MININEC Notes

The original MININEC code on which ANTENNA MODEL is based is known to have several problems. These problems are described here, along with our fixes for them.

	The results obtained exhibit a frequency bias, which gets worse as the wire diameter increases. The results obtained are not correct for the frequency specified, but are correct for a slightly lower frequency. ANTENNA MODEL handles this with an algorithm modification.
	There are problems when wires meet at an angle of less than 180 degrees—sometimes called the "bent wire" effect. This is very noticeable for quad and delta loops. The problem diminishes as the angle increases from 90 to 180 degrees. The problem has been described in a number of ways, but we believe the best explanation is that the model simply converges very slowly. ANTENNA MODEL handles this problem with an algorithm modification and increased segmentation density. ANTENNA MODEL recognizes these situations and applies the algorithm modification, and when automatic segmentation is employed, increases the segmentation density appropriately.
	There are problems when a short segment is adjacent to a much longer segment. This can happen, for instance, at the end of a folded dipole. This has led to the practice of "segment length tapering" and the rule-of-thumb that adjacent segments should not differ in length by a ratio of more than two to one. This problem is related to the next problem, which is that the code completely breaks down and gives ridiculous answers when two wires meet at an acute angle of less than 28 degrees (a skinny rhombic), or when two parallel wires are spaced closer than 0.23 times the segment length on the wires. Both of these problems are caused by using the wrong numerical method to perform some function integrations. ANTENNA MODEL handles these problems by using a slower, but accurate, numerical method to perform the integrations.
	The MININEC model for real ground isn't as sophisticated and accurate as NEC's Reflection Coefficient and Sommerfeld-Norton ground models. We overcome this limitation by simply including these NEC ground models in addition to the usual MININEC ground model. (ANTENNA MODEL's implementation of Sommerfeld-Norton ground is accurate for wire segments down to about 0.02 wavelengths above ground.)

If the MININEC code has these problems, then why do we use it? Because the MININEC code has some advantages, as well.

There is no problem placing a source or load at the intersection of two wires (a common NEC problem). "Split sources" and "split loads" are not required.

The MININEC code handles wires with stepped radii (read: Yagi elements made of telescoping tubing) very well. There is no need for "equivalent monotaper elements" as required with NEC-based programs.

ANTENNA MODEL implements current sources without resorting to networks and hidden "remote" wires.

NEC-based programs are susceptible to reduced accuracy when models employ angular junctions of unequal-diameter wires. ANTENNA MODEL isn't.

NEC-based programs are susceptible to reduced accuracy when models employ closely spaced parallel unequal-diameter wires. ANTENNA MODEL isn't.