He used single pole double throw switches to create a resonant dipole for each band 40-10 meters. The whole antenna is a full-sized 18 gauge dipole from 10-20 and the 30/40 meter inclusion simply inserted inductance in series with each leg across an open switch fed with RG 174 coax. A great antenna but I wanted to see if I could reduce some weight.
While I haven't created the necessary inductance to get to 40 meters, the result of my fiddling has yielded a full-sized dipole that is resonant on 6, 10, 12, 15, 17, 20, 30 meters and fits inside a quart-sized Baggie using a continuous length of twisted pair from a piece of CAT 5e cable.
Now, when I say "resonant", what I mean is very near 50 Ohms at the antenna terminals, which is an important distinction when using CAT 5e twisted pair. Its characteristic impedance is 100-Ohms and something inside your brain says that it won't be matched to a 50-Ohm system. You are correct but I will show you the way to make this as efficient as possible using an antenna tuner at the rig, and a switchable mini-balun, with the payoff of having a 3 ounce, 6-30 meter antenna system for your shiny new QRP rig, while only sacrificing 0.8 dB at the most!
The worst case is, in my case, on 10 meters where the EZNEC calculated Z is 42 - j37 Ohms. The VK1OD twisted pair calculator reveals that when a 42 - j37 Ohm load is placed at the end of 20.5 feet (6.25m) of twisted pair line you get 0.8 dB of mismatch loss in addition to 0.5 dB matched line loss for a total loss of 1.3 dB at the worst case frequency and worst case load impedance.
An 82% efficiency at the worst possible load and frequency, 10 meters, in my case. If you payed closer attention to getting 50 Ohms on all bands the only loss incurred would be the loss of the Tx line, balun and tuner.
Now, notice that the input impedance (Zin) is very near 200 Ohms which is a very efficient match when you switch the mini-balun to the 4:1 setting.
Adding up, systems losses in this case yields 0.5 dB line loss, 0.8 dB mismatch loss, 0.2 dB balun loss, 1.5 dB total system loss! A very acceptable trade considering portability, cost and efficiency.
A closer examination of the Smith chart reveals a neat relationship between a 50 Ohm source, 100 Ohm Tx line and 200 Ohm load. A reactive transformer (Q-match) is formed and used to our advantage. No matter where your impedance comes up with varying lengths of Tx line, the input Z will always be between 50 and 200 Ohms, an easy match for the switchable mini-balun attached to any auto tuner.
Now all that's left is to build one. If you pull the rip cord found inside most cables, you'll be able to free the pairs fairly quickly, the hard part is un-twisting the pair to your highest frequency and cut the remaining inter-switch wires to length and unravel those separately. BTW, don't use Phil's lengths in his document because they are inaccurate, use the 468/f(MHz) formula and you'll be fine.