Technical Note:

Measuring E-fields 25-1000 MHz with the NIST dipole-detector.

by Chris Scott

Much of the material was adapted from NIST technical note numbers 1309 and 1098. Thanks to Galen Koepke, Electronics Engineer, Fields & Interference Metrology Group, NIST, for helping me reproduce the standard calibration procedures. 

There is often a requirement in RF and antenna work for gain standard antennas. What is the incident field strength in volts-per-meter present within the intrinsic 376.73 ohm impedance of free space within a subject volume of space? This information is often necessary both for comparative gain measurements and for absolute measurements. Desirable gain-standard antenna qualities include stable and accurately known gain, a high degree of dimensional stability, and clearly defined polarization. Ease of reproduction has additional merit.  Before we delve too deeply into this experimental apparatus, note that it is of limited use making broadcast field strength measurements, since it is a broadband sensor and cannot distinguish between different frequencies.  If general purpose field strength measurement is desired, see our adjustable calibrated dipole product.

Reflections notwithstanding, a thin half-wave dipole has a gain of 2.14 dB over a theoretical spherically radiating (isotropic) antenna. This represents the pattern of the antenna. The half-wave dipole is well accepted as a natural, reliably reproducible reference standard. Can it be used to measure incidental fields? Sure, assuming that it's matched to a balanced, well characterized transmission line run several wavelengths perpendicular to the dipole element to prevent induced currents. Only then, can a calibrated receiver or detector accurately measure the antenna terminal voltage or power, after being adjusted for mismatch and line loss. This creates a fragile test environment and introduces other sources of error.

Worst case error and total uncertainty may be calculated simply as the addition of the percentages, or decibels. Since the sources of error are for the most part unrelated and independent, with each error unlikely to be at its extreme value in the same direction, this approach is very conservative. A more realistic method is the root-sum-square calculation. Finding the rss uncertainty follows its name; square the components, sum those squares, take the square root. In any case the goal is to reduce or eliminate sources of error.

The antenna and detector described in this note represent the standard hardware used for measuring the calibrated E-field at the NIST open-field antenna test range. The worst case total uncertainty is one decibel, or using the rss method, .6 dB, according to NIST scientists.

The NIST standard receive antenna is a resonant half-wave dipole with a high impedance detector-voltmeter installed across the gap of the center insulating support, which is usually made of Teflon. The detector diode is a high-burnout Schottky microwave type. Hewlett-Packard is one manufacturer of these units. The Schottky diode has a low turn-on voltage and a high (70v) peak reverse voltage rating. This type of diode provides a very high shunt impedance (greater than one-hundred Megohms) for the signal levels used.

An innovative feature of the dipole-detector is the use of high-resistance interconnection wires. At about one-thousand ohms per inch, these carbon-impregnated plastic wires don't interfere with the incident field. The termination must ultimately be a high impedance voltmeter; most modern digital units qualify, but should be carefully calibrated.

This arrangement reduces two sources of error considerably. The disturbance of the incident field is much less due to the use of semi-conducting wires, and the impedance matching problems are eliminated, since there is virtually no load.

The voltmeter is calibrated using a modified coaxial tee connector energized by the internal fifty Megahertz calibration source of an rf power meter. This circuit is shown in the calibration schematic. The response of the detector is essentially frequency independent, the low frequency cutoff region mainly being limited by the value of the bypass capacitors used. Much lower calibration frequencies can also be used. The resistor-capacitor filter network produces a dc output voltage very close to the peak value of the radio-frequency input voltage. A typical NIST calibration chart is shown.

A center insulator with banana connectors installed in both ends is one way to allow for frequency change. Cut-to-length rods may then be used for various frequencies. Alternatively, telescoping whips may be installed, permitting quick adjustment over a wide range. The formula for calculating the proper length of the elements according to Schelkunoff's algorithm is shown. A MathCAD spreadsheet is also available for these calculations. In the case of the telescoping whip elements, the equivalent diameter can be approximated by integrating diameter along their lengths. This is not extremely critical to accuracy, since a small reactive component in the source impedance remains negligible compared to the load resistance.

Note that in the presence of strong interfering signals, the dipole-detector system cannot be used for accurate measurements, due to the broadband detection. A noteworthy variation of this antenna is where the dipole length is adjusted to less than about one-ninth wavelength; instead of the current distribution being sinusoidal, it becomes triangular, and remains that way, extending the flat frequency response down to much lower frequencies.

Many broadband "isotropic" power density meters sum the output of three orthogonal dipoles like this to sense the E-field component. The effective length must of course, be calculated differently, due to the change in current distribution.

So exactly how many volts per meter is in that ether? Build the standard dipole-detector and know. 

 view cs&a HOMEPAGE