The Quasi-Random Wire Antenna Revisited

A confluence of several factors have put me back onto these things, following up on an earlier success.  First, their simplicity.  Second, I want an easy NVIS backpacking wire that will cover 160–30m.  Third, I was curious about the analogous results for common shortwave broadcast bands, and developed a program to do the calculation there, which naturally led back to doing the same calculations for the ham bands... you get the picture.  Finally, LDG has extended their balun/unun line to a handy 9:1 unun and I had to get and try one.  A non-resonant random wire in free space has an average impedance around 600 ohms, so a 9:1 transformer should drop it into the 67 ohm range, which is only a small mismatch into 50 ohm coax.

First, the calculations.  Similarly to one from the guy at U Del., my code calculates the "bad" half-wavelength integral lengths corresponding to each of the the band edges, and then eliminates all values in between.  My code uses a knock out array to keep track of the allowable values, and after all selected bands are considered and the corresponding bad areas knocked out, the few remaining values are output.  To convert frequencies to feet, the usual 468 ft*MHz value is used, which includes the usual 0.95 velocity factor.  I added an option for an additional 0.965 velocity factor for insulated wire.  The code rounds inward to the nearest even foot, to eliminate marginal values.  In this calculation, the maximum wire length considered is 140'.  Longer lengths are possible, but this is enough for a 160m antenna.  Finally, where the U Del program limits things to 4 harmonics, the new program calculates how many harmonics can fit a total wire length and goes from there.  Sometime, this went as high as 16.  User input consists of the range of bands to be considered, whether to use the WARC bands, and whether the wire is bare or insulated.

Without showing any nice graphics here (tired; too much work; not needed), the results of the two programs agreed well as long as the insulation velocity factor wasn't included and the harmonics were limited to 4.  Turn these improvements back on, and the results varied a little.  But to cut to the chase, for bands 160m-30m, 136'–140' of insulated wire were the shortest practical lengths.  Drop the 160m band and limit ourselves to 80m-30m, and the shortest practical lengths were 69'–85'.

To test these numerical results, two 20 ga. wires at the average value of the above mentioned bands were cut; i.e., 138' and 75'.  For each test, the wire was hung at approximately 6' above ground from tree branches using loose zip-ties.  The wire did not contact the branches.  This was a reasonable representation of a temporary field set-up.  The feed end of the wire was attached to the hot post of the 9:1 unun.  Optionally, a ~30' ground/counterpoise wire could be attached to the ground post on the unun.  Both configurations were tried in testing, across all bands.  A 25' length of LMR-240 50 ohm coax was used to connect to an MFJ HF-VHF SWR meter.  The meter was ungrounded, however the coax shield was entirely along the ground and likely provided significant inductive coupling to the moist grassy earth.

Across all covered bands (i.e., 160, 80, 60, 40, & 30m, with 160m omitted for the shorter wire), both wires exhibited useable antenna properties.  SWRs were typically in the 2–4 range, and never exceeded 6.  Values for R (radiation resistance) ranged from 22 to 110 ohms for the ungrounded configuration, with X (reactance) values from 20 to 80 ohms.  Generally, the R values were ~50% greater than the X values.  For the grounded configuration similar results were seen; however, the X values were usually ~50% greater than the R values, indicating a lower radiation efficiency.  While sweeping between bands, peaks of several hundred ohms were seen, indicating that the "bad" half-wavelength integral values had been successfully avoided for the amateur bands and buried harmlessly in the inter-band gaps.

Bottom Line: For 160–30m, a 138' quasi-random wire strung horizontally at 6' worked well in SWR meter testing, and similarly for 80–30m, a 75' quasi-random wire worked well too.  The "bad" wire lengths were avoided for the amateur bands, but could be seen as peaks in SWR while sweeping between bands.  For all cases, the ungrounded unun configuration worked better than with the ~30' counterpoise/ground attached.

Testing will continue, beginning with transmission tests and A/B comparisons to an existing NVIS dipole.  Additionally, shortwave broadcast band values and reception testing will be performed and presented in the near future.

Oh man, that post was dry as dust.  Sometimes though, "Just the facts ma'am" is the only approach to take.