RF field level measurements will always have a significant amount of measurement uncertainty even when the measurements are made by a skilled surveyor using the best available instruments. Measurement uncertainty has three major components.
Measurement uncertainty due to the instrumentation
Perturbation of the field by the surveyor
Time and spatial variations in the field
Virtually all RF safety measurements are made with broadband instruments comprised of a probe and a meter. The accuracy of a survey instrument is almost entirely driven by the accuracy of the probe. Most probe specifications are expressed in dB. A parameter that has a 1.0 dB tolerance means the value could be off by 26 percent. In contrast, even a simple meter should be accurate within a maximum of 5 percent.
Frequency deviation is often the parameter that contributes the most to measurement uncertainty, but it is not the only parameter that should be considered. The FCC Regulations and all of the other major worldwide standards have exposure limits that vary as a function of frequency. The growth of wireless services and deployment of digital television have both led to a growing number of sites that have multiple emitters operating at frequencies with different Maximum Permissible Exposure (MPE) limits. This has led to the use of shaped-frequency response probes as the primary tool used for surveys of wireless and broadcast sites.
Shaped-frequency response probes are designed so that sensitivity at the point of detection varies over their frequency range. The goal is to match a standard, such as the FCC Regulations, as closely as possible. Narda Microwave holds the patent on this technology, which is accomplished in a manner similar to designing a filter. It is impossible to make the sensitivity match the MPE limits exactly. The greatest errors tend to occur at the transition points, where the MPE limits change from a constant to a slope or vice versa. In the FCC Regulations for Occupational/Controlled exposure, these transition points occur at 3 MHz, 30 MHz, 300 MHz, and 1,500 MHz.
It is almost impossible to define the overall accuracy of a survey instrument because accuracy depends not only on the instrument but also on what is being measured and the techniques that are used to make the measurements. The best that one can do is to estimate the overall measurement uncertainty for a particular set of conditions and determine that the uncertainty is valid providing that the equipment is used properly. There are several parameters that should be considered when attempting to establish the level of measurement uncertainty. The parameters are listed in order of importance; i.e., the parameters that can result in the greatest uncertainty are listed first. The list assumes that the probe is well designed and appropriate for the measurement task.
RMS Detection. It is not important to have a probe work in the “square law” region if you are measuring within a single frequency. However, measurements at a multiple-emitter site should be made with a probe’s detectors functioning within the square law region. The potential error caused by your probe working outside of the square law region is as much as 10 dB. In other words, the instrument might overstate the actual field strength by a factor as high as 10 to 1.
Frequency Sensitivity. Be sure to look for guaranteed maximum frequency deviation—many monitors and probes use “typical” values. Some shaped-frequency response probes and monitors use an artificial reference point rather than the standard or regulation. Many of these products can have errors as high as 10 dB.
Calibration Frequencies. Single-frequency calibration makes the assumption that all similar probes function within specified tolerance limits. This is a big assumption.
Ellipse Ratio. This parameter defines the measurement variation that occurs when you rotate the probe about its axis and point it in different directions. Good probes achieve guaranteed deviations that are no greater than ±0.75 dB.
Isotropic Response. This parameter, which includes ellipse ratio, defines the measurement variation that occurs when signals come from different directions relative to the probe. The guaranteed values should not exceed ±1.5 dB.
The guaranteed maximum frequency deviation of the two shaped-frequency response probes used by RF Safety Solutions—the Narda 8700 series—is ±2 dB. These probes are calibrated at 14 different frequencies from 300 kHz to 3 GHz. The other parameters, such as ellipse ratio and isotropic response, are less significant than frequency deviation but cannot be ignored. A good rule of thumb when making measurements in multi-signal environments with this type of equipment is to assume an uncertainty of ±3 dB. Other shaped-frequency response probes, such as the Type 25 FCC-shaped probe used with the EMR series of meters, are calibrated at a single frequency and donot have a guaranteed maximum frequency response error. The frequency deviation of these probes can be far greater than ±2 dB.
If measurements are made where there is only a single emitter or where all emitter frequencies are very close to each other, as is the case at a site with only cellular service, a correction factor can be used to reduce the amount of measurement uncertainty. In more complex environments, the additional calibration frequencies only provide a rough indication of frequency deviation. This can reduce overall measurement uncertainty to about ±1 dB. It is important to understand the characteristics of the probe to use the correction factors correctly. A shaped-frequency response probe has its greatest errors in the transition regions near 3 MHz, 30 MHz, 300 MHz, and 1,500 MHz where the probe’s sensitivity changes sharply over frequency. The use of correction factors is less accurate when one attempts to interpolate between two calibration frequencies near the transition regions of the probe.