Over the years since I started this column, I have had requests for a listing of basic guidelines and limits applicable to NEC. I have had occasion to create some presentation graphics covering some of this information, and I shall present them in this column.
The .GIF graphics can be extracted from the text and placed in a Word or Word Perfect document, one per page at full paper width. Then the result may be printed or saved. Only the ones useful to you should be extracted. As well, you may run them through a graphic program. such as Paint Shop Pro or equivalent, for printing--and even revising to suit your specific needs. Alternatively, you may take notes on aspects of NEC limitations and guidelines as they apply to your projected modeling work.
The set of graphics is neither comprehensive nor complete in detail. Most of the sheets are taken from the NEC-2 manual, but apply also generally to NEC-4. The key exception is the fact that NEC-4 permits wires underground, with rules for the penetration of a wire into the ground. Fig. 7 below covers the NEC-4 situation in an example of a monopole with buried radials.
Since the graphics themselves contain the text, little commentary is required. Actually, each graphic requires a full column of commentary, but that has mostly appeared in past columns.
For fuller information, there are several useful sources. Of course, the NEC-2 and NEC-4 user manuals are the primary sources. For modeling with NEC-2, Basic Antenna Modeling: A Hands-On Tutorial As well, ARRL offers an on-line course in modeling with NEC-2. Both of these latter sources offer exercise model files.
Fig. 1: Some Absolute NEC Limits for Wires and Segmentation: Although few MF and HF models will approach these limitations, they become especially important in the modeling of antennas for VHF and upward, where the anticipated element diameter may become a very appreciable fraction of a wavelength. See column #3 for more information on NEC limitations.
Fig. 2: Some Conservative Wire and Segmentation Recommendations for Newer Modelers: The experienced modeler may safely by-pass these recommendations, although most models should remain within these guidelines in order to achieve an AGT (Average Gain Test) value that is close to ideal. See Fig. 11 for further information on the AGT.
Fig. 3: Some Modeling Practices to Embrace: Not all good practices appear in this brief set of guidelines, but the listed suggestions may help you develop your own extended list of good practices. Good practices do not ensure a good model, but they do help to eliminate oversights that seem to defy detection once the full model is complete.
Fig. 4: Some Modeling Practices to Avoid: As with the list of good practices, this list of practices to avoid is incomplete, but a potential foundation for a user-specific list of things to avoid. Some practices, such as using a stepped-diameter element, have correction features in some implementations of NEC-2. NEC-4 has overcome much of the inaccuracy involved in stepped diameter elements, but AGT values for large diameter changes may still be disappointing. The remaining practices to avoid apply equally to NEC-2 and NEC-4, with the exception of the ground warning, which applies to NEC-2 only.
Fig. 5: Limits for Wires Near the Ground: The Sommerfeld-Norton ground calculation system is very accurate for wire very near the ground, and the limits for wire proximity to ground are small. However, a ground radial system very near the ground in NEC-2 provides only a crude indication of the buried ground radial system for which it may substitute and which is available in NEC-4. See column #11 for more on ground radial systems. The increased speed of current-generation PCs has largely made the use of the reflection coefficient method of calculating ground effects irrelevant, since the S-N ground calculations do not significantly slow the calculation process down.
Fig. 6: Selecting the Right Ground for the Right Job: In NEC-2, a wire (vertical monopole) touching the ground may not yield correct results with either the S-N or the reflection coefficient ground. Hence, resorting to a perfect ground for comparative results between models may be necessary. However, for more accurate results that take into account the properties of the ground, NEC-4 is preferred, since one may directly model a buried ground radial system.
Fig. 7: NEC-4 Ground Penetration Rules, Using a Vertical Monopole and Radials as a Sample Case: The sample in this figure combines the rules for ground penetration in NEC-4 with an example of element length tapering to avoid adjacent segments that differ too much in length. The method shown uses manual tapering with separate wires for each segment length. However, it is possible to use the GC input to automate the process within sngle wires for each element. Whichever system is used, the source wire and the wire between Z=0 and the junction of the radials should be individual wires to allow for maximum control over the model and to avoid junction errors as one modifies the model. Although the wire penetrating ground may have a segment junction at the required Z=0 point, users achieve maximum model control by making Z=0 a wire junction.
Fig. 8: Transmission Line (TL) Limitations: As non-radiating elements of the model, transmission lines are subject to many limitations. Where a model requires transmission lines outside regions of high current and low rates of current change, many lines may be modeled using physical (GW) wires. See column #21 and #22 for more on transmission lines.
Fig. 9: Connections of Loads, Sources, and Transmission Lines on the Same Segment: Applying mutliple loads, sources, and/or transmission lines on a single segment is often a source of confusion. (In addition, some implementations of NEC may permit only a single load and/or a single source on a chosen segment.) See columns #4 and #5 for more on sources. See columns #6, #13-#17, and #46 for more on loads. Since a load is in series with a transmission line, placing a load on the wire used to terminate a transmission line will not place the load in parallel with a source at the near end of the transmission line. One must use the admittance facilties of the TL entry, although these values are not frequency nimble. To obtain a load in parallel with a source that will change reactance as the frequency changes requires other types of work-arounds.
Fig. 10: The Convergence Test: For a fuller account of the convergence test, see column #1. The convergence test, like the AGT, is a necessary but not a sufficient condition of model adequacy.
Fig. 11: The Average Gain Test: For a fuller account of the average gain test, see column #20. Since both the convergence test and the average gain test are necessary but not sufficient conditions of model adequacy, they together yield at best a good indication of model adequacy, but not a decisive judgment. The use of experimental results as well as a full evaluation of the model in terms of all program limits remain recommended additional checks on models.
Converting the AGT number into a value in decibels is simply a matter of 10 times the common log of the AGT value. Use only the AGT value obtained in free-space (or over perfect ground) for lossless wires and no resistive components to any loads in the model. If the source impedance is very close to have no reactance, then the basic AGT value times the reported source resistance value will provide a more correct source resistance value. The positive AGT value in dB may be subtracted from the reported gain value and a negative AGT value in dB may be added to the reported gain value to yield a more nearly correct gain figure for many models.
There are additional rules and provisions within NEC. There are, as well, numerous very specific situations that might create a problem if not modeled carefully. Nevertheless, I hope this collection of graphics--whether in whole or in part--provides a few handy reminders that help you avoid potential pitfalls.
Go to Main Index