in parallel with the input and output, from the center connection of the two series resistors to the ground return line. The T-pad attenuator has the same impedance looking from either end. Because its design is symmetrical, it is a more versatile impedance-matching device. Input and output may be transposed. It can be used to match either equal or unequal transmission lines. The two series resistive elements generally have the same value. But they can have different values if intended to join circuits that have different impedances. When used to match unequal impedances, the T-pad attenuator is known as a taper-pad attenuator. The balanced T-pad attenuator is comprised of two equivalent T-pads connected together, sharing in common the same two series resistors with ground center points that are in parallel with input and output. This combination forms a single four-terminal, two-port passive attenuator. The balanced T-pad attenuator is known as an H-pad attenuator because its
AN EXAMPLE OF A 30 DB PAD IN A Π CONFIGURATION. THIS DEVICE IS FROM MINICIRCUITS AND WORKS UP TO 18 GHZ WHILE EXHIBITING A MAXIMUM VSWR OF 1.11:1.
eeworldonline.com | designworldonline.com
BASICS OF PADS
THE BASIC ATTENUATOR PAD CONFIGURATIONS.
AN EXAMPLE OF A 100-W POWER ATTENUATOR. THIS DEVICE, FROM COAXIAL DYNAMICS, WORKS UP THROUGH 2.4 GHZ WITH A MAXIMUM VSWR OF 1.35:1. POWER MEASUREMENTS ARE PROBABLY THE MOST COMMON APPLICATION FOR ATTENUATORS.
schematic resembles that letter. The bridged T-pad attenuator uses an additional resistor in the series line. Notice that it bridges (i.e. is connected in parallel across) the two series-connected resistors. The bridged T-pad attenuator reduces the signal by altering the characteristic impedance of the circuit. That is because the signal flies across the T-pad attenuator. The two original T-pad resistive elements remain equal to the input and output impedances. You could say that the bridged-pad matches itself to the characteristic impedance of the two transmission lines. Additionally, it is feasible to substitute a potentiometer or a resistive switch for two of the resistive elements, thereby creating a variable attenuator network. Rather than using a potentiometer, a stepped bridged-T attenuator can be created by using switched, fixed value resistors. The π-pad passive attenuator schematic generally consists of one series impedance element. On the input and output sides are two impedance elements connected in parallel to the ground return line. The π-pad attenuators are used in RF and microwave transmission lines. At these frequencies it is essential to avoid the use of any inductive elements such as wire-wound resistors. The π-pad attenuators are available in unbalanced and balanced configurations. They are symmetrical and can be used between equal or unequal impedances as matching devices. When configured between circuits of equal impedance, the impedance elements have the same value, but to match circuits of unequal impedance, different values are chosen. The balanced π-pad attenuator requires an additional impedance in the common ground line, and for that reason it is called an O-pad attenuator. In the O (balanced) configuration, the value of the upper series impedance is divided by two, as is the lower series impedance. Now that there are two half-valued series impedances, however, the overall series eeworldonline.com | designworldonline.com
impedance remains the same as in the π-pad configuration. Attenuators are balanced or unbalanced depending upon the type of transmission line. For example, attenuators used with coaxial lines are unbalanced. Attenuators used with unshielded twisted pairs as in Ethernet circuits are balanced. Because most attenuator circuits consist only of passive elements, they are linear as well as reciprocal. If the circuit is symmetrical, then input and output ports may be reversed. RF attenuators generally take the form of precision coaxial connectors. Waveguides are required for frequencies in excess of 30 GHz, with wavelengths less than one centimeter. An optical attenuator, of which a fiber optic attenuator is a special case, reduces the power level of an optical signal. You can make a continuously variable, free-space (non-fiber) optical attenuator by removing the lenses from a pair of polarized sunglasses. When held together so the polarizing lines are parallel, all incident light will pass through and attenuation is minimum. When held so that the polarizing lines are perpendicular, no light passes and attenuation is maximum. When rotated, attenuation varies between these extremes. In fiber optic communications, an optical attenuator can be installed inline to match signal levels, preventing harmful reflections and data loss. Optical attenuators can also be inserted temporarily to test optical power levels. Various expedients involve introducing a small gap by loosening a connector, although this can cause reflections. Another quick solution is to wrap several turns of optical fiber around a pencil so a portion of the light is not conveyed by means of total internal reflection.
REFERENCES
A SIMPLE WAY OF KEEPING FOUR THE S PARAMETERS STRAIGHT.
Basics of network analyzers, https://www. testandmeasurementtips.com/do-you-really-need-avna-when-a-scalar-network-analyzer-might-do-faq/ 6 • 2021
DESIGN WORLD — EE NETWORK
49
AD INDEX TEST
&
MEASUREMENT HANDBOOK • JUNE 2021
Digi-Key..................................................... Cover, IFC
LEMO USA, Inc......................................................... 9
Electrocube, Inc...................................................... 29
Newark, An Avnet Company.................................. 39
Ironwood Electronics.............................................. 41
Pico Technology North America............................ BC
Keystone Electronics Corp............................... 1, IBC
Teledyne LeCroy..................................................... 19
SALES
LEADERSHIP TEAM Neel Gleason ngleason@wtwhmedia.com 312.882.9867 @wtwh_ngleason
Jami Brownlee jbrownlee@wtwhmedia.com 224.760.1055 Mike Caruso mcaruso@wtwhmedia.com
Courtney Nagle cseel@wtwhmedia.com 440.523.1685
469.855.7344 Bill Crowley bcrowley@wtwhmedia.com
Jim Powers jpowers@wtwhmedia.com 312.925.7793 @jpowers_media
610.420.2433 Jim Dempsey jdempsey@wtwhmedia.com 216.387.1916 Mike Francesconi mfrancesconi@wtwhmedia.com 630.488.9029
Publisher Mike Emich memich@wtwhmedia.com 508.446.1823 @wtwh_memich Managing Director Scott McCafferty smccafferty@wtwhmedia.com 310.279.3844 @SMMcCafferty EVP Marshall Matheson mmatheson@wtwhmedia.com 805.895.3609
WO
RLD
@mmatheson
@D 50
DESIGN WORLD — EE NETWORK
6 • 2021
G ESI
N
eeworldonline.com | designworldonline.com
DW THiNK Custom_6-21p.indd 1
5/3/21 11:33 AM
PicoScope 6000E Series NEW! 1 GHz and 750 MHz models ®
AND Introducing our first Pico active high-performance probes with Intelligent Probe Interface
A smarter scope for faster debug Visit www.picotech.com/A852 to find out more Email: sales@picotech.com. Errors and omissions excepted. Please contact Pico Technology for the latest prices before ordering.