Select the Right RF Power Sensor for Your Application
By Richard R. Hawkins, President, LadyBug Technologies

Most power sensors fall into one of three categories: thermistor, thermocouple, or diode detector. Thermistor detectors measure power using DC substitution. RF energy, absorbed by a thermistor in one leg of a bridge, results in a resistance change and bridge imbalance. DC is applied, rebalancing the bridge and resulting in a voltage proportional to RF power. In thermocouple detectors, absorbed RF energy heats the thermocouple junction, resulting in a voltage proportional to power. While both of these designs are useful, they suffer from limitations--primarily sensitivity (usually limited to -30 dBm) and slow response (>100msec).

Diodes mitigate the limitations inherent in thermistors and thermocouples. Diode-based detectors produce a voltage proportional to input power. They offer quicker response time and greater sensitivity, allowing them to measure very low power levels. Coupling greater sensitivity with multipath designs enables extremely wide-dynamic-range products. Together, these advantages make diodes the preferred design.

Traditionally, RF power sensors were used in conjunction with a separate power meter. They are slower, more costly, occupy greater rack space and require a complex “zeroing” and calibration procedure before use. Further, they were often limited to measuring only true average RF power. The introduction of the USB RF power sensor (see Figure 1) brings several advantages to making RF power measurements. These sensors are fast, low cost, compact, sophisticated, and easy to use: simply connect to a computer and go. And some USB sensors have a full set of measurement capability, from simple CW to modulation analysis to full-featured pulse profiling and time-domain gated analysis.

Applications for Power Sensors
RF power measurements are fundamental to the life cycle of any RF or microwave product and are required in almost all environments. There are general requirements during early investigation, development, verification, manufacture, field deployment, repair, recalibration and field service. RF measurements are also required in bench top, portable, remote sensing as well as embedded system environments. Specific application areas include: aerospace, defense, terrestrial, satellite, wireless communications, point-to-point communications, wired communications, radar, GPS, antenna test, automated test, component test, instrument test, etc.

RF power sensors historically have been used for terminating power measurements and for in-line power measurements for signal-level monitoring. However, with USB sensors now measuring at 2000 points per second, it is now practical to make real-time multiport scalar network analyzer reflection and transmission measurements.

USB Sensors vs. Classic Sensors and Meters Cost and Flexibility
Classic power sensor-meter combinations run between $5K and $8K for the meter alone and an additional $1.5K to $5K for the sensor. A USB sensor ranges from only about $1.25K to $5K. A computer is needed for display and recording of the power measurements, but often one is already available--rendering USB sensors very economical. And a single computer can be used for multiple sensors.

USB sensors provide flexibility due to their compact size and light weight, enabling them to be used in applications in which classic sensors and meters are completely unsuited. With no need for tethering to a meter, the door is open to uses involving multiple sensors embedded not only in test systems, but also in field-deployed operational systems.

Rack Space
Classic power meters consume rack space, and one meter can support no more than two sensors. USB sensors require only a single computer--and multipoint power as well as multiport scalar measurements can easily be made through the use of a USB hub, making port replication simple and inexpensive.

Remote Operation
Classic sensors require a custom cable to connect the sensor to the meter. The custom nature and expense of these cables limits the distance between the measuring sensor and the display meter. This cable is typically about 3 feet, with very expensive extensions available. For USB sensors, the length from the point of measurement to the USB port on a computer or hub can be up to 5 meters--and, with a USB extender, can be up to thousands of meters. These lengths allow for easy operation at antenna sites, embedding in systems and other remote-sensing applications. Connections are made using standard USB cables available for under $20.

No-Zero, No-Cal
Historically, meter-sensor applications required a zero and calibration before use. Even most USB power sensors require some zeroing. LadyBug offers the only wide-dynamic-range sensor on the market offering “no-zero, no-cal”; these units require no zeroing or calibration whatsoever before use. This offers much greater convenience, but the larger benefit is in improved stability and accuracy (see Figure 2). LadyBug sensors undergo the most rigorous factory calibration in the industry, including a NIST-traceable calibration across the full operational temperature range. The sensor measures the temperature for each power reading, employing a patented technique to ensure that the correct factory temperature calibration factor is applied to each measurement.

USB Sensors: Reduced Overall Error
Several errors make up the accuracy or measurement uncertainty for a power measurement. Classic power-meter power-sensor applications exhibit greater error than do USB sensors. Below is a list of the most significant errors; each of these is transferred to the uncertainty of the final power measurement. Total uncertainty is typically achieved by combining the errors.

Calibrator – typical error 2%
The calibrator is used to transfer a reference to a power sensor. Uncertainty of the calibrator is transferred directly to the final measurement, regardless of frequency or level. Because USB sensors do not require a calibrator, this error is not present.

Sensor calibrator mismatch – typical error 2%
In the calibration process, RF power is removed and the power sensor is connected to the power meter calibrator. Because neither the calibrator nor the power sensor match is perfect, there is a significant mismatch error transferred to the final measurement during calibration. As USB sensors do not require a calibrator, this error is not present.

Meter instrumentation error – typical error 0.5%
A power meter is calibrated separate from the sensor, resulting in a set of measurement errors independent of the power sensor. A USB sensor does not suffer from this error since a separate meter is not present; the error is removed during the power-sensor factory calibration.

Sensor to DUT mismatch and repeatability – typical error 2 to 6%
Mismatch between the DUT and the power sensor is typically the largest of all errors in a power measurement. It occurs because of the imperfect match of the DUT and power sensor. This error is present for all sensors.

Calibration factors – typical error 1.5 to 3%
Calibration of the power sensor results in a calibration factor and an associated error. The calibration factor encompasses all errors associated with level calibration of the power sensor during manufacture or recalibration.

Linearity – typically as high as 7% for older sensors and 3% for newer sensors
Linearity error is the additional error that accumulates over power from a reference point.

Noise – 0.2% to 10% or more at -60 dBm
Noise appears as measurement jitter and it is most significant at lower power levels: about 20 dB above the noise floor. LadyBug sensors have impressively low noise: typically better than 0.2% at -60 dBm. Both traditional meter-sensor solutions and other USB solutions have well over 10% under the same conditions.

Temperature
Changes in temperature can result in large errors, particularly for very wide temperature variations (see Figure 2). Sensors featuring no-zero and no-cal have temperature sensitivity much lower than those classic sensors requiring a zero and cal, and those USB sensors requiring only a zero.

Dynamic Range vs. Instantaneous Dynamic Range
Dynamic range is the range of power over which a sensor is capable of making useful power measurements. Generally, this means true average power measurements. Both classic power meter-sensors and USB power sensors with a usable dynamic range of 80 dB (typically +20 dBm to -60 dBm) or more are readily available. Further, attenuators, couplers, etc. are often used to shift the dynamic range of a power sensor.

Instantaneous dynamic range is key, particularly when measuring signal on/off ratios, signals with high levels of modulation, or large transitions. For both classic power meters and power sensors and USB power sensors, the instantaneous dynamic range rarely exceeds 50 dB and can be as low as 30 dB. This limitation occurs because a single detector’s dynamic range is limited to 40 or 50 dB. And, since most sensors measure using one range at a time, the user is limited to the dynamic range of that single path. However, there are USB sensors whose instantaneous dynamic range exceeds 80 dB--accomplished by measuring all paths simultaneously.

Frequency Range
Power sensors are available in a variety of frequency ranges that cover bands from 9 kHz to 110 GHz. The most common bands are from 10 MHz to 26.5 GHz. Variations in the frequency response of the sensor are accounted for in the calibration table stored within the sensor.

Zero-and-Cal
The zero-and-cal process for traditional power sensors involves multiple disconnections from the measurement point and connections to an external calibration source. By themselves, zero-and-cal do not provide an accurate power measurement even when executed after each change in temperature. Newer sensors that have an internal zero-and-cal capability don’t require an external calibration source but also exhibit sensitivity to temperature. In this case, the internal zero-and-cal is automatically performed periodically but not for each measurement. The sensitivity of a USB sensor with internal zero is shown in Figure 4. Notice with non-temperature-compensated USB power sensors measuring at low power levels, measurement error climbs higher than 1.5 dB and represents a measurement error of over 40%.

Measurement Capabilities and Features Measurements
Classic power meters typically measure true average power only. Today, many USB sensors offer this same capability, along with a host of others often found in solutions costing well over $10K. Measurements include true average CW, modulation analysis (average, peak, pulse, duty cycle, crest factor) and full featured time domain gated power measurements. Data logging, averaging, relative power, offsets, frequency response correction and other functionality is included.

Measurement Speed vs. Settled Measurement Speed
Sensors typically will have their sample rate specified. However, a more important parameter is settled measurements per second. The number of readings or measurements per second varies widely between different sensors. But, making unsettled measurements at a high rate is rather pointless. Settled measurements are very dependent on the noise present and the integration time required to get a settled measurement. LadyBug’s USB power sensors make a settled measurement at -60dBm in about one millisecond. The typical settling time for a classic sensor at this power level is generally greater than one second and can be up to four seconds or more.

Triggering
Triggering is essential for synchronization of power measurements to system or other events, such as with a signal generator, network analyzer, signal pulse, etc. In these applications full-featured triggering can be required. Full-featured triggering includes trigger in, trigger out and triggering on incoming RF signals (internal trigger). Most classic and USB power sensors will receive a trigger. However, few will issue a trigger or trigger on the incoming RF. Modern USB power sensors offer full-featured triggering. The LB480A pulse profiling sensor will reliably trigger on signals below -55dBm. (Trigger ports are shown in Figure 5.)

Software for USB sensors
An essential requirement for any USB device is software. The same is true for USB power sensors. Users must be able to get up and running quickly with easy-to-use power meter panels (Figure 6) and programming users must be able to integrate the power sensor into a variety of application environments, such as VB.NET, C++, National Instruments LabVIEW, Agilent VEE, and others. The power meter panels and the programming environment drivers must be able to make all required measurements for the application. It is generally recommended that the user ensure supplied software supports multiple sensors.

Choose the Right Sensor
When selecting a power sensor, much depends on the application. But, beyond the needed dynamic range and frequency coverage there are many aspects to consider: required measurements, the programming software environment, ease of use, accuracy and uncertainty that can affect measurements. Will there be wide temperature variations? Will the sensor be connected and re-connected often? Will many different devices be measured such that mismatch may be an issue?

Where portability is important, or the sensor is to be embedded in a system, or remote measurements are needed, a USB sensor may be the optimal choice. For more information, please visit our website.

LadyBug Technologies
www.ladybug-tech.com
TXTLINX.COM86
Email this article to a friend!