The quarter wave cable sections should be made using top quality double screened cable with known properties. Only N-connectors of good quality should be used (as BNC can couse trouble when stressed or bended). The exact shortening factor of the used cable must be known. Total mechanical cable lengths (including connections) can be calculated from it.
The length of the piece of coax itself can be found by subtracting the mechanical length of the connecting plugs and T-pieces from the total mechanical length. So, looking at fig.3d, the total mechanical length of the 1/4 lambda sections should be determened between:
* the coupling link itself and the center of the T-piece
* the centers of the T-pieces.
The total physical cablelength l can be calculated by the formula:
l = (7500/f) . V [cm]
where V is the shorteningfactor of the cable + connectors used, and f is the frequency in MHz.
Determening the physical length of 1/4 wave sections using cable with unknown shortening factor.
A 6dB power pad connected directly on the transmitter output should protect the transmitter from severe VSWR during experiments. A long piece of lossy 50 ohms coax (RG58 or RG174) could also be used.
Shorten the cable in little pieces, until the RF-power meter shows a minimum.
The definitive total physical length of a finished cable must be measured including the physical length of all connections used. This is critical on 70cms.
So shorten the cable a little more, to compensate for the extra physical length of the connectors and T-pieces, which will be installed later. Example: about 2.5 cm for each (BNC connector+half BNC T-piece).