tuner setup

Most people don't seem to understand what a antenna tuner actually does.

" It will only make your rig see 50 Ohms and prevent it from cutting back the output power."
" It will not make the antenna any better, nor will it improve the SWR on the line "
Although all the above is true, this does NOT imply that whenever an antenna is not matched
to the line, power will be lost. In fact the power reflected from the antenna will be re-reflected
by the tuner and adds up with the output power of the rig. The forward power therefore is always
higher then the output power of the rig. This may seem strange to most people, but it can easily
be shown by measuring forward and reflected power at the tuner output.
If you think this is all nonsence please read Walter Maxwell's Another look at Reflections
On a loss less line no power will be lost. On practical lines, a higher SWR also means a higher line loss.
If this is a problem, one should consider using open line, with a relatively high characteristic impedance.
Since matched line loss is inversely proportional to the characteristic impedance (on HF) 450 Ohm
slotted line has about 10 times lower loss than a standard grade 50 Ohm coax.

Z-match

Most tuners use 3 variable components ( 1 L and 2 C's ) to match the impedance at the input of the line. However, only two components are necessary.  This means that an almost infinite number of settings will produce a match, but not all of those will give good efficiency.

Z-match circuit
 A simple L-network should be able to match any impedance that pops up at its terminals. Only one combination of component values will give a match, and the efficiency is always at its optimum.

The classical Z-match is a variant of a simple L-network. The Z-match consists of 2 coils both with a link coil. L3/L4 is used for 80m and 40m, and L1/L2 is used for 30m to 10 m. The open line feeders may be connected by a switch to either L2 or L4. Only 2 controls; C1 and C2 are used to obtain a match.
low frequency equivalent circuit
 The low frequency part of the circuit looks like a simple inductively coupled L-network. Coil L3 itself is fixed, but the inductance can be varied with the parallel capacitor. The advantage of this arrangement is that you don't need a variable coil, either roller type or switched, and that the effective inductance can be much higher (near resonance) than the inductance of L3 itself. Under these conditions however, the current trough the coil increases and the efficiency will drop.
Since it is fixed, L3 (and L4) can be made from thick, well conducting material to ensure high Q and minimal losses.
high frequency equivalent circuit
 The equivalent circuit of the high frequency part is not that different.
C2 is a split stator type with its rotor connected to ground. This means that the centre of L1 may be considered as a "virtual ground". Again the equivalent circuit acts as a simple L-network. The only difference is that L1 is fed at a "tap" Louis Varney, G5RV proposed to modify the original Z-match and feed both L3 and L1 at a tap close to ground. This would benefit the efficiency. I did try that, but could not find any significant difference between the original and the modified circuit (other then component value).