OwenDuffy.net 

A very popular multiband wire antenna with hams is the 133' dipole, centre fed with an arbitrary length of open wire line and matched with an balun/ATU combination. This antenna is regularly promoted in ARRL publications, eg (Ford 2004).
But is it the nobrainer that is projected in those publications?
A correspondent wrote that he had escaped the big city, and he was pleased that he could graduate from a G5RV now that he had a large yard available.
The 'improvement' was to be a full size 80m half wave (133') centre fed dipole, fed with home brew open wire line with Zo around 600Ω. The location of the dipole was governed by convenient trees, and the transmission line was dictated pretty much by the distance to the building entry (14m). Obviously, the transmission line cannot be shorter than the minimum distance to reach from feed point to building, and it isn't practical to accommodate excess length.
So, he flashed it up, literally! Arcs and sparks in the balun/ATU at less than 500W PEP on the low end of 80m.
Why?
(Ford 2004) doesn't mention the risk of flashover. In fact the only mention
of difficulties is [y]ou may discover that you cannot achieve an acceptable
SWR on some bands, no matter how much you adjust the tuner
. It will be seen
that this is a symptom of the same underlying problem, extreme impedances that
might be presented at some feed line Zo, lengths, and frequencies.
So, let's review the design.
Invariably the dipole is cut to formula, and hams typically obsess over whether the formula is accurate for their antenna height, wire diameter, stranded vs solid, insulation type and diameter, whether the loop around the insulator is counted etc... but in taking such a component focused view, they ignore the feed line as a critically important factor in the SYSTEM.
Fig 1 shows an NEC model of a 133' dipole at 10m above average ground. Feed point impedance is plotted, and resonance occurs at about 3.55MHz. A triumph for the formula, 133' gives us a resonant dipole, it works... or does it?
Lets take a system view, and include the 14m of 600Ω feed line.
Fig 2 shows impedance seen by the balun/ATU looking into 14m of feed line. Feed point impedance at the CW end of the band is 204+j1063.
Calculating the peak voltage at 1000W into the feed line using Calculate Vpk and Irms for given power at complex load , we get 3389V. This is a very high voltage that will tax most balun/ATUs.
Is there a fix?
In fairness, (Ford 2004) does say [c]hanging the length of the feed
line may resolve this problem
, though not explaining how or why. It is not
the only solution, and often not a practical solution (as in this case).
Shorter feed line results in lower voltage, and certain longer feed line lengths will give lower voltage. A method for designing for voltage is given at Avoiding flashover in baluns and ATUs . But it needn't be hit and miss, see that article for a design approach to possible solutions.
One of the solutions that will spring to mind is to use a 4:1 voltage balun (as fitted to probably most commercial ATUs). That might help to obtain a 'match', and it does reduce the voltage on the ATU side due to its transformation ratio, but its own losses are quite likely to consume substantial power with this kind of load and so reduce the power available to the feedline... so reducing the voltage at the expense of efficiency.
In this case, the OM shortened the dipole to about 100' (yes, that is G5RV length) and had a working and efficient system.
Feed line type and length is a critical component of this antenna system, yet they are ignored in most descriptions and recommendations.
Without even considering the higher bands, this antenna is not the nobrainer 'works every time' antenna that is commonly portrayed.
Version  Date  Description 
1.01  26/12/2011  Initial. 
1.02  
1.03  
1.04  
1.05 
© Copyright: Owen Duffy 1995, 2019. All rights reserved. Disclaimer.